Page MenuHome GnuPG
Files F34102346

No OneTemporary
Actions

View Options

This file is larger than 256 KB, so syntax highlighting was skipped.
diff --git a/g10/ChangeLog b/g10/ChangeLog
index 88227e37a..7bbe53601 100644
--- a/g10/ChangeLog
+++ b/g10/ChangeLog
@@ -1,13603 +1,13608 @@
+2009-12-15 Werner Koch <wk@g10code.com>
+
+ * tdbio.c (tdbio_set_dbname): Do not call log_fatal after creating
+ the directory. Fixes bug#1169. Reported by Daniel Leidert.
+
2009-09-28 Werner Koch <wk@g10code.com>
* trustdb.c (get_validity_info): Take care of a NULL PK. Fixes
bug#1138.
(get_validity_string): Ditto.
2009-09-02 Werner Koch <wk@g10code.com>
* app-openpgp.c (do_decipher): Compute required Le.
* iso7816.c (iso7816_decipher): Add new arg LE.
* compress-bz2.c (do_uncompress): Detect unexpected EOF. Fixes
bug#1011.
2009-08-26 Werner Koch <wk@g10code.com>
* keyedit.c (menu_revsig): Check for signature right away. Fix
Debian-bug#543530.
2009-08-13 Werner Koch <wk@g10code.com>
* cardglue.h: Add new error code mappings.
From GnuPG-2:
* app-openpgp.c (change_keyattr_from_string): New.
(do_setattr): Support KEY-ATTR.
* ccid-driver.c (print_pr_data): Fix 64 bit compat problem.
* ccid-driver.c (parse_ccid_descriptor): Enable hack for SCR 3320.
* card-util.c (generate_card_keys): Remove special case for
GnuPG-2. Ask for the keysize and change it.
(card_generate_subkey): Ask for the keysize and change it.
(get_info_for_key_operation): Read KEY-ATTR.
(show_keysize_warning, ask_card_keysize): New.
(do_change_keysize): New.
2009-08-11 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_spawn): Try and detect mis-coded Latin1
and convert it to UTF8. Whether the heuristics succeed or not,
the resulting string must be valid UTF8 as LDAP requires that.
This is bug 1055.
2009-08-03 Werner Koch <wk@g10code.com>
* card-util.c (generate_card_keys): Ask for off-card keys only if
the card supports it.
(get_info_for_key_operation): Read EXTCAP.
(card_store_subkey): Check for non matching sizes.
* cardglue.c (agent_scd_writecert, agent_scd_readcert): New stubs.
* misc.c (not_in_gpg1_notice): New.
2009-07-30 Werner Koch <wk@g10code.com>
* misc.c (md5_digest_warn): New.
(print_digest_algo_note): Divert to new function.
* sig-check.c (do_check): Print MD5 warning.
2009-07-31 David Shaw <dshaw@jabberwocky.com>
* gpg.c (main): --pgp6 includes --disable-mdc.
2009-07-29 Werner Koch <wk@g10code.com>
* keygen.c (keygen_set_std_prefs): Remove RMD-160 from the list.
Change order to SHA-256, SHA-1, SHA-384, SHA-512, SHA-224.
(gen_dsa): Use a 256 bit Q for 2048 bit P. Round to FIPS allowed
values in non-expert mode.
(ask_algo): Add arg R_SUBKEY_ALGO. Change return value semantics.
Change presented order of algorithms. Make RSA+RSA the default.
(generate_keypair, generate_subkeypair): Adjust for change.
2009-07-23 Werner Koch <wk@g10code.com>
* keygen.c (generate_keypair): Allow Elgamal > 3072 in BOTH mode.
Reported by Jeroen Schot. Fixes bug#1091.
(ask_keysize): Add new arg PRIMARY_KEYSIZE. Set DSA default to
2048. Print a different prompt for a subkey.
(gen_dsa): Remove check for dsa2 option.
* trustdb.c (how_to_fix_the_trustdb): New.
* tdbio.c (tdbio_invalid): Print hints on how to fix the trustdb.
* gpg.c (main) <aFixTrustDB>: Print hints.
2009-07-23 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_import_ldap): Try a DNS-SD lookup to find
a domain-specific LDAP server before resorting to keys.{domain}.
2009-07-22 Werner Koch <wk@g10code.com>
* cardglue.h (struct agent_card_info_s): Add field EXTCAP.
* cardglue.c (agent_learn): Read KEY-ATTR.
(learn_status_cb): Parse EXTCAP.
2009-07-21 Werner Koch <wk@g10code.com>
* app-common.h, app-openpgp.c, iso7816.c, iso7816.h, apdu.c,
* apdu.h, ccid-driver.c, ccid-driver.h, card-util.c: Update from
GnuPG 2.0 SVN revision 5084.
* cardglue.h (GCRY_MD_SHA256): Add more GCRY_MD constants.
(gcry_handler_progress_t): Add definition.
(struct agent_card_info_s): Add fields apptype, is_v2, key_attr.
* cardglue.c (learn_status_cb): Set them.
(agent_release_card_info): Release APPTYPE.
(unescape_status_string, send_status_direct): New.
(gcry_mpi_release, gcry_mpi_set_opaque): New.
(gcry_md_algo_name): New.
(open_card): s/initialized/ref_count/.
(agent_learn): Pass new new flag arg to learn_status.
(agent_scd_genkey): Add new arg createtime.
* keygen.c (gen_card_key, gen_card_key_with_backup): Add new arg
TIMESTAMP.
(write_direct_sig, write_selfsigs, write_keybinding)
(make_backsig): Ditto.
(do_generate_keypair): Pass timestamp to all signing functions.
(generate_card_subkeypair): Ditto.
* keyedit.c (menu_backsign): Pass a new timestamp to all backsisg.
* gpg.c (main): Disable keypad support.
* options.h (struct): Add field disable_keypad.
2009-07-17 Werner Koch <wk@g10code.com>
* keyring.c (keyring_rebuild_cache): Replace the assert by a
proper error message and allow to delete a bad keyblock.
2009-07-16 Werner Koch <wk@g10code.com>
* misc.c (has_invalid_email_chars): Let non-ascii characters pass
through. Fixes bug#1087.
2009-06-24 Werner Koch <wk@g10code.com>
* passphrase.c (passphrase_to_dek): Do not deref a PW of NULL.
* revoke.c (gen_revoke): Remove unused malloc of PK.
2009-06-17 Werner Koch <wk@g10code.com>
* parse-packet.c (parse): Use a casted -1 instead of a 32 bit
constant to check for a garbled package. Fixes bug#1040.
2009-06-05 David Shaw <dshaw@jabberwocky.com>
* gpg.c (main): Remove Camellia restriction.
2009-06-02 Werner Koch <wk@g10code.com>
* gpgv.c (tty_cleanup_after_signal): Add new stub.
2009-05-22 Werner Koch <wk@g10code.com>
* signal.c (got_fatal_signal): Call tty_cleanup_after_signal.
2009-05-20 Werner Koch <wk@g10code.com>
* gpg.c (main): Fix --fingerprint/--with-fingerprint command
detection. Fixes bug#1044.
2009-05-11 Werner Koch <wk@g10code.com>
* apdu.c (send_le): Replace log_error by log_info. Fixes bug#1043.
* skclist.c (build_sk_list): Use log_info for "duplicated entry".
Fixes bug#1045.
* encode.c (encode_simple, encode_crypt): Print empty file warning
only in verbose mode. Closes bug#1039.
* sign.c (write_plaintext_packet):
2009-05-06 Werner Koch <wk@g10code.com>
* keyring.c (keyring_get_keyblock): Fix memory leak due to
ring_trust packets. Fixes bug#1034.
* getkey.c (finish_lookup): Remove dead code.
2009-05-05 Werner Koch <wk@g10code.com>
* keygen.c (read_parameter_file): Add keyword "Creation-Date".
(output_control_s): s/create/creation/.
(enum para_name): Add pCREATIONDATE, pKEYCREATIONDATE. Remove
pCREATETIME.
(generate_keypair): Do not set old pCREATETIME.
(parse_creation_string): New.
(proc_parameter_file): Set pCREATIONDATE.
(do_generate_keypair): Remove arg TIMESTAMP. Set it using
pKEYCREATIONDATE.
(get_parameter_u32): Set a default pKEYCREATIONDATE.
* tdbio.c (lookup_hashtable): Add const to function args.
(cmp_trec_fpr): Ditto.
(tdbio_search_trust_byfpr): Remove cast.
* parse-packet.c (parse): Remove special treatment for compressed
new style packets. Fixes bug#931.
2009-04-03 Werner Koch <wk@g10code.com>
* gpgv.c (main): Pass readonly flag to keydb_add_resource.
* keydb.c (keydb_add_resource): Add arg READONLY.
2009-03-20 David Shaw <dshaw@jabberwocky.com>
* keyring.c (rename_tmp_file): Force a fsync (via iobuf_ioctl) on
secret keyring files to be extra safe on filesystems that may not
sync data and metadata together (ext4). Also check return code
from the cache invalidation to make sure we're safe over NFS and
similar.
2009-03-14 David Shaw <dshaw@jabberwocky.com>
* gpgv.c (strusage): Fix name of program in "Syntax" line.
2009-02-09 Werner Koch <wk@g10code.com>
* keylist.c (print_capabilities): Take care of cert-only keys.
Fixes bug#998.
2009-02-03 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_typemap): gpgkeys_hkp handles hkps as
well.
2008-12-21 David Shaw <dshaw@jabberwocky.com>
* gpg.c (main): Properly handle UTF8 usernames with --sign-key and
--lsign-key.
2008-12-11 Werner Koch <wk@g10code.com>
* sig-check.c (check_revocation_keys): Close message digest.
(signature_check2): Switch to SHA-1 for SIG_ID computation. This
is to match 2.0.10.
2008-12-09 Werner Koch <wk@g10code.com>
* keygen.c (proc_parameter_file): Check that key and subkey usages
are allowed.
2008-11-18 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (validate_one_keyblock): Fix the trust signature
calculations so that we lower the trust depth of signatures to fit
within the current chain, rather than discarding any signature
that does not fit within the trust depth.
2008-10-03 David Shaw <dshaw@jabberwocky.com>
* main.h, mainproc.c (check_sig_and_print),
keylist.c (list_keyblock_print), pkclist.c (do_edit_ownertrust),
keyedit.c (menu_showphoto), photoid.c (generate_photo_id,
show_photos), misc.c (pct_expando): Add %v and %V expandos so
that displaying photo IDs can show the attribute validity
tag (%v) and string (%V). Originally by Daniel Gillmor.
2008-09-24 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu): Fix bug where a modified keyring loses
its modified status after a "clean" or "minimize" that doesn't
need to do anything.
2008-09-23 David Shaw <dshaw@jabberwocky.com>
* pkclist.c (select_algo_from_prefs): Redo function to rank prefs
and pick a consensus winner across all keys.
2008-09-04 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_import_cert): Allow keyserver URLs in
addition to full URLs in CERT records.
2008-08-11 Werner Koch <wk@g10code.com>
* keygen.c (ask_expire_interval): Check for time overflow of an
u32. Fixes bug #947. Use SIZEOF_TIME_T for the 2039--2106 check.
2008-08-01 Werner Koch <wk@g10code.com>
* tdbio.c (open_db) [!EROFS]: Move closing parens out of the
ifdef. Reported by Ken Takusagawa.
2008-04-30 Werner Koch <wk@g10code.com>
* getkey.c (parse_auto_key_locate): Ignore nodefault and local
methods.
2008-04-17 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (parse_key): Add constant for Camellia-192.
2008-04-12 David Shaw <dshaw@jabberwocky.com>
* getkey.c (merge_selfsigs_subkey): If there are multiple 0x19
backsigs, take the most recent one.
2008-03-25 Werner Koch <wk@g10code.com>
* keyserver.c (parse_keyrec): Take care of char defaulting to
unsigned when using hextobyte.
* options.skel: Make the default keyserver keys.gnupg.net.
2008-03-23 David Shaw <dshaw@jabberwocky.com>
* import.c (collapse_uids): Fix bug 894: possible memory
corruption around deduplication of user IDs.
2008-03-17 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (parse_key): Parse a secret key encrypted with
Camellia.
2008-02-09 Marcus Brinkmann <marcus@g10code.de>
* gpg.c (main): New variable default_configname. Use it if
save_configname is NULL (can happen if default configfile does
not exist). Move default configname determination to ...
(get_default_configname): ... this new function.
2008-01-26 Werner Koch <wk@g10code.com>
* card-util.c (get_manufacturer): Add vendor 004 and support for
the unmanaged range.
2007-12-14 Werner Koch <wk@g10code.com>
* gpg.c (main): Set opt.no_homedir_creation already on the first
option parsing pass. This is required so that the gpg_dearmor
tool won't create a ~/.gnupg.
2007-12-12 Werner Koch <wk@g10code.com>
* misc.c (print_pubkey_algo_note): Print a warning for type 20 keys.
(openpgp_pk_test_algo, openpgp_pk_algo_usage): Allow type 20 fro
encryption only with option --rfc2440.
* mainproc.c (proc_pubkey_enc): Ditto.
2007-12-12 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (sanitize_regexp): New. Protect against dangerous
regexps (malloc bombs) by force-commenting any characters aside
from the ones we explicitly want.
(check_regexp): Use it here before passing the regexp to
regcomp().
2007-12-11 Werner Koch <wk@g10code.com>
* mainproc.c (proc_pubkey_enc): Allow type 20 Elgamal key for
decryption.
2007-12-10 Werner Koch <wk@g10code.com>
* cardglue.h (gpg_error_from_syserror): New.
* app-openpgp.c (do_decipher): Take care of cryptograms shorter
that 128 bytes. Fixes bug#851.
* import.c (auto_create_card_key_stub): Do not clear the entire
fingerprint. This finally makes the stub creation work. My past
tests seemed to work because there was a key with a all zero
fingerprint available (Elgamal signing keys).
2007-12-08 Werner Koch <wk@g10code.com>
* misc.c (openpgp_pk_algo_usage): Allow Elgamal type 20 for
encryption.
2007-12-03 Werner Koch <wk@g10code.com>
* keygen.c (ask_key_flags): Add a translation remark and implement
a workaround.
2007-11-28 David Shaw <dshaw@jabberwocky.com>
* sig-check.c (do_check): Code to try both the incorrect and
correct SHA-224 DER prefixes when verifying a signature. See the
change itself for more discussion.
* main.h, seskey.c (do_encode_md): Rename to pkcs1_encode_md and
make non-static.
* gpg.c (print_algo_names): New.
(list_config): Use it here for the "ciphername" and "digestname"
config items so we can get a script-parseable list of the names.
2007-11-19 Werner Koch <wk@g10code.com>
* keyedit.c (keyedit_menu): String grammar fix.
2007-11-12 Werner Koch <wk@g10code.com>
* status.c (do_get_from_fd, do_shm_get): s/bool/getbool/ to
overcome problems with Mac OS 10.5 which seems to include
stdbool.h silently.
2007-10-27 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (parse_onepass_sig): Sigclass is hex, so include
the 0x.
* keyserver.c (keyserver_spawn): Cosmetic, but looks cleaner.
2007-10-23 Werner Koch <wk@g10code.com>
Switched entire package to GPLv3+.
2007-10-22 David Shaw <dshaw@jabberwocky.com>
* gpg.c (main): Add --require-cross-certification to
--openpgp/--rfc4880 mode.
* gpg.c (main): Disable --rfc2440-text and --force-v3-sigs by
default. Enable --require-cross-certification by default.
--openpgp (--rfc4880) is the same as --rfc2440 except with
"--enable-dsa2 --no-rfc2440-text --escape-from-lines".
2007-10-21 David Shaw <dshaw@jabberwocky.com>
* misc.c (compliance_option_string, compliance_failure): Minor
cleanup.
* armor.c (is_armor_header): Comment about 4880.
2007-10-17 David Shaw <dshaw@jabberwocky.com>
* options.h, gpg.c (main), misc.c (compliance_option_string): Add
--rfc4880, and make --openpgp an alias to it. --rfc2440 now
stands alone. For now, use the old 2440 defaults for 4880.
* keyedit.c (keyedit_menu): Use compliance_option_string() instead
of printing the compliance modes here.
2007-09-02 David Shaw <dshaw@jabberwocky.com>
* import.c (collapse_uids): Significant speedup for de-duping user
IDs.
2007-08-24 Werner Koch <wk@g10code.com>
* keyring.c (keyring_register_filename): Use same_file_p.
Suggested by Kurt Fitzner.
2007-07-19 David Shaw <dshaw@jabberwocky.com>
* gpg.c (main): Fix typo. Noted by John Clizbe.
* armor.c (parse_header_line): Improve test so that the header
test only allows "Hash" in the signed data section.
2007-07-18 David Shaw <dshaw@jabberwocky.com>
* armor.c (is_armor_tag): New. Detect if an armor header matches
2440bis-21.
(parse_header_line): Call it here, as bis-21 requires warning the
user (but continuing to process the message) when seeing an
unknown header.
2007-07-09 Werner Koch <wk@g10code.com>
* gpg.c: Better print an extra warning if Camellia has been
configured.
2007-07-06 David Shaw <dshaw@jabberwocky.com>
* encode.c (encode_crypt): Missed one call to
setup_plaintext_name(). This is bug#809.
2007-06-12 David Shaw <dshaw@jabberwocky.com>
* sign.c (mk_notation_policy_etc): expect all sigs that this is
called for are >=v4.
(write_signature_packets, make_keysig_packet): Only call it for
>=v4 sigs. This allows --force-v3-sigs and --force-v4-certs to
enable or disable notations, policies, and keyserver URLs. This
is bug#800.
2007-05-09 Werner Koch <wk@g10code.com>
* openfile.c (overwrite_filep, open_outfile) [W32]: Need to use
just "nul". Though, I am pretty sure that some MSDOS versions
grok the extra /dev/.
2007-05-07 Werner Koch <wk@g10code.com>
* openfile.c (open_outfile, overwrite_filep) [W32]: Use "/dev/nul".
2007-04-16 David Shaw <dshaw@jabberwocky.com>
* packet.h, mainproc.c (reset_literals_seen): New function to
reset the literals count.
* verify.c (verify_one_file), decrypt.c (decrypt_messages): Call
it here so we allow multiple literals in --multifile mode (in
different files - not concatenated together).
* gpg.c, pkclist.c: #include <strings.h> for strcasecmp if it is
present. Note that autoconf protects us against a strings.h that
cannot be used together with string.h.
2007-04-03 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (parse_marker): New. Enforce that the marker
contains 'P', 'G', 'P', and nothing but.
(parse): Call it here.
(skip_packet): No longer need to handle marker packets here.
2007-03-14 David Shaw <dshaw@jabberwocky.com>
* keyserver.c: Windows Vista doesn't grok X_OK and so fails
access() tests. Previous versions interpreted X_OK as F_OK
anyway, so we'll just use F_OK directly.
2007-03-08 Werner Koch <wk@g10code.com>
* plaintext.c (handle_plaintext): Add two extra fflush for stdout.
2007-03-06 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu): If we modify the keyblock (via
fix_keyblock() or collapse_uids()) make sure we reprocess the
keyblock so the flags are correct. Noted by Robin H. Johnson.
* getkey.c (fixup_uidnode): Properly clear flags that don't apply
to us (revoked, expired) so that we can reprocess a uid.
2007-03-01 David Shaw <dshaw@jabberwocky.com>
* options.h, gpg.c (main), mainproc.c (check_sig_and_print): Allow
multiple sig verification again as this is protected via the
multiple-messages code.
2007-02-26 Werner Koch <wk@g10code.com>
* gpg.c (main): Add verify option show-primary-uid-only.
* options.h (VERIFY_SHOW_PRIMARY_UID_ONLY): New.
* mainproc.c (check_sig_and_print): Implement it.
* status.h (STATUS_ERROR): New status code.
* status.c (get_status_string): Ditto.
* mainproc.c (proc_plaintext): Emit it if multiple messages are
detected.
2007-02-20 David Shaw <dshaw@jabberwocky.com>
* mainproc.c (literals_seen): New.
(proc_plaintext): Error out if more than one plaintext packet is
encountered
* options.h, gpg.c: New option --allow-multiple-messages and --no
variant.
2007-02-13 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (parse_signature): It's hex.
* getkey.c (merge_selfsigs_subkey): Avoid listing the contents of
a backsig when list mode is on. Noted by Timo Schulz.
2007-02-04 Werner Koch <wk@g10code.com>
* parse-packet.c (parse_signature): Limit bytes read for an
unknown alogorithm. Fixes Debian bug#402592.
2007-02-01 David Shaw <dshaw@jabberwocky.com>
* main.h, keygen.c (ask_expire_interval, parse_expire_string):
Pass in the time to use to calculate the expiration offset, rather
than querying it internally. Change all callers.
2007-01-31 David Shaw <dshaw@jabberwocky.com>
* keygen.c (do_generate_keypair, proc_parameter_file)
(generate_keypair, generate_subkeypair): Pass a timestamp through
to all the gen_xxx functions.
* keyedit.c (sign_uids): Another multiple to single timestamp
operation.
* sign.c (write_plaintext_packet): Take timestamp from outside.
Change all callers.
(sign_file, clearsign_file, sign_symencrypt_file): Calculate one
timestamp for the whole transaction. Different timestamps for
each step can leak performance information.
* main.h, keygen.c (ask_expiredate), keyedit.c (menu_expire):
Remove the need for ask_expiredate().
2007-01-30 David Shaw <dshaw@jabberwocky.com>
* keygen.c (make_backsig, write_direct_sig, write_selfsigs)
(write_keybinding): Have the various selfsigs created during key
generation use the timestamp of their parent key. This helps
avoid leaking information about the performance of the machine
where the keys were generated. Noted by Christian Danner.
2007-01-15 Werner Koch <wk@g10code.com>
* parse-packet.c (read_protected_v3_mpi): Make sure to stop
reading even for corrupted packets.
* keygen.c (generate_user_id): Need to allocate one byte more.
Reported by Felix von Leitner.
2006-12-14 David Shaw <dshaw@jabberwocky.com>
* options.skel: Tweak some examples to match reality and update
the RFC for CERT now that it is out of draft.
2006-12-13 David Shaw <dshaw@jabberwocky.com>
* Makefile.am: Install options.skel via dist_pkgdata_DATA so that
"make uninstall" works properly.
2006-12-13 Werner Koch <wk@g10code.com>
* gpgv.c, gpg.c: Do not include the now removed g10defs.h file.
2006-12-11 Werner Koch <wk@g10code.com>
* seskey.c (encode_session_key, do_encode_md): Use new
mpi_nlimb_hint_from_nbytes function.
* sign.c (do_sign): Ditto.
* Makefile.am (AM_CPPFLAGS): Define GNUPG_LIBDIR.
* gpgv.c (i18n_init): s/G10_LOCALEDIR/LOCALEDIR/.
* gpg.c (i18n_init): Ditto.
2006-12-07 Werner Koch <wk@g10code.com>
* Makefile.am (AM_CPPFLAGS): Define GNUPG_DATADIR.
* openfile.c (copy_options_file): Use log_info and not log_error
to avoid error returns in case of a missing skeleton file.
2006-12-04 Werner Koch <wk@g10code.com>
* filter.h (armor_filter_context_t): New field REFCOUNT.
* armor.c (new_armor_context, release_armor_context)
(push_armor_filter): New.
(armor_filter): Call releae_armor_context for IOBUFCTRL_FREE.
* import.c (import): Use the new function here instead of the
old hack using the iobuf_push_filter2.
* keyserver.c (keyserver_spawn): Ditto.
2006-12-03 Werner Koch <wk@g10code.com>
* keyedit.c (menu_clean): Made strings translatable.
2006-12-03 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_clean): Show "already minimized" rather than
"already clean" when a minimized key is minimized again. From
Dirk Traulsen.
2006-12-02 David Shaw <dshaw@jabberwocky.com>
* options.h, gpg.c (main), passphrase.c (passphrase_to_dek): Add
--passphrase-repeat option to control how many times gpg will
re-prompt for a passphrase to ensure the user has typed it
correctly. Defaults to 1.
2006-12-02 Werner Koch <wk@g10code.com>
* encr-data.c: Allocate DFX context on the heap and not on the
stack. Changes at several places. Fixes CVE-2006-6235.
2006-11-27 Werner Koch <wk@g10code.com>
* openfile.c (ask_outfile_name): Fixed buffer overflow occurring
if make_printable_string returns a longer string. Fixes bug 728.
2006-10-23 Werner Koch <wk@g10code.com>
* gpg.c (main): New command --gpgconf-list.
2006-10-12 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (parse_symkeyenc): Show the unpacked as well as
the packed s2k iteration count.
* main.h, options.h, gpg.c (encode_s2k_iterations, main),
passphrase.c (hash_passphrase): Add --s2k-count option to specify
the number of s2k hash iterations.
2006-10-06 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_spawn): Write the 16-digit keyid rather
than whatever key selector the user used on the command line.
2006-10-02 Werner Koch <wk@g10code.com>
* encr-data.c (decrypt_data, mdc_decode_filter): Check the MDC
right here and don't let parse-packet handle the MDC.
2006-08-21 Werner Koch <wk@g10code.com>
* skclist.c (is_insecure): Also test for uppercase version of the
insecure string.
2006-07-31 Werner Koch <wk@g10code.com>
* openfile.c (open_outfile) [USE_ONLY_8DOT3]: Search backwards for
the dot. Fixes bug 654.
* passphrase.c (agent_open): Use log_info instead of log_error to
allow a fallback without having gpg return an error code. Fixes
bug #655.
* encode.c (encode_crypt_files): Invalidate the whole fd cache.
This is a workaround for problems in iobuf's stupid fd cache.
* decrypt.c (decrypt_messages): Ditto.
* verify.c (verify_files): Ditto.
2006-07-26 Werner Koch <wk@g10code.com>
* keygen.c (gen_card_key_with_backup): Initialize sk_{un}protected.
* import.c (import): Initialize KEYBLOCK.
* pkclist.c (edit_ownertrust): Intialize trust to avoid gcc
warning.
* parse-packet.c (parse_comment): Cap comments at 65k.
(parse_gpg_control): Skip too large control packets.
2006-06-28 David Shaw <dshaw@jabberwocky.com>
* keydb.h, pkclist.c (select_algo_from_prefs, algo_available):
Pass a union for preference hints rather than doing void * games.
* sign.c (sign_file): Use it here.
* sign.c (sign_file): When signing with multiple DSA keys, one
being DSA1 and one being DSA2 and encrypting at the same time, if
the recipient preferences give a hash that can work with the DSA2
key, then allow the DSA1 key to be promoted rather than giving up
and using hash_for().
* pkclist.c (algo_available): Automatically enable DSA2 mode when
handling a key that clearly isn't DSA1 (i.e. q!=160).
2006-06-28 Werner Koch <wk@g10code.com>
* import.c (check_prefs_warning): Fix change for better
translatability.
* app-openpgp.c (do_writekey): Fixed computation of memmove
length. This led to garbled keys if E was larger than one byte.
Thanks to Achim Pietig for hinting at the garbled E.
2006-06-27 Werner Koch <wk@g10code.com>
* gpg.c (reopen_std) [HAVE_W32_SYSTEM]: Do not use it.
2006-06-22 David Shaw <dshaw@jabberwocky.com>
* options.h, gpg.c (main), keygen.c (keygen_upd_std_prefs,
keygen_add_std_prefs, proc_parameter_file): Add
--default-keyserver-url to specify a keyserver URL at key
generation time, and "Keyserver:" keyword for doing the same
through a batch file.
* sign.c (do_sign): Accept a truncated hash even for DSA1 keys (be
liberal in what you accept, etc).
2006-06-12 David Shaw <dshaw@jabberwocky.com>
* import.c (import_one): Add a flag (from_sk) so we don't check
prefs on an autoconverted public key. The check should only
happen on the sk side. Noted by Dirk Traulsen.
2006-06-09 David Shaw <dshaw@jabberwocky.com>
* keygen.c (gen_card_key): Add optional argument to return a
pointer (not a copy) of the stub secret key for the secret key we
just generated on the card.
(generate_card_subkeypair): Use it here so that the signing key on
the card can use the card to generate the 0x19 backsig on the
primary key. Noted by Janko Heilgeist and Jonas Oberg.
* parse-packet.c (parse_user_id): Cap the user ID size at 2048
bytes. This prevents a memory allocation attack with a very large
user ID. A very large packet length could even cause the
allocation (a u32) to wrap around to a small number. Noted by
Evgeny Legerov on full-disclosure.
2006-05-25 David Shaw <dshaw@jabberwocky.com>
* keygen.c (gen_dsa): Allow generating DSA2 keys
(allow specifying sizes > 1024 when --enable-dsa2 is set). The
size of q is set automatically based on the key size.
(ask_keysize, generate_keypair): Ask for DSA size when
--enable-dsa2 is set.
2006-05-24 David Shaw <dshaw@jabberwocky.com>
* exec.c (make_tempdir): Fix bug with a temporary directory on
Win32 that is over 256 bytes long. Noted by Israel G. Lugo.
2006-05-23 David Shaw <dshaw@jabberwocky.com>
* gpg.c (reopen_std): New function to reopen fd 0, 1, or 2 if we
are called with them closed. This is to protect our
keyring/trustdb files from corruption if they get attached to one
of the standard fds. Print a warning if possible that this has
happened, and fail completely if we cannot reopen (should never
happen).
(main): Call it here.
2006-05-22 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (dump_sig_subpkt, parse_signature),
build-packet.c (build_sig_subpkt_from_sig), getkey.c
(fixup_uidnode, merge_selfsigs_main, merge_selfsigs_subkey),
keygen.c (keygen_add_key_expire): Fix meaning of key expiration
and sig expiration subpackets - zero means "never expire"
according to 2440, not "expire instantly".
* getkey.c (get_pubkey_byname), import.c (import_one): Fix key
selection problem when auto-key-locate returns a list of keys, not
all of which are usable (revoked, expired, etc). Noted by Simon
Josefsson.
2006-04-26 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (direct_uri_map): New.
(keyserver_spawn): Used here to add "_uri" to certain gpgkeys_xxx
helpers when the meaning is different if a path is provided
(i.e. ldap).
(keyserver_import_cert): Show warning if there is a CERT
fingerprint, but no --keyserver set.
2006-04-22 David Shaw <dshaw@jabberwocky.com>
* keyserver.c: Fix build problem with platforms that stick libcurl
in a place not in the regular include search path.
2006-04-20 David Shaw <dshaw@jabberwocky.com>
* options.h, gpg.c (main): Add --enable-dsa2 and --disable-dsa2.
Defaults to disable.
* pkclist.c (algo_available): If --enable-dsa2 is set, we're
allowed to truncate hashes to fit DSA keys.
* sign.c (match_dsa_hash): New. Return the best match hash for a
given q size.
(do_sign, hash_for, sign_file): When signing with a DSA key, if it
has q==160, assume it is an old DSA key and don't allow truncation
unless --enable-dsa2 is also set. q!=160 always allows truncation
since they must be DSA2 keys.
(make_keysig_packet): If the user doesn't specify a
--cert-digest-algo, use match_dsa_hash to pick the best hash for
key signatures.
2006-04-19 David Shaw <dshaw@jabberwocky.com>
* gpg.c (print_mds), armor.c (armor_filter, parse_hash_header):
Add SHA-224.
* sign.c (write_plaintext_packet), encode.c (encode_simple):
Factor common literal packet setup code from here, to...
* main.h, plaintext.c (setup_plaintext_name): Here. New. Make sure
the literal packet filename field is UTF-8 encoded.
* options.h, gpg.c (main): Make sure --set-filename is UTF-8
encoded and note when filenames are already UTF-8.
2006-04-18 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_backsign): Give some more verbose errors when we
have no need to backsign.
2006-04-11 David Shaw <dshaw@jabberwocky.com>
* options.skel, photoid.c (get_default_photo_command): Find an
image viewer at runtime. Seems FC5 doesn't have xloadimage.
2006-04-08 David Shaw <dshaw@jabberwocky.com>
* getkey.c (parse_auto_key_locate): Fix dupe-removal code.
* keyedit.c (menu_backsign): Allow backsigning even if the secret
subkey doesn't have a binding signature.
* armor.c (radix64_read): Don't report EOF when reading only a pad
(=) character. The EOF actually starts after the pad.
* gpg.c (main): Make --export, --send-keys, --recv-keys,
--refresh-keys, and --fetch-keys follow their arguments from left
to right. Suggested by Peter Palfrader.
2006-04-08 Werner Koch <wk@g10code.com>
* mainproc.c (list_node): Print ring trust value only if not empty
and --with-colons has been given.
2006-04-05 Werner Koch <wk@g10code.com>
* getkey.c (user_id_not_found_utf8): New.
(get_primary_uid, get_user_id): Use it. Fixes Debian bug #205028
in the right way.
2006-04-03 Werner Koch <wk@g10code.com>
* import.c (check_prefs_warning): Merged strings for better
translation.
* gpg.c (main) [__GLIBC__]: Default to libpcsclite.so.1.
* status.h, status.c (STATUS_BEGIN_SIGNING): New. Suggested by
Daiki Ueno.
* textfilter.c (copy_clearsig_text): Issue new status code.
* sign.c (sign_file, sign_symencrypt_file): Ditto.
2006-03-31 David Shaw <dshaw@jabberwocky.com>
* getkey.c (get_pubkey_byname): Fix missing auto_key_retrieve
unlock. Fix strings to not start with a capital letter as per
convention.
2006-03-30 David Shaw <dshaw@jabberwocky.com>
* main.h, seskey.c (encode_md_value): Modify to allow a q size
greater than 160 bits as per DSA2. This will allow us to verify
and issue DSA2 signatures for some backwards compatibility once we
start generating DSA2 keys.
* sign.c (do_sign), sig-check.c (do_check): Change all callers.
* sign.c (do_sign): Enforce the 160-bit check for new signatures
here since encode_md_value can handle non-160-bit digests now.
This will need to come out once the standard for DSA2 is firmed
up.
2006-03-28 Werner Koch <wk@g10code.com>
* openfile.c (overwrite_filep): Fix small cpr issue. Noted by
Daiki Ueno.
2006-03-22 David Shaw <dshaw@jabberwocky.com>
* getkey.c (parse_auto_key_locate): Silently strip out duplicates
rather than causing an error.
2006-03-22 Werner Koch <wk@g10code.com>
* sig-check.c (signature_check2): Changed warning URL to include faq.
* misc.c (idea_cipher_warn): Ditto.
2006-03-22 David Shaw <dshaw@jabberwocky.com>
* mainproc.c (get_pka_address): Fix bug introduced as part of
sig_to_notation conversion. Noted by Peter Palfradrer.
2006-03-21 Werner Koch <wk@g10code.com>
* cardglue.c (agent_scd_pksign): Allow the use of ripemd-160 along
with scdaemon.
2006-03-16 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_import_cert): Handle the IPGP CERT type
for both the fingerprint alone, and fingerprint+URL cases.
* getkey.c (get_pubkey_byname): Minor cleanup.
2006-03-13 David Shaw <dshaw@jabberwocky.com>
* keyserver-internal.h, keyserver.c (keyserver_import_pka): Use
the same API as the other auto-key-locate fetchers.
* getkey.c (get_pubkey_byname): Use the fingerprint of the key
that we actually fetched. This helps prevent problems where the
key that we fetched doesn't have the same name that we used to
fetch it. In the case of CERT and PKA, this is an actual security
requirement as the URL might point to a key put in by an attacker.
By forcing the use of the fingerprint, we won't use the attacker's
key here.
* keyserver-internal.h, keyserver.c (keyserver_spawn,
keyserver_work, keyserver_import_cert, keyserver_import_name,
keyserver_import_ldap): Pass fingerprint info through.
* main.h, import.c (import_one): Optionally return the fingerprint
of the key being imported.
(import_keys_internal, import_keys_stream, import): Change all
callers.
2006-03-12 David Shaw <dshaw@jabberwocky.com>
* sig-check.c (signature_check2): Print the backsig warning when
there is no backsig present. Give a URL for more information.
* keyedit.c (menu_backsign): Small tweak to work properly with
keys originally generated with older GnuPGs that included comments
in the secret keys.
2006-03-10 Werner Koch <wk@g10code.com>
* card-util.c (get_manufacturer): Added Vendor 3
2006-03-09 David Shaw <dshaw@jabberwocky.com>
* build-packet.c (string_to_notation): Add ability to indicate a
notation to be deleted with a '-' prefix.
* keyedit.c (menu_set_notation): Use it here to allow deleting a
notation marked with '-'. This works with either "-notation" or
"-notation=value".
2006-03-08 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_set_notation): New function to set notations on
self-signatures.
(keyedit_menu): Call it here.
(tty_print_notations): Helper.
(show_prefs): Show notations in "showpref".
* mainproc.c (get_pka_address)
* keylist.c (show_notation): Remove
duplicate code by using notation functions.
* packet.h, build-packet.c (sig_to_notation)
* keygen.c (keygen_add_notations): Provide printable text for
non-human-readable notation values.
* packet.h, build-packet.c (sig_to_notation)
* keygen.c (keygen_add_notations): Tweak to handle non-human-readable
notation values.
* options.h, sign.c (mk_notation_policy_etc)
* gpg.c (add_notation_data): Use it here for the various notation
commands.
* packet.h, main.h, keygen.c (keygen_add_notations)
* build-packet.c (string_to_notation, sig_to_notation)
(free_notation): New "one stop shopping" functions to handle
notations and start removing some code duplication.
2006-03-08 Werner Koch <wk@g10code.com>
* mainproc.c (do_check_sig): Use log_error for standalone revocations.
2006-03-07 David Shaw <dshaw@jabberwocky.com>
* options.h, mainproc.c (check_sig_and_print), gpg.c (main):
pka-lookups, not pka-lookup.
* options.h, gpg.c (main), keyedit.c [cmds], sig-check.c
(signature_check2): Rename "backsign" to "cross-certify" as a more
accurate name.
* options.h, gpg.c (main, parse_trust_model), pkclist.c
(check_signatures_trust), mainproc.c (check_sig_and_print,
pka_uri_from_sig), trustdb.c (init_trustdb): Some tweaks to PKA so
that it is a verify-option now.
2006-03-07 Werner Koch <wk@g10code.com>
* mainproc.c (proc_signature_packets): Return any_sig_seen to caller.
(check_sig_and_print): Option to partly allow the old behaviour.
* gpg.c: New option --allow-multisig-verification.
2006-03-06 David Shaw <dshaw@jabberwocky.com>
* sign.c (make_keysig_packet): Don't use MD5 for a RSA_S key as
that is not a PGP 2.x algorithm.
* mainproc.c (proc_compressed): "Uncompressed" is not a valid
compression algorithm.
2006-03-06 Werner Koch <wk@g10code.com>
* mainproc.c (check_sig_and_print): Made the composition test more
tight. This is due to another bug report by Tavis Ormandy.
(add_onepass_sig): Simplified.
2006-03-05 Werner Koch <wk@g10code.com>
* plaintext.c (handle_plaintext): Replace assert by explict error
conflict message. Reported by Tavis Ormandy.
2006-03-02 Werner Koch <wk@g10code.com>
* cardglue.c (check_card_serialno): Don't ask in batch mode.
2006-03-01 David Shaw <dshaw@jabberwocky.com>
* getkey.c (parse_auto_key_locate): Error if the user selects
"cert" or "pka" when those features are disabled.
* misc.c (has_invalid_email_chars): Fix some C syntax that broke
the compilers on SGI IRIX MIPS and Compaq/DEC OSF/1 Alpha. Noted
by Nelson H. F. Beebe.
2006-02-27 David Shaw <dshaw@jabberwocky.com>
* options.skel: Document auto-key-locate and give a pointer to
Simon Josefsson's page for CERT.
2006-02-24 David Shaw <dshaw@jabberwocky.com>
* keydb.h, getkey.c (release_akl), gpg.c (main): Add
--no-auto-key-locate.
* options.h, gpg.c (main): Keep track of each keyserver registered
so we can match on them later.
* keyserver-internal.h, keyserver.c (cmp_keyserver_spec,
keyserver_match), gpgv.c: New. Find a keyserver that matches ours
and return its spec.
* getkey.c (get_pubkey_byname): Use it here to get the
per-keyserver options from an earlier keyserver.
2006-02-23 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (parse_keyserver_options): Only change max_cert if
it is used.
* options.c, gpg.c (main), keyserver.c (keyserver_spawn): No
special treatment of include-revoked, include-subkeys, and
try-dns-srv. These are keyserver features, and GPG shouldn't get
involved here.
* keyserver.c (parse_keyserver_uri, add_canonical_option): Always
append options to the list, as ordering may be significant to the
user.
* gpg.c (add_notation_data): Fix reversed logic for isascii check
when adding notations. Noted by Christian Biere.
* options.h, keyserver.c (add_canonical_option): New.
(parse_keyserver_options): Moved from here.
(parse_keyserver_uri): Use it here so each keyserver can have some
private options in addition to the main keyserver-options
(e.g. per-keyserver auth).
2006-02-22 David Shaw <dshaw@jabberwocky.com>
* options.h, keyserver-internal.h, keyserver.c
(keyserver_import_name), getkey.c (free_akl,
parse_auto_key_locate, get_pubkey_byname): The obvious next step:
allow arbitrary keyservers in the auto-key-locate list.
* options.h, keyserver.c (parse_keyserver_options): Remove
auto-cert-retrieve as it is no longer meaningful. Add
max-cert-size to allow users to pick a max key size retrieved via
CERT.
* options.h, gpg.c (main), mainproc.c (check_sig_and_print),
keyserver.c (keyserver_opts): Rename auto-pka-retrieve to
honor-pka-record to be consistent with honor-keyserver-url.
* options.h, keydb.h, g10.c (main), getkey.c
(parse_auto_key_locate): Parse a list of key access methods.
(get_pubkey_byname): Walk the list here to try and retrieve keys
we don't have locally.
2006-02-21 David Shaw <dshaw@jabberwocky.com>
* getkey.c (get_pubkey_byname): Fix minor security problem with
PKA when importing at -r time. The URL in the PKA record may
point to a key put in by an attacker. Fix is to use the
fingerprint from the PKA record as the recipient. This ensures
that the PKA record is followed.
* keyserver-internal.h, keyserver.c (keyserver_import_pka): Return
the fingerprint we requested.
* gpgv.c: Stub keyserver_import_ldap.
* keyserver-internal.h, keyserver.c (keyserver_import_ldap):
Import using the PGP Universal trick of asking
ldap://keys.(maildomain) for the key.
2006-02-20 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (parse_keyserver_uri): Include the scheme in the uri
even when we've assumed "hkp" when there was no scheme.
2006-02-20 Werner Koch <wk@g10code.com>
* apdu.c (open_pcsc_reader): As a precaution set LIST to NULL
after free.
2006-02-14 Werner Koch <wk@gnupg.org>
* verify.c (verify_signatures): Print warning also for NO_DATA.
* mainproc.c (struct mainproc_context): New field any_sig_seen.
(add_signature): Set it.
(proc_signature_packets): Test and return NO_DATA.
2006-02-09 Werner Koch <wk@g10code.com>
* gpg.c (main) <oLockNever>: Disable random locking.
2006-02-06 Werner Koch <wk@g10code.com>
* ccid-driver.c, ccid-driver.h: Updated from GnuPG 1.9. Changes:
* ccid-driver.h (CCID_DRIVER_ERR_NO_KEYPAD): New.
* ccid-driver.c (send_escape_cmd): New args RESULT, RESULTLEN and
RESULTMAX. Changed all callers.
(ccid_transceive_escape): New.
* ccid-driver.c (special_transport): New
(ccid_open_reader, do_close_reader, ccid_shutdown_reader)
(bulk_out, bulk_in): Add support for CardMan 4040 reader.
* ccid-driver.c (scan_or_find_devices): Factored most code out to
(scan_or_find_usb_device): .. new.
(make_reader_id): Fixed vendor mask.
2006-01-24 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (parse_keyserver_uri): If there is a path present,
set the direct_uri flag so the right keyserver helper is run.
2006-01-22 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_spawn): Include the EXEEXT so we can find
keyserver helpers on systems that use extensions.
* misc.c (path_access) [HAVE_DRIVE_LETTERS]: Do the right thing
with drive letter systems.
2006-01-17 David Shaw <dshaw@jabberwocky.com>
* keydb.h, passphrase.c (next_to_last_passphrase): New. "Touch" a
passphrase as if it was used (move from next_pw to last_pw).
* pubkey-enc.c (get_session_key): Use it here to handle the case
where a passphrase happens to be correct for a secret key, but yet
that key isn't the anonymous recipient (i.e. the secret key could
be decrypted, but not the session key). This also handles the
case where a secret key is located on a card and a secret key with
no passphrase. Note this does not fix bug 594 (anonymous
recipients on smartcard do not work) - it just prevents the
anonymous search from stopping when the card is encountered.
2006-01-07 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_refresh): Fix problem when more than one
key in a refresh batch has a preferred keyserver set. Noted by
Nicolas Rachinsky.
2006-01-01 David Shaw <dshaw@jabberwocky.com>
* mainproc.c (check_sig_and_print), keyserver.c
(keyserver_import_pka), card-util.c (fetch_url): Always require a
scheme:// for keyserver URLs except when used as part of the
--keyserver command for backwards compatibility.
* sign.c (write_signature_packets): Lost a digest_algo line.
* sign.c (hash_for): Add code to detect if the sk lives on a smart
card. If it does, only allow 160-bit hashes, a la DSA. This
involves passing the *sk in, so change all callers. This is
correct for today, given the current 160-bit q in DSA, and the
current SHA-1/RIPEMD160 support in the openpgp card. It will
almost certainly need changing down the road.
* app-openpgp.c (do_sign): Give user error if hash algorithm is
not supported by the card.
2005-12-23 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_import_pka): New. Moved from
getkey.c:get_pubkey_byname which was getting crowded.
* keyserver.c (keyserver_import_cert): Import a key found in DNS
via CERT records. Can handle both the PGP (actual key) and IPGP
(URL) CERT types.
* getkey.c (get_pubkey_byname): Call them both here.
* options.h, keyserver.c (parse_keyserver_options): Add
"auto-cert-retrieve" option with optional max size argument.
* gpgv.c: Stubs.
* keyserver-internal.h, keyserver.c (keyserver_spawn,
keyserver_work, keygerver_getname): New keyserver_getname function
to fetch keys by name.
* getkey.c (get_pubkey_byname): Call it here to enable locating
keys by full mailbox from a keyserver a la PKA. Try PKA first,
though, as it is likely to be faster.
2005-12-20 Werner Koch <wk@g10code.com>
* gpg.c: New option --allow-pka-lookup.
(parse_trust_model): Add "+pka" variants.
(main): Make KEYSERVER_AUTO_PKA_RETRIEVE teh default.
* options.h (opt): New fields PKA_TRUST_INCREASE and
ALLOW_PKA_LOOKUP.
* status.h (STATUS_PKA_TRUST_BAD, STATUS_PKA_TRUST_GOOD): New.
* pkclist.c (check_signatures_trust): Increase trust due to valid
PKA only if that new option has been set. Issue new status lines.
* trustdb.c (init_trustdb): Print info if this option is active.
* getkey.c (get_pubkey_byname): Honor allow-pka-lookup.
* mainproc.c (pka_uri_from_sig): Ditto.
* trustdb.c (validate_keys): Print no "ultimately trusted keys
found" only in non-quiet mode.
2005-12-19 David Shaw <dshaw@jabberwocky.com>
* getkey.c (merge_selfsigs_main): All primary keys can certify.
2005-12-18 David Shaw <dshaw@jabberwocky.com>
* gpg.c (main): Restore convert-sk-to-pk as programs rely on it.
* keyid.c (usagestr_from_pk): Remove special PUBKEY_USAGE_CERT
flag. It's no longer needed.
2005-12-14 David Shaw <dshaw@jabberwocky.com>
* gpg.c (main): Don't default to import-options convert-sk-to-pk.
It causes confusing warning messages when importing a PGP-exported
key that contains a secret key without selfsigs followed by the
public key.
2005-12-08 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_fetch): Switch on fast-import before we
--fetch-keys so we don't rebuild the trustdb after each fetch.
2005-12-08 Werner Koch <wk@g10code.com>
* gpg.c (main): Check for DBCS lead byte when converting the
homedir. By Kazuyoshi Kakihara. Fixes PR561.
* keyserver.c (keyserver_fetch): Made strings translatable.
2005-12-08 David Shaw <dshaw@jabberwocky.com>
* options.h, keyserver.c (curl_cant_handle, keyserver_spawn)
(keyserver_fetch): Set a flag to indicate that we're doing a direct
URI fetch so we can differentiate between a keyserver operation
and a URI fetch for protocols like LDAP that can do either.
2005-12-07 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_spawn): Don't print "searching for key
00000000" when fetching a URI.
* keyserver-internal.h, keyserver.c (keyserver_fetch): New. Fetch
an arbitrary URI using the keyserver helpers.
* gpg.c (main): Call it from here for --fetch-keys.
2005-12-07 Werner Koch <wk@g10code.com>
* pkclist.c (do_we_trust): Add NOTREACHED comment.
2005-11-20 David Shaw <dshaw@jabberwocky.com>
* main.h, keylist.c (print_revokers): New. Print the "rvk"
designated revoker record. Moved from
keyedit.c:show_key_with_all_names_colon.
* keylist.c (list_keyblock_colon): Use it here ...
* keyedit.c (show_key_with_all_names_colon): ... and here.
2005-11-19 David Shaw <dshaw@jabberwocky.com>
* free-packet.c (copy_secret_key): Copy secret key into secure
memory since we may unprotect it.
* main.h, g10.c (main), revoke.c (gen_desig_revoke): Add local
user support so users can use -u with --desig-revoke. This
bypasses the interactive walk over the revocation keys.
2005-11-17 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu, menu_clean): Simplify clean options to
just "clean", and add "minimize".
* import.c (parse_import_options): Make help text match the export
versions of the options.
* options.h, export.c (parse_export_options, do_export_stream):
Reduce clean options to two: clean and minimize.
* trustdb.h, trustdb.c (clean_one_uid): New function that joins
uid and sig cleaning into one for a simple API outside trustdb.
2005-11-13 David Shaw <dshaw@jabberwocky.com>
* armor.c (parse_header_line): A fussy bit of 2440: header lines
are delimited with a colon-space pair. Therefore a line such as
"Comment: " (with a trailing space) is actually legal, albeit not
particularly useful.
2005-11-11 David Shaw <dshaw@jabberwocky.com>
* trustdb.h, trustdb.c (clean_key): New function to handle key
cleaning from one convenient place.
* options.h, import.c (parse_import_options,
clean_sigs_from_all_uids, import_one): Reduce clean options to
two: clean and minimize.
* parse-packet.c (setup_user_id): Remove.
(parse_user_id, parse_attribute): Just use xmalloc_clear instead.
* trustdb.c (clean_uid_from_key, clean_uids_from_key):
Significantly simpler implementation.
2005-11-10 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu, menu_clean_sigs_from_uids): Add
"minimize" command.
* packet.h, keyedit.c (menu_clean_uids_from_key), trustdb.c
(clean_uids_from_key): Fix display bug where sigs cleaned for
other reasons caused a uid to appear as if it had been compacted.
* packet.h: Move some flags to a bitfield. Change all callers.
* options.h, import.c (parse_import_options,
clean_sigs_from_all_uids, import_one): Add import-minimal option.
Similar to export-minimal, except it works on the way in.
* trustdb.h, trustdb.c (clean_sigs_from_uid): Add flag to remove
all non-selfsigs from key during cleaning. Change all callers.
* export.c (do_export_stream): Use it here for export-minimal so
we don't need additional minimize code in the export path.
2005-11-06 David Shaw <dshaw@jabberwocky.com>
* options.skel: Add a section for --encrypt-to. This is Debian
bug 336211 by Javier Fernández-Sanguino Peña.
2005-11-05 David Shaw <dshaw@jabberwocky.com>
* Makefile.am: Include @LIBUSB_CPPFLAGS@ in our CPPFLAGS.
Strictly speaking this should be only in gpg_CPPFLAGS, but then we
have to compile everything twice for gpg and gpgv.
* apdu.c (open_pcsc_reader): Fix double free.
* gpg.c (main) [__APPLE__]: Default the PCSC driver to the OS X
location. Suggested by Patty A. Hardy.
2005-11-02 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (clean_sigs_from_uid): Include sigs from unavailable
keys in the sigs that are cleaned. Suggested by Dirk Traulsen and
many others.
2005-11-01 David Shaw <dshaw@jabberwocky.com>
* import.c (import_one): Do collapse_uids() before we do any
cleaning so keyserver mangled keys with doubled user IDs can be
properly cleaned - possibly sigs on the different user IDs cancel
each other out.
* import.c (parse_import_options), export.c
(parse_export_options): List "xxx-clean" before the longer options
so we don't end up with a partial match on the longer options.
* trustdb.c (clean_uids_from_key): Return proper number of cleaned
user IDs. Don't count user IDs as cleaned unless we actually
delete something.
2005-10-27 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_addrevoker), getkey.c (finish_lookup): Fix
problem with adding a cert-only designated revoker. Code was
looking for a key with sign ability, and not cert ability. Noted
by Timo Schulz.
2005-10-27 Werner Koch <wk@g10code.com>
* gpg.c [__CYGWIN__]: Set default driver to winscard.dll.
* apdu.c, apdu.h: Updated from gnupg 1.9. Changes are:
* apdu.c [__CYGWIN__]: Make cygwin environment similar to _WIN32.
Suggested by John P. Clizbe.
* apdu.h (SW_HOST_NO_KEYPAD): New.
* apdu.c (host_sw_string): Support new code.
(reader_table_s): New field CHECK_KEYPAD.
(new_reader_slot, open_ct_reader, open_pcsc_reader)
(open_ccid_reader, open_rapdu_reader): Initialize it.
(check_ccid_keypad): New.
(apdu_check_keypad): New.
(apdu_send_le): Factored all code out to ...
(send_le): .. new. Takes an additional arg; changed all callers
of the orginal function to use this one with a NULL for the new
arg.
(apdu_send_simple_kp): New.
(ct_send_apdu, pcsc_send_apdu, my_rapdu_send_apdu)
(send_apdu_ccid): New arg PININFO.
(send_apdu_ccid): Use the new arg.
2005-10-26 David Shaw <dshaw@jabberwocky.com>
* keygen.c (proc_parameter_file): Default key and subkey usage
flags to algo capabilities if parameter file doesn't specify them.
Noted by Timo Schulz.
2005-10-18 Werner Koch <wk@g10code.com>
* cardglue.c (pin_cb): Fixed prompt for repeated PIN. Return
G10ERR_CANCELED and not just -1.
(status_sc_op_failure): New. Use it where we issue that status.
(pin_cb): Append serial number to the need-pin status message.
(agent_scd_change_pin): Add arg SERIALNO. Changed all callers.
(agent_scd_writekey): Ditto.
(agent_scd_setattr): Ditto.
(agent_scd_genkey): Ditto.
(agent_scd_checkpin): Pass serialno to the pin_cb.
* keygen.c (parse_expire_string): Allow setting the expire
interval using a "seconds=<n>" syntax. This is useful for
debugging.
2005-10-17 Werner Koch <wk@g10code.com>
* export.c (do_export_stream): Factored some code out to ...
(skip_subkey_p): .. new.
(subkey_in_list_p, release_subkey_list): New.
(new_subkey_list_item): New.
(do_export_stream): Export exactly specified subkeys into one
keyblock.
2005-10-13 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu, menu_backsign): New "backsign" command
to add 0x19 backsigs to old keys that don't have them.
* misc.c (parse_options): Fix build warning.
* main.h, keygen.c (make_backsig): Make public.
2005-10-12 David Shaw <dshaw@jabberwocky.com>
* options.h, getkey.c (merge_selfsigs_subkey), gpg.c (main),
sig-check.c (signature_check2): Add --require-backsigs and
--no-require-backsigs. Currently defaults to
--no-require-backsigs.
2005-10-11 David Shaw <dshaw@jabberwocky.com>
* getkey.c (merge_selfsigs_subkey), sig-check.c
(signature_check2), keygen.c (make_backsig): Did some backsig
interop testing with the PGP folks. All is well, so I'm turning
generation of backsigs on for new keys. Checking for backsigs on
verification is still off.
2005-10-05 Werner Koch <wk@g10code.com>
* g10.c: Renamed to ..
* gpg.c: ..this.
* Makefile.am: Adjusted accordingly.
2005-09-22 Werner Koch <wk@g10code.com>
* sign.c (write_plaintext_packet): Don't print an empty file
warning if the file is actually too large.
* encode.c (encode_simple,encode_crypt): Ditto.
* progress.c (handle_progress): Adjusted for iobuf_get_filelength
change.
* photoid.c (generate_photo_id): Ditto.
2005-09-20 Werner Koch <wk@g10code.com>
* mainproc.c (proc_symkey_enc): Take care of a canceled passphrase
prompt.
2005-09-19 David Shaw <dshaw@jabberwocky.com>
* keylist.c (reorder_keyblock, do_reorder_keyblock): Reorder
attribute IDs as well as regular text IDs.
* plaintext.c (ask_for_detached_datafile): Use make_filename() on
filename so tilde expansion works.
2005-09-14 David Shaw <dshaw@jabberwocky.com>
* main.h, misc.c (parse_options): Add the ability to have help
strings in xxx-options commands.
* keyserver.c (keyserver_opts), import.c (parse_import_options),
export.c (parse_export_options), g10.c (parse_list_options, main):
Add help strings to xxx-options.
2005-09-10 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (show_names): Moved name display code out from
show_key_with_all_names.
(keyedit_menu): Call it here for pref and showpref so they can
show only the selected user ID. Suggested by Timo Schulz.
2005-09-07 Werner Koch <wk@g10code.com>
* cardglue.h (GPG_ERR_TOO_LARGE): New.
* apdu.c, apdu.h, iso7816.c, iso7816.h
* ccid-driver.c, ccid-driver.h: Updated from GnuPG 1.9 source.
Changes are:
* iso7816.c (iso7816_select_path): New.
* iso7816.c (iso7816_read_binary): Use Le=0 when reading all
data. Handle 6C00 error and take 6B00 as indication for EOF.
* apdu.h (SW_EXACT_LENGTH_P): New.
* apdu.c (new_reader_slot, reset_pcsc_reader, pcsc_get_status)
(open_pcsc_reader): Set new reader state IS_T0.
(apdu_send_le): When doing T=0 make sure not to send Lc and Le.
Problem reported by Carl Meijer.
(apdu_send_direct): Initialize RESULTLEN.
* misc.c (parse_options): Allow meta option "help" to list all
options and to exit the program.
2005-09-02 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (enum_sig_subpkt, parse_signature,
parse_attribute_subpkts): Make a number of warnings verbose items.
These fire on many slightly mangled keys in the field, so the
warning is becoming burdensome.
2005-09-01 David Shaw <dshaw@jabberwocky.com>
* photoid.h, photoid.c (generate_photo_id): Allow passing in a
suggested filename.
* keyedit.c (keyedit_menu, menu_adduid): Call it here so "addphoto
filename" works.
2005-08-31 David Shaw <dshaw@jabberwocky.com>
* photoid.c (generate_photo_id): Enable readline completion and
tilde expansion for the JPEG prompt.
2005-08-30 Werner Koch <wk@g10code.com>
* passphrase.c (agent_open): Print a warning and not an error in
case of a missing agent. Should fix Debian bug #325578.
2005-08-26 David Shaw <dshaw@jabberwocky.com>
* misc.c (openpgp_pk_algo_usage): Default to allowing CERT for
signing algorithms.
* keyedit.c (sign_uids): Don't request a signing key to make a
certification.
* keygen.c (do_add_key_flags): Force the certify flag on for all
primary keys, as the spec requires primary keys must be able to
certify (if nothing else, which key is going to issue the user ID
signature?)
(print_key_flags): Show certify flag.
(ask_key_flags, ask_algo): Don't allow setting the C flag for
subkeys.
* keyid.c (usagestr_from_pk), getkey.c (parse_key_usage):
Distinguish between a sign/certify key and a certify-only key.
* keyedit.c (ask_revoke_sig): Add a revsig --with-colons mode.
Suggested by Michael Schierl.
2005-08-21 David Shaw <dshaw@jabberwocky.com>
* Makefile.am: No need to link with curl any longer.
* main.h, misc.c (path_access): New. Same as access() but does a
PATH search like execlp.
* keyserver.c (curl_can_handle): Removed. Replaced by...
(curl_cant_handle): We are now relying on curl as the handler of
last resort. This is necessary because PGP LDAP and curl LDAP are
apples and oranges.
(keyserver_typemap): Only test for ldap and ldaps.
(keyserver_spawn): If a given handler is unusable (as determined
by path_access()) then try gpgkeys_curl.
* exec.h, exec.c (make_tempdir, expand_args, exec_write,
exec_read): Minor cleanup to use bitfield flags instead of a bunch
of integers.
2005-08-20 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Add aliases sign-with->local-user and
user->recipient to make switching from PGP command line to GPG
easier.
2005-08-19 David Shaw <dshaw@jabberwocky.com>
* options.skel: Remove the surfnet LDAP keyserver from the list of
samples since it is being shut down.
* getkey.c (classify_user_id): Disable the '.' and '+' search
modes since they aren't supported yet.
2005-08-05 David Shaw <dshaw@jabberwocky.com>
* g10.c (main), passphrase.c (set_passphrase_from_string): New
--passphrase command line option. Only useful in very special
circumstances.
2005-08-05 Werner Koch <wk@g10code.com>
* gpgv.c (keyserver_import_fprint): New stub.
* keygen.c (ask_user_id): Moved email checking code out to ..
* misc.c (is_valid_mailbox): .. new.
* mainproc.c (get_pka_address): Use it here.
* getkey.c (get_pubkey_byname): Add falback to auto-retrieve a key
via the PKA mechanism.
* options.h (KEYSERVER_AUTO_PKA_RETRIEVE): New.
* keyserver.c (keyserver_opts): Ditto.
* mainproc.c (check_sig_and_print): Use it here to retrieve keys
from a PKA DNS record.
* pkclist.c (build_pk_list): Add comments to this function;
re-indented it.
2005-08-04 David Shaw <dshaw@jabberwocky.com>
* keygen.c (proc_parameter_file): Sanity check items in keygen
batch file. Noted by Michael Schierl.
* pkclist.c (do_edit_ownertrust): Don't allow ownertrust level 0.
Noted by Michael Schierl.
* keygen.c (write_keyblock): Don't try and build deleted kbnodes
since we start our tree with one.
2005-08-04 Werner Koch <wk@g10code.com>
* export.c (do_export_stream): Skip on-card keys when only subkeys
are to be exported. It does not make sense to replace the on-card
key stub by a no-key stub.
* revoke.c (gen_revoke): Check for non-online keys.
* seckey-cert.c (is_secret_key_protected): Return -3 for
non-online key stubs. The old code assumes that a protection
algorithm is still set but in some cases this one is 0 and thus it
won't be possible to decide whether it is unprotected or
protected.
2005-07-28 Werner Koch <wk@g10code.com>
* Makefile.am (other_libs): Add SRVLIBS.
* parse-packet.c (can_handle_critical_notation): We know about
pka-address@gnupg.org.
* packet.h (PKT_signature): New fields PKA_INFO and PKA_TRIED.
(pka_info_t): New.
* free-packet.c (cp_pka_info): New.
(free_seckey_enc, copy_signature): Support new fields.
* mainproc.c (get_pka_address, pka_uri_from_sig): New.
(check_sig_and_print): Try to get the keyserver from the PKA
record.
* pkclist.c (check_signatures_trust): Adjust the trust based on
the PKA.
* gpgv.c (parse_keyserver_uri): New stub.
* keygen.c (has_invalid_email_chars): Moved to ..
* misc.c (has_invalid_email_chars): .. here and made global.
2005-07-27 Werner Koch <wk@g10code.com>
* export.c (do_export_stream): Make two strings translatable.
2005-07-26 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_typemap): Special-case LDAP since curl
will report that it can handle it, and we don't want it to.
2005-07-26 Werner Koch <wk@g10code.com>
* passphrase.c (agent_get_passphrase): Make sure to release the
saved codeset.
(agent_open): Add arg ORIG_CODESET and switch back to it in case
of error. Changed all callers.
2005-07-22 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (sign_uids): Don't prompt for setting signature expiry
to match key expiry unless --ask-cert-expire is set. Suggested by
Peter Palfrader.
2005-07-22 Werner Koch <wk@g10code.com>
* g10.c, options.h: New option --exit-on-status-write-error.
* status.c (write_status_text): Make use of this option.
2005-07-22 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main): Removed option --no-interactive-selection.
* keyedit.c (keyedit_menu): Use --interactive to enable the uid
walking when signing a key with no uids specified to sign.
* keylist.c (list_keyblock_print): Fix silly typo. Noted by Greg
Sabino Mullane.
2005-07-20 Werner Koch <wk@g10code.com>
* openfile.c (open_outfile): Disable FD caching for created files.
* encode.c (encode_simple, encode_crypt): Disable FD caching for
input files.
* verify.c (verify_one_file): Ditto.
* decrypt.c (decrypt_messages): Ditto. This is bug #479.
* misc.c (get_libexecdir) [W32]: Changed to return the value of
program used to create the process.
* keyserver.c (keyserver_spawn) [DISABLE_KEYSERVER_PATH]: Don't
change the exec-path at all.
2005-07-20 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (curl_can_handle): New. Do a runtime check against
libcurl to see if it can handle a particular protocol.
(keyserver_typemap): Call it here.
* Makefile.am: Pull in libcurl for curl_version_info() if used.
2005-07-19 Werner Koch <wk@g10code.com>
* g10.c, options.h: New option --limit-card-insert-tries.
* cardglue.c (open_card): Use it.
* export.c (parse_export_options): New option
export-reset-subkey-passwd.
(do_export_stream): Implement it.
* misc.c (get_libexecdir): New.
* keyserver.c (keyserver_spawn): Use it
2005-07-18 Werner Koch <wk@g10code.com>
* tdbio.c (open_db): Check for EROFS. Suggested by Bryce Nichols.
2005-07-08 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (clean_uids_from_key): Don't keep a valid selfsig
around when compacting a uid. There is no reason to make an
attacker's job easier - this way they only have a revocation which
is useless in bringing the uid back.
* keydb.h, kbnode.c (undelete_kbnode): Removed. No longer needed.
* import.c (chk_self_sigs): Allow a uid revocation to be enough to
allow importing a particular uid (no self sig needed). This
allows importing compacted uids.
2005-06-20 David Shaw <dshaw@jabberwocky.com>
* keygen.c (save_unprotected_key_to_card): Better fix for gcc4
warning.
2005-06-20 Werner Koch <wk@g10code.com>
* g10.c, options.h: New option --no-interactive-selection.
* keyedit.c (keyedit_menu): Use it.
2005-06-18 Werner Koch <wk@g10code.com>
* parse-packet.c (parse_signature): Use log_info for messages
about missing timestamp or keyid. In case we don't use that key
there won't be no further error and thus gpg does not need to
return with an error.
2005-06-13 David Shaw <dshaw@jabberwocky.com>
* keygen.c (save_unprotected_key_to_card): Fix gcc4 warning.
* options.h, import.c (parse_import_options, import_one): Add
import-clean-uids option to automatically compact unusable uids
when importing. Like import-clean-sigs, this may nodify the local
keyring.
* trustdb.c (clean_uids_from_key): Only allow selfsigs to be a
candidate for re-inclusion.
2005-06-12 David Shaw <dshaw@jabberwocky.com>
* options.h, import.c (parse_import_options,
clean_sigs_from_all_uids, import_one): Add import-clean-sigs
option to automatically clean a key when importing. Note that
when importing a key that is already on the local keyring, the
clean applies to the merged key - i.e. existing superceded or
invalid signatures are removed.
* getkey.c (merge_selfsigs_main, merge_selfsigs_subkey): Make sure
that even after keys may be merged together, we only have one
chosen selfsig.
2005-06-09 David Shaw <dshaw@jabberwocky.com>
* options.h, import.c (parse_import_options, delete_inv_parts):
import-unusable-sigs is now a noop.
* options.h, export.c (do_export_stream), keyedit.c (keyedit_menu,
menu_clean_subkeys_from_key), trustdb.h, trustdb.c
(clean_subkeys_from_key): Remove subkey cleaning function. It is
of very limited usefulness since it cannot be used on any subkey
that can sign, and can only affect multiple selfsigs on
encryption-only subkeys.
* keydb.h, kbnode.c (undelete_kbnode): New function to undelete a
kbnode.
* trustdb.c (clean_uids_from_key): Further tweak the algorithm so
that the last good selfsig is kept when the chosen selfsig is a
revocation.
2005-06-08 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (clean_uids_from_key), keyedit.c
(menu_clean_uids_from_key): Tweak algorithm to preserve the last
selfsig which helps prevent uid resurrections.
* getkey.c (fixup_uidnode, merge_selfsigs_main): Handle both
expired and revoked uids in fixup_uidnode(). No need to special
case in merge_selfsigs_main(). This also means that an expired
uid will have its selfsig tagged with chosen_selfsig.
2005-06-07 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), export.c (parse_export_options,
do_export_stream): Add export-options export-clean-sigs,
export-clean-uids, export-clean-subkeys, and export-clean which is
all of the above. Export-minimal is the same except it also
removes all non-selfsigs. export-unusable-sigs is now a noop.
2005-06-06 Werner Koch <wk@g10code.com>
* cardglue.c (open_card): Emit new CARDCTRL status 5 for no reader
available.
2005-06-02 Werner Koch <wk@g10code.com>
* app-openpgp.c (do_writekey): Typo fix.
* status.c, status.h: Removed STATUS_BAD_PASSPHRASE_PIN.
2005-06-01 David Shaw <dshaw@jabberwocky.com>
* signal.c [HAVE_DOSISH_SYSTEM]: Fix unused function warnings on
mingw32. Noted by Joe Vender.
* passphrase.c [_WIN32]: Remove unused variables.
2005-05-31 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_clean_uids_from_key,
menu_clean_subkeys_from_key), trustdb.c (clean_uids_from_key,
clean_subkeys_from_key): Fix mingw32 build warnings. Noted by Joe
Vender.
2005-05-31 Werner Koch <wk@g10code.com>
* keydb.h [!ENABLE_AGENT_SUPPORT]: Define dummy types.
* cardglue.c (assuan_strerror, assuan_transact): Dummy functions
if not build with agent support.
* armor.c (check_input): Don't bail out on invalid header lines
unless in struict rfc2440 mode. Suggested by Richard Patterson.
2005-05-30 Werner Koch <wk@g10code.com>
* tlv.c: Add hack to compile without gpg-error.h.
2005-05-30 David Shaw <dshaw@jabberwocky.com>
* trustdb.h, trustdb.c (clean_subkeys_from_key): New. Walk
through the subkeys on a key, and mark any that aren't usable for
deletion. Note that a signing subkey is never marked for deletion
since these keys are still useful after expiration or revocation.
* keyedit.c (menu_clean_subkeys_from_key): New function to call
clean_subkeys_from_key() on a key. Note that the strings here are
not marked for translation yet. The UI is still in flux, and
there is no point in annoying the translators twice.
(keyedit_menu): Call it here as part of the "clean" command.
2005-05-29 David Shaw <dshaw@jabberwocky.com>
* trustdb.h, trustdb.c (clean_uids_from_key): New. Walk through
the user IDs on a key, and mark any that aren't valid for
deletion.
* keyedit.c (menu_clean_uids_from_key): New function to call
clean_uids_from_key() on a key.
(keyedit_menu): Call it from here as part of the "clean" command.
2005-05-26 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Default {export|import}-unusable-sigs to off until
the "clean" UI can be finished.
2005-05-24 Werner Koch <wk@g10code.com>
* passphrase.c (ask_passphrase): Unescape the description string.
* cardglue.c (unescape_status_string): Removed. Changed all
caller to use ...
* misc.c (unescape_percent_string): New.
* g10.c (add_notation_data): Check number of at-signs.
2005-05-23 Werner Koch <wk@g10code.com>
* app-openpgp.c, app-common.h: Again updated from gnupg 1.9 CVS.
* cardglue.c (open_card): Check USE_AGENT.
(agent_scd_checkpin): Implemented Assuan part.
(agent_scd_change_pin): Ditto.
* g10.c (main): Option --debug-ccid-driver may now be given
several times increase the debug level.
* ccid-driver.c (parse_ccid_descriptor): Mark SCR335 FW version
5.14 as good.
(do_close_reader): Never do a reset. The caller should instead
make sure that the reader has been closed properly. The new retry
code in ccid_slot_status will make sure that the readersatrts up
fine even if the last process didn't closed the USB connection
properly.
(ccid_get_atr): For certain readers try switching to ISO mode.
Thanks to Ludovic Rousseau for this hint and the magic numbers.
(print_command_failed): New.
(bulk_in): Use it here. Add new arg NO_DEBUG.
(ccid_slot_status): Disabled debugging.
2005-05-21 Werner Koch <wk@g10code.com>
* cardglue.c (send_status_info): Make CTRL optional.
(agent_scd_writekey, inq_writekey_parms): New.
(agent_openpgp_storekey): Removed.
* cardglue.h: Add a few more error code mappings.
* keygen.c (copy_mpi): Removed.
(save_unprotected_key_to_card): Changed to use agent_scd_writekey.
* app-common.h, app-openpgp.c, tlv.c, tlv.h: Updated from newer
version in gnupg 1.9 CVS.
2005-05-20 Werner Koch <wk@g10code.com>
* ccid-driver.c (ccid_transceive): Arghhh. The seqno is another
bit in the R-block than in the I block, this was wrong at one
place. Fixes bug #419 and hopefully several others.
2005-05-19 Werner Koch <wk@g10code.com>
* app-common.h, app-openpgp.c, tlv.c, tlv.h: Updated from newer
version in gnupg 1.9 CVS.
2005-05-18 Werner Koch <wk@g10code.com>
* passphrase.c (agent_open): Made global and add arg TRY.
(agent_close): Made global.
* app-common.h (app_t): Add a field to store the Assuan context.
2005-05-13 David Shaw <dshaw@jabberwocky.com>
* build-packet.c (do_comment): Removed.
(build_packet): Ignore comment packets.
* export.c (do_export_stream): Don't export comment packets any
longer.
* options.h, g10.c (main): Remove --sk-comments and
--no-sk-comments options, and replace with no-op.
2005-05-11 David Shaw <dshaw@jabberwocky.com>
* keygen.c (write_selfsigs): Rename from write_selfsig. Write the
same selfsig into both the pk and sk, so that someone importing
their sk (which will get an autoconvert to the pk) won't end up
with two selfsigs.
(do_generate_keypair): Call it from here.
* parse-packet.c (can_handle_critical_notation): New. Check for
particular notation tags that we will accept when critical.
Currently, that's only preferred-email-encoding@pgp.com, since we
know how to handle it (pass it through to a mail program).
(can_handle_critical): Call it from here.
(parse_one_sig_subpkt): Sanity check that notations are
well-formed in that the internal lengths add up to the size of the
subpacket.
2005-05-07 Werner Koch <wk@g10code.com>
* ccid-driver.c (do_close_reader): Don't do a reset before close.
Some folks reported that it makes the SCR335 hang less often.
Look at the source on how to re-enable it.
2005-05-06 David Shaw <dshaw@jabberwocky.com>
* main.h, keygen.c (parse_expire_string, ask_expire_interval),
sign.c (sign_file, clearsign_file, sign_symencrypt_file), g10.c
(main), keyedit.c (sign_uids): Use seconds rather than days
internally to calculate expiration. We no longer need the
day-based code as we don't generate v3 keys.
* sign.c (sign_file, clearsign_file, sign_symencrypt_file): Use
the default sig expire value when signing in batchmode.
2005-05-05 David Shaw <dshaw@jabberwocky.com>
* Makefile.am, packet.h, main.h, comment.c: Remove comment.c. We
don't use any of these functions any longer.
* keygen.c (start_tree): New function to "prime" a KBNODE list.
(do_generate_keypair): Use it here rather than creating and
deleting a comment packet.
* keygen.c (gen_elg, gen_dsa): Do not put public factors in secret
key as a comment.
* options.h, encode.c (encode_simple, encode_crypt), keygen.c
(do_create): Remove disabled comment packet code.
* keygen.c (keygen_set_std_prefs): Add SHA256 and BZip2 to default
preferences.
* options.h, g10.c (main): Add new --default-sig-expire and
--default-cert-expire options. Suggested by Florian Weimer.
* main.h, keygen.c (parse_expire_string, ask_expire_interval): Use
defaults passed in, or "0" to control what default expiration is.
* keyedit.c (sign_uids), sign.c (sign_file, clearsign_file,
sign_symencrypt_file): Call them here, so that default expiration
is used when --ask-xxxxx-expire is off.
2005-05-03 Werner Koch <wk@g10code.com>
* passphrase.c (agent_get_passphrase): Add new arg CACHEID.
Changed all callers.
(ask_passphrase): Add new arg CACHEID and use it in agent mode.
Changed all callers.
(passphrase_clear_cache): New arg CACHEID. Changed all callers.
* cardglue.c (format_cacheid): New.
(pin_cb): Compute a cache ID.
(agent_scd_pksign, agent_scd_pkdecrypt): Use it.
(agent_clear_pin_cache): New.
* card-util.c (change_pin): Clear the PIN cache.
(check_pin_for_key_operation): Ditto.
2005-04-24 David Shaw <dshaw@jabberwocky.com>
* trustdb.h, trustdb.c (mark_usable_uid_certs): Add flags for the
no-pubkey and chosen revocation cases.
(clean_uid): New function to clean a user ID of unusable (as
defined by mark_usable_uid_certs) certs.
* keyedit.c (keyedit_menu, menu_clean_uids): Call it here for new
"clean" command that removes unusable sigs from a key.
* trustdb.h, keyedit.c (keyedit_menu, menu_select_uid_namehash):
Allow specifying user ID via the namehash from --with-colons
--fixed-list-mode --list-keys. Suggested by Peter Palfrader.
2005-04-21 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (sign_uids, keyedit_menu): When the user requests to
sign a key without specifying which user IDs to sign, and declines
to sign all user IDs, walk through the set of user IDs and prompt
for which to sign.
* mainproc.c (symkey_decrypt_seskey): There is no need to have an
extra check for a bad passphrase and/or unknown cipher algorithm
here. We'll fail quite happily later, and usually with a better
error message to boot.
2005-04-20 Werner Koch <wk@g10code.com>
* sign.c (sign_file, sign_symencrypt_file): Allow for hash
debugging.
2005-04-16 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_spawn): Free some memory.
* sign.c (hash_for): Comments.
2005-04-11 Werner Koch <wk@g10code.com>
* g10.c (main, add_notation_data, add_policy_url)
(add_keyserver_url): Use isascii() to protect the isfoo macros and
to replace direct tests. Possible problems noted by Christian
Biere.
* keyserver.c (parse_keyserver_uri): Ditto.
2005-04-07 Werner Koch <wk@g10code.com>
* g10.c (main): Declare --pipemode deprecated.
* misc.c (deprecated_command): New.
* ccid-driver.c (ccid_slot_status): Fixed debug messages.
* card-util.c (card_edit): Add command "verify". Enhanced admin
command to allow optional arguments "on", "off" and "verify".
(card_status): Print private DOs in colon mode.
* app-openpgp.c (do_check_pin): Add hack to allow verification of
CHV3.
2005-04-01 Werner Koch <wk@g10code.com>
* keygen.c (keygen_set_std_prefs): Explain the chosen order of
AES key sizes.
2005-04-01 David Shaw <dshaw@jabberwocky.com>
* mainproc.c (proc_plaintext): Properly handle SIG+LITERAL
(old-style PGP) signatures that use hashes other than SHA-1,
RIPEMD160, or MD5.
2005-03-31 David Shaw <dshaw@jabberwocky.com>
* exec.h, exec.c (set_exec_path): Remove some dead code and change
all callers. We no longer need to append to $PATH.
2005-03-31 Werner Koch <wk@g10code.com>
* passphrase.c (agent_open): Dropped support for W32 - is was
never actually used. Removed support for the old non-assuan
protocol; there has never been a matured implementation and
gpg-agent is now arround for quite some time. Rewritten to make
use of the Assuan code from ../util.
(gpga_protocol_codes): Removed.
(readn): Removed.
(agent_close): Simplified for use with Assuan.
(agent_get_passphrase, passphrase_clear_cache): Removed support
for old protocol. Use only with ENABLE_CARD_SUPPORT defined.
(agent_send_all_options): Take assuan context instead of a file
descriptor.
(agent_send_option): Likewise. Use assuan_transact.
* passphrase.c (writen, readaline): Removed.
* g10.c (main): Print a warning if --use-agent has been used but
it has not been build with support for it.
* keydb.c (keydb_add_resource): Clarify meaning of flags. Add new
flag 4. Use log_info for errors registering the default secret key.
* g10.c (main): Flag the default keyrings.
2005-03-30 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_spawn): Don't mess about with the $PATH.
Rather, call keyserver helpers with the full path. This fixes
some PATH-inspired DLL problems on W32. Noted by Carlo Luciano
Bianco.
2005-03-30 Werner Koch <wk@g10code.com>
* cardglue.c (pin_cb): Print a warning if the info string hack is
not there. This may happen due to typos in the translation.
2005-03-22 Werner Koch <wk@g10code.com>
* misc.c (w32_shgetfolderpath) [W32]: Changed declaration of
function ptr. Noted by Tim Costello.
* apdu.c [W32]: Changed declaration of dlopened function pointers.
2005-03-21 David Shaw <dshaw@jabberwocky.com>
* gpgv.c: Stubs for tty_enable_completion() &
tty_disable_completion().
* openfile.c (ask_outfile_name): Enable readline completion when
prompting for an output filename.
* plaintext.c (ask_for_detached_datafile): Enable readline
completion when prompting for a detached sig datafile.
2005-03-21 Werner Koch <wk@g10code.com>
* keyedit.c (command_generator, keyedit_completion): Changed
indentation.
* card-util.c (command_generator, card_edit_completion): Ditto.
2005-03-19 David Shaw <dshaw@jabberwocky.com>
* card-util.c (command_generator, card_edit_completion)
[GNUPG_MAJOR_VERSION==1 && HAVE_LIBREADLINE]: New functions to
enable command completion in the --card-edit menu.
(card_edit): Call them here.
2005-03-18 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (command_generator, keyedit_completion)
[HAVE_LIBREADLINE]: New functions to enable command completion in
the --edit-key menu.
(keyedit_menu): Call them here.
2005-03-17 David Shaw <dshaw@jabberwocky.com>
* getkey.c (get_seckey_byname2): If no explicit default key is
set, don't pick a disabled default. Noted by David Crick.
* Makefile.am: Calculate GNUPG_LIBEXECDIR directly. Do not
redefine $libexecdir.
* options.h, keyserver.c (parse_keyserver_options)
(keyserver_spawn): Don't treat 'verbose' and 'include-disabled' as
special. Just pass them through silently to the keyserver helper.
2005-03-16 Werner Koch <wk@g10code.com>
* ccid-driver.c (parse_ccid_descriptor): Make SCM workaround
reader type specific.
(scan_or_find_devices): Do not check the interface subclass in the
SPR532 kludge, as this depends on the firmware version.
(ccid_get_atr): Get the Slot status first. This solves the
problem with readers hanging on recent Linux 2.6.x.
(bulk_in): Add argument TIMEOUT and changed all callers to pass an
appropriate one. Change the standard timeout from 10 to 5 seconds.
(ccid_slot_status): Add a retry code with an initial short timeout.
(do_close_reader): Do an usb_reset before closing the reader.
2005-03-14 Werner Koch <wk@g10code.com>
* card-util.c (card_status): Use isotimestamp and not the
localized asctimestamp to match the timezone used in the key
information.
* cardglue.c (pin_cb): Disable debug output.
2005-03-11 Werner Koch <wk@g10code.com>
* keygen.c (gen_card_key_with_backup): Write status line with the
backup filename.
* status.h, status.h (STATUS_BACKUP_KEY_CREATED): New.
2005-03-10 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (parse_keyserver_options): Accept honor-http-proxy
as an alias for http-proxy.
* delkey.c (do_delete_key, delete_keys): Fix problem with --expert
preventing --delete-secret-and-public-keys from deleting secret
keys.
2005-03-10 Werner Koch <wk@g10code.com>
* keyedit.c (keyedit_menu) [W32]: Run the trustdb stale check
earlier.
2005-03-07 Werner Koch <wk@g10code.com>
* cardglue.c (agent_scd_pkdecrypt, agent_scd_pksign)
(agent_scd_genkey, agent_scd_setattr, agent_scd_change_pin)
(agent_scd_checkpin, agent_openpgp_storekey): Make sure to send a
SC_OP_FAILURE after card operations which might change data.
* card-util.c (change_pin): Send a SC_OP_SUCCESS after a PIN has
been changed.
(change_name): Removed a debug output.
* status.h, status.c: New codes BAD_PASSPHRASE_PIN, SC_OP_FAILURE
and SC_OP_SUCCESS.
2005-02-24 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu): Only print the key signing hint when
signing from a place where it is useful (i.e. --edit-key and not
--sign-key).
2005-02-16 Werner Koch <wk@g10code.com>
* card-util.c (fetch_url): Fetch the key from the default
keyserver if no URL is available.
2005-02-15 Werner Koch <wk@g10code.com>
* passphrase.c (agent_get_passphrase): Don't call free_public_key
if PK is NULL.
(passphrase_clear_cache): Ditto. Removed debug output.
(passphrase_to_dek): Ditto.
2005-02-13 Werner Koch <wk@g10code.com>
* keyedit.c (cmds): Limit code to 80 columns. Add command
BKUPTOCARD.
2005-02-09 David Shaw <dshaw@jabberwocky.com>
* encr-data.c (decrypt_data): Use it here to turn off the "quick
check" bytes for PK decryptions. This is in regards to the Mister
and Zuccherato attack on OpenPGP CFB mode.
* mainproc.c (proc_symkey_enc): Set a flag to indicate that a
particular session key came from a passphrase and not a PK.
2005-02-08 Werner Koch <wk@g10code.com>
* misc.c (w32_shgetfolderpath): New.
(default_homedir): Use it to avoid problems under Windows95.
2005-02-06 David Shaw <dshaw@jabberwocky.com>
* trustdb.h, trustdb.c (trustdb_check_or_update): New. If the
trustdb is dirty and --interactive is set, do an --update-trustdb.
If not interactive, do a --check_trustdb unless
--no-auto-check-trustdb is set.
* import.c (import_keys_internal): Moved from here.
* keyserver.c (keyserver_refresh): Call it here after all
refreshing has happened so that we don't rebuild after each
preferred keyserver set of imports, but do one big rebuild at the
end. This is Debian bug #293816, noted by Kurt Roeckx.
2005-02-04 David Shaw <dshaw@jabberwocky.com>
* getkey.c (merge_selfsigs_subkey): Merged away definition from
the backsigs code.
2005-01-31 David Shaw <dshaw@jabberwocky.com>
* keygen.c (do_generate_keypair): Write the auth key to the card
before the encryption key. This is a partial workaround for a PGP
bug (as of this writing, all versions including 8.1), that causes
it to try and encrypt to the most recent subkey regardless of
whether that subkey is actually an encryption type. In this case,
the auth key is an RSA key so it succeeds.
2005-01-27 David Shaw <dshaw@jabberwocky.com>
* keyid.c (keyid_from_sk, keyid_from_pk): Use 0xFFFFFFFFFFFFFFFF
instead of 0x0000000000000000 for the invalid key ID since
all-zeroes is reserved for the anonymous recipient.
* keyedit.c (change_passphrase), keygen.c (generate_subkeypair):
Fix a string ;)
2005-01-27 Werner Koch <wk@g10code.com>
* parse-packet.c (listfp): New.
(set_packet_list_mode): Intialize it to stdout or stderr depending
on a global option. Made all printing in list mode use LISTFP.
* keygen.c (generate_subkeypair): Detect primary key on-card and
ask for the passphrase. Return an error if the primary key is a
plain stub.
* keyedit.c (change_passphrase): Don't ever change any stub key.
Print a note if a key consists of only stub keys. Reported by
Dany Nativel. These are bugs #401 and #402.
2005-01-26 Werner Koch <wk@g10code.com>
* ccid-driver.c (parse_ccid_descriptor): Need the CSM workaround
also for newer firmware versions. Need to get a list of fixed
firmware versions and use that.
2005-01-26 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (parse_keyserver_uri): Allow RFC-2732 IPv6 [literal
address] syntax in keyserver URLs.
(keyserver_typemap): Map ftps if we are supporting it.
2005-01-25 Werner Koch <wk@g10code.com>
* keygen.c (do_generate_keypair): Don't continue after an error;
fixed at two places. Why at all didn't I used a goto to cleanup,
tsss?
* app-openpgp.c (get_cached_data): New arg GET_IMMEDIATE to bypass
the cache. Changed all callers.
(get_one_do): Bypass the cache if the value would have been read
directly for v1.1 cards. It makes things a bit slower but only for
1.0 cards and there are not that many cards out in the wild. This
is required to fix a caching bug when generating new keys; as a
side effect of the retrieval of the the C4 DO from the 6E DO the
chaced fingerprint will get updated to the old value and later
when signing the generated key the checking of the fingerprint
fails becuase it won't match the new one. Thanks to Moritz for
analyzing this problem.
(verify_chv3): Removed the CHV status reread logic because we
won't cache the C4 DO anymore.
2005-01-21 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (free_keyserver_spec): Fix small leak.
(keyserver_typemap): Map https if we are supporting it.
2005-01-20 Werner Koch <wk@g10code.com>
* cardglue.c (open_card): Issue new CARDCTRL(4) status.
* gpgv.c (tty_fprintf): New stub.
* card-util.c (card_status): Create a secret key stub on the fly
and print more information about a card key.
* import.c (pub_to_sec_keyblock, auto_create_card_key_stub): New.
* getkey.c (get_seckeyblock_byfprint): New.
* keylist.c (print_card_key_info): New.
* g10.c (i18n_init) [W32]: Pass registry key to gettext
initialization.
* gpgv.c (i18n_init) [W32]: Ditto.
2005-01-18 Werner Koch <wk@g10code.com>
* misc.c (default_homedir): New. Taken from gnupg 1.9.15.
* g10.c (main): Use it.
* gpgv.c (main): Ditto.
* keylist.c (public_key_list): Do a trustdb staleness check before
opening the keyring.
(secret_key_list): Ditto.
2005-01-10 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu): Move command strings outside the
function to get ready for the readline completion code.
* passphrase.c (readline, agent_send_option, agent_open,
agent_get_passphrase, passphrase_clear_cache): Rename readline()
to readaline() to keep readline library namespace clear.
2005-01-06 David Shaw <dshaw@jabberwocky.com>
* filter.h, armor.c (armor_filter): Use the eol string from the
armor filter context instead of hardcoding '\n' or '\r\n'. If no
eol string is provided, default to '\n' or '\r\n' as appropriate.
(is_armor_header): Trim tabs in armor header lines as well.
* keyserver.c (keyserver_spawn): Use it here to force '\n' line
endings since the keyserver output file gets a LF->CRLF expansion
on win32.
2005-01-05 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Typo.
* armor.c (is_armor_header): Allow CR and LF (not just actual
spaces) in an armor header line (-----BEGIN etc). This is needed
due to CRLF issues on win32. As before, --openpgp makes it
strict.
2005-01-03 David Shaw <dshaw@jabberwocky.com>
* Makefile.am: Use @LIBUSB@ instead of @LIBUSB_LIBS@
* import.c (delete_inv_parts): Comments on import-unusable-sigs.
2005-01-01 David Shaw <dshaw@jabberwocky.com>
* options.h, import.c (parse_import_options, delete_inv_parts):
Add import-unusable-sigs flag to enable importing unusable
(currently: expired) sigs.
* options.h, export.c (parse_export_options, do_export_stream):
Add export-unusable-sigs flag to enable exporting unusable
(currently: expired) sigs.
2004-12-29 David Shaw <dshaw@jabberwocky.com>
* packet.h, getkey.c (merge_selfsigs_main, sig_to_revoke_info),
keyid.c (revokestr_from_pk), keyedit.c (show_key_with_all_names):
Show who revoked a key (either the same key or a designated
revoker) and when.
2004-12-28 Werner Koch <wk@g10code.com>
* ccid-driver.c (find_endpoint): New.
(scan_or_find_devices): Add new args to return endpoint info and
interface number.
(ccid_open_reader, ccid_shutdown_reader): Take care of these new
args.
(bulk_in, bulk_out): Use the correct endpoints.
(ccid_transceive_apdu_level): New.
(ccid_transceive): Divert to above.
(parse_ccid_descriptor): Allow APDU level exchange mode.
(do_close_reader): Pass the interface number to usb_release_interface.
2004-12-24 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_typemap): Only map HTTP and FTP if
libcurl has specifically been selected to handle them.
2004-12-22 David Shaw <dshaw@jabberwocky.com>
* options.h, keyserver.c (parse_keyserver_uri): Properly parse
auth data from URLs and pass to keyserver helpers.
* keyserver.c (keyserver_typemap): New. Map certain keyserver
types to a common type (e.g. ldaps -> ldap). If we are building
with curl, map both http and ftp to curl.
* build-packet.c (build_sig_subpkt): Only allow one preferred
keyserver subpacket at a time.
2004-12-21 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_set_keyserver_url): Make sure we only operate on
the chosen selfsig so we don't accidentally promote an older
selfsig to chosen. Discovered by Simon Josefsson and 'Todd'.
* keygen.c (ask_expire_interval): Fix typo.
2004-12-20 David Shaw <dshaw@jabberwocky.com>
* keylist.c (list_keyblock_print): Secret key listings should
always show everything (expired UIDs, revoked subkeys, etc, etc).
* keyedit.c (keyedit_menu): Add additional help for the "sign"
flags.
2004-12-20 Werner Koch <wk@g10code.com>
* keygen.c (ask_expire_interval): For better translations chnage 2
strings.
* seckey-cert.c (do_check): Handle case when checksum was okay but
passphrase still wrong. Roman Pavlik found such a case.
2004-12-20 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu): Invisible alias "passwd" as
"password".
* passphrase.c: Don't check for __CYGWIN__, so it is treated as a
unix-like system.
* options.h, g10.c (main), textfilter.c (standard): Use new option
--rfc2440-text to determine whether to filter "<space>\t\r\n" or
just "\r\n" before canonicalizing text line endings. Default to
"<space>\t\r\n".
2004-12-19 David Shaw <dshaw@jabberwocky.com>
* keygen.c (keygen_get_std_prefs): Set reference count when
creating the temporary user ID.
* keyedit.c (keyedit_menu): Merge updpref and setpref. Keep
updpref as an invisible alias. Add invisible alias for revphoto.
Fix small memory leak when using "setpref" (not all of the uid was
freed).
(menu_revkey): Trigger a trust rebuild after revoking a key.
Don't allow revoking an already-revoked whole key.
(menu_revsubkey): Don't allow revoking an already-revoked subkey.
2004-12-18 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_revkey): Rename to menu_revsubkey.
(menu_revkey): New. Revoke a whole key.
(keyedit_menu): Call it here for when 'revkey' is used without any
subkeys selected. This is to be consistent with the other
functions which are "selected part if selected, whole key if not".
* signal.c: Use only HAVE_LIBREADLINE to detect readline
availability.
* Makefile.am: Link with readline where necessary.
2004-12-17 Werner Koch <wk@g10code.com>
* passphrase.c (agent_get_passphrase): Define NREAD locally as
size_t or int.
* keylist.c (list_keyblock_print): Make field width an int.
* keyedit.c (show_key_with_all_names): Ditto.
2004-12-16 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Add --require-secmem/--no-require-secmem to cause
gpg to exit if it cannot lock memory. Also remove --nrsign-key
and --nrlsign-key since this can better be done via --edit-key.
2004-12-15 David Shaw <dshaw@jabberwocky.com>
* apdu.c (apdu_send_le, apdu_send_direct), keylist.c
(status_one_subpacket, print_one_subpacket): Fix some compiler
warnings.
* g10.c (main): Fix --compression-algo to take a string argument
like --compress-algo.
* trustdb.c (uid_trust_string_fixed): For safety, check for a pk.
2004-12-14 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu): Re-remove the N_() markers.
* trustdb.c (uid_trust_string_fixed): Show uids as revoked if the
key is revoked.
* keyedit.c (show_key_with_all_names): Don't show validity for
secret key UIDs.
* keyedit.c (parse_sign_type): New. Figure out the flags (local,
nonrevoke, trust) for a signature.
(keyedit_menu): Call it here so we can mix and match flags, and
don't need "nrltsign", "ltsign", "tnrsign", etc, etc, etc.
2004-12-14 Werner Koch <wk@g10code.com>
* passphrase.c (agent_get_passphrase): Removed debug output
* keyserver.c (keyserver_work, keyserver_spawn): Map ldaps to ldap.
* keyedit.c (keyedit_menu): Removed the N_() markers from the
command names.
* card-util.c (card_edit): Ditto.
2004-12-13 Werner Koch <wk@g10code.com>
* passphrase.c (read_passphrase_from_fd): Fixed memory leak.
Noted by Andrei Darashenka.
2004-12-11 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (parse_preferred_keyserver): Force preferred
keyserver subpackets to have a URI scheme specified.
2004-12-10 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), textfilter.c (standard): Use --rfc2440
or --openpgp directly to determine the end of line hashing rule.
* trustdb.c (uid_trust_string_fixed): Show uids as expired if the
key is expired.
2004-12-10 Werner Koch <wk@g10code.com>
* app-openpgp.c (send_fprtime_if_not_null): New.
(do_getattr): Add KEY_TIME.
(do_learn_status): Print KEY_TIME.
* cardglue.c (learn_status_cb): Parse KEY-TIME.
* card-util.c (card_status): Print creation time if available.
2004-12-09 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), textfilter.c (len_without_trailing_ws):
Removed (not used).
(standard): 2440 says that textmode hashes should canonicalize
line endings to CRLF and remove spaces and tabs. 2440bis-12 says
to just canonicalize to CRLF. So, we default to the 2440bis-12
behavior, but revert to the strict 2440 behavior if the user
specifies --rfc2440. In practical terms this makes no difference
to any signatures in the real world except for a textmode detached
signature.
2004-12-09 Werner Koch <wk@g10code.com>
* passphrase.c (agent_get_passphrase): New args CUSTOM_PROMPT and
CUSTOM_DESCRIPTION. Changed all callers.
* app-openpgp.c (do_getattr, do_learn_status, do_setattr): Support
the new private DOs.
(do_change_pin): Add a "N" prefix to the strings so that the
callback can act accordingly for a new PIN. Unfortunately this
breaks existing translations but I see no wother way to overvome
this.
* cardglue.c (learn_status_cb): Ditto.
(agent_release_card_info): Ditto.
(struct pin_cb_info_s): Removed and changed all users.
(pin_cb): Reworked.
* card-util.c (card_status): Print them
(card_edit): New command PRIVATEDO.
(change_private_do): New.
2004-12-09 David Shaw <dshaw@jabberwocky.com>
* keygen.c (ask_algo): Add a choose-your-own-capabilities option
for DSA.
2004-12-07 David Shaw <dshaw@jabberwocky.com>
* keygen.c (ask_keysize): Change strings to always use %u instead
of hardcoding key sizes. Bump default to 2048. Bump minimum down
to 512, where possible, but require --expert to get there. DSA is
always 1024 unless --expert is given.
2004-11-29 David Shaw <dshaw@jabberwocky.com>
* getkey.c (parse_key_usage): New function to parse out key usage
flags. Set PUBKEY_USAGE_UNKNOWN to handle flags that we don't
understand.
(fixup_uidnode, merge_selfsigs_main, merge_selfsigs_subkey): Call
it from here to remove duplicate code.
2004-11-26 David Shaw <dshaw@jabberwocky.com>
* export.c (do_export_stream): Allow export-minimal to work with
secret keys, even though a non-selfsig secret key signature is
rare.
* options.h, export.c (parse_export_options, do_export_stream),
import.c (parse_import_options, import_keys_internal): Make the
import-options and export-options distinct since they can be mixed
together as part of keyserver-options.
2004-11-24 David Shaw <dshaw@jabberwocky.com>
* options.h, export.c (parse_export_options, do_export_stream):
Add "export-minimal" option to disregard any sigs except selfsigs.
* trustdb.c (uid_trust_string_fixed): Use a string that can be
atoi-ed, but also has a comment for the translator.
* trustdb.h, trustdb.c (uid_trust_string_fixed): New. Return a
fixed-size translatable string similar to trust_value_to_string.
This allows for easier lining up of displays.
* keyedit.c (show_key_with_all_names), keylist.c
(list_keyblock_print): Use it here to print validity strings.
* gpgv.c: Stub.
2004-11-18 Werner Koch <wk@g10code.com>
* g10.c (S_IRGRP) [HAVE_DOSISH_SYSTEM]: Define to 0.
2004-11-17 Werner Koch <wk@g10code.com>
* g10.c (open_info_file): New.
(main): Unconditionally implement --status-file, --logger-file,
--attribute-file, --passphrase-file, --command-file. This is not
generally useful but easy to support and might make scripting
under Windows easier.
2004-11-11 Werner Koch <wk@g10code.com>
* passphrase.c (readn): Fixed test against EINTR.
2004-11-05 Werner Koch <wk@g10code.com>
* app-openpgp.c: Made more strings translatable.
(verify_chv3, do_change_pin): Add a special prefix to the prompt
of the Admin PIN prompts.
* passphrase.c (ask_passphrase): Add arg TRYAGAIN_TEXT. Changed
call callers.
* cardglue.c (pin_cb): Make use of the OPAQUE arg to pass
arguments to the PIN callback. Use this to implement a way to
check for correct PIN repetition. Changed all callers to pass an
opaque argument. Improved detection of Admin PIN prompts.
2004-11-04 David Shaw <dshaw@jabberwocky.com>
* plaintext.c (handle_plaintext): Don't try and create a
zero-length filename when using --use-embedded-filename with input
that has no filename (clearsigned or message generated from a
pipe).
* encode.c (encode_simple, encode_crypt), progress.c
(handle_progress), sign.c (write_plaintext_packet): Fix a few
inconsistent calls (NULL filename means a pipe here, so don't
bother to check it twice).
2004-11-03 David Shaw <dshaw@jabberwocky.com>
* misc.c (print_digest_algo_note): The latest 2440bis drafts
deprecates MD5, so give a warning.
(print_pubkey_algo_note, print_cipher_algo_note,
print_digest_algo_note): Give the algorithm name in the
experimental algo warning.
2004-11-03 Timo Schulz <twoaday@g10code.com>
* passphrase.c (readn, writen): Use w32_strerror instead
of just showing the error number.
* misc.c [_WIN32]: Fix warning about missing prototypes.
2004-10-28 David Shaw <dshaw@jabberwocky.com>
* skclist.c (build_sk_list): Don't need to warn about
PGP-generated Elgamal signing keys since we no longer support any
Elgamal signing keys.
* sign.c (sign_file, clearsign_file): Use "writing to" instead of
"writing to file" to match other strings.
* pkclist.c (check_signatures_trust): Fix typo. Noted by Moray
Allan. This is Debian bug #278708.
* passphrase.c (ask_passphrase, passphrase_to_dek): "password" ->
"passphrase".
* keyedit.c (show_key_with_all_names): Show designated revoker as
part of translatable string.
2004-10-28 Werner Koch <wk@g10code.com>
* Makefile.am (other_libs): New. Also include LIBICONV. Noted by
Tim Mooney.
2004-10-28 Werner Koch <wk@g10code.com>
* apdu.c (open_pcsc_reader): Removed bad free in error handler.
2004-10-27 David Shaw <dshaw@jabberwocky.com>
* card-util.c, delkey.c, keygen.c, plaintext.c, keyedit.c,
passphrase.c, revoke.c: Collapse the two different "can't do that
in batch mode" strings into one.
* keylist.c (status_one_subpacket): New. Send the subpacket data
to the --status interface.
* card-util.c (card_edit): Show when admin is enabled or not.
* status.h, status.c: New STATUS_SIG_SUBPACKET type.
* build-packet.c (build_sig_subpkt): Multiple keyserver URLs are
allowed.
* keyring.c: Make some strings translatable.
* exec.c, sign.c: Change "can't open file" to "can't open" and
"can't create file" to "can't create" to be consistent with other
strings so we don't have to translate both.
* delkey.c, export.c, keyedit.c, pkclist.c, revoke.c, skclist.c:
Fix a few missed possible \"username\" quotes.
2004-10-26 Werner Koch <wk@g10code.com>
* app-openpgp.c (verify_chv3): The minimum length for CHV3 is
8. Changed string to match the other ones.
* passphrase.c (agent_send_all_options): Try to deduce the ttyname
from stdin.
2004-10-22 Werner Koch <wk@g10code.com>
* card-util.c (fetch_url): Disable for gnupg 1.9
(card_generate_subkey): Ditto.
(card_store_subkey): Ditto.
2004-10-21 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), mainproc.c (check_sig_and_print):
Rename verify-option show-validity to show-uid-validity to match
the similar list-option.
* app-openpgp.c (verify_chv3): Fix typo.
2004-10-21 Werner Koch <wk@g10code.com>
* app-common.h (app_openpgp_storekey): Add prototype.
* app-openpgp.c (do_sign): Replace asprintf by direct allocation.
This avoids problems with missing vasprintf implementations.
* card-util.c (generate_card_keys): Add a #warning for gnupg 1.9
and use the same string there.
2004-10-20 David Shaw <dshaw@jabberwocky.com>
* g10.c (parse_list_options): Fix non-constant initializer so we
can build with C89.
2004-10-17 David Shaw <dshaw@jabberwocky.com>
* keylist.c (print_one_subpacket): The flags field should be hex.
2004-10-17 Werner Koch <wk@g10code.com>
* passphrase.c (agent_get_passphrase): Cast UIDLEN to int. Noted
by Christian Cornelssen.
2004-10-16 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (parse_one_sig_subpkt, enum_sig_subpkt): Don't
BUG() on unknown subpackets. Rather, just return them silently.
2004-10-15 Werner Koch <wk@g10code.com>
* status.h (STATUS_NEED_PASSPHRASE_PIN): New.
* status.c (get_status_string): Added.
* passphrase.c (ask_passphrase): Moved status printing to ..
* cardglue.c (pin_cb): .. here and issue new status message.
* keyedit.c (sign_uids): Don't include the leading LF in the
translatable string but print them separately.
* apdu.c (apdu_open_remote_reader) [_WIN32]: We don't have ENOSYS.
* app-openpgp.c (parse_login_data): New.
(app_select_openpgp): Call it.
(do_setattr): Reparse it after change.
* pkclist.c (do_edit_ownertrust): Add a note to translators.
* keygen.c (ask_user_id): Ditto.
* helptext.c: Typo fix.
2004-10-14 David Shaw <dshaw@jabberwocky.com>
* keylist.c (list_keyblock_print): Show the fingerprint after the
key, not after the first user ID.
* keyedit.c (show_key_with_all_names): Don't show validity if
we're just printing user IDs for signing.
* armor.c (fake_packet): Properly handle the case where the line
is dash-space (i.e. a blank line that was quoted). Give a warning
for bad dash escaping.
2004-10-14 Werner Koch <wk@g10code.com>
* export.c (do_export_stream) [ENABLE_SELINUX_HACKS]: Don't allow
secret key export.
* import.c (import_secret_one) [ENABLE_SELINUX_HACKS]: Likewise
* misc.c (is_secured_filename): New.
* keydb.c (maybe_create_keyring)
* tdbio.c (tdbio_set_dbname)
* plaintext.c (handle_plaintext)
* openfile.c (copy_options_file, open_outfile)
* exec.c (exec_write)
* keygen.c (do_generate_keypair, gen_card_key_with_backup)
* sign.c (sign_file, clearsign_file)
* keyring.c (create_tmp_file, do_copy): Check for secured files
before creating them.
* keygen.c (print_status_key_created, read_parameter_file):
s/unsigned char/byte/ due to a strange typedef for RISC OS. Noted
by Stefan.
2004-10-13 David Shaw <dshaw@jabberwocky.com>
* armor.c (fake_packet): Allow arbitrary dash-escaped lines as per
2440bis-10. This is bug #158.
* keyserver.c (keyserver_work): Handle keyserver timeouts.
* pkclist.c (do_edit_ownertrust): Different prompt when we're
using direct trust since the meaning is different.
* keyedit.c (trustsig_prompt): Change the strings to match the
ones in pkclist.c:do_edit_ownertrust to make translation easier.
* trustdb.c (trust_model_string, get_validity): Add direct trust
model which applies to the key as a whole and not per-uid.
* options.h, g10.c (parse_trust_model): New.
(main): Call it from here to do string-to-trust-model.
2004-10-13 Werner Koch <wk@g10code.com>
* tdbdump.c (import_ownertrust): Removed all log_error_f and
reworded the messages.
* dermor.c: Include i18n.h. Made 2 strings translatable.
* misc.c (register_secured_file, is_secured_file)
(unregister_secured_file): New.
* keyring.c (do_copy, rename_tmp_file): Implement the SELinux hacks.
(keyring_register_filename): Ditto.
* tdbio.c (open_db): Ditto.
* openfile.c (copy_options_file, open_sigfile): Ditto.
* verify.c (verify_signatures, verify_one_file): Ditto.
* photoid.c (generate_photo_id): Ditto.
* keygen.c (read_parameter_file): Ditto.
* import.c (import_keys_internal): Ditto.
* decrypt.c (decrypt_message, decrypt_messages): Ditto.
* dearmor.c (dearmor_file, enarmor_file): Ditto.
* g10.c (main, print_mds): Ditto.
* exec.c (exec_write, exec_read): Ditto.
* card-util.c (change_login): Ditto.
* encode.c (encode_simple, encode_crypt): Ditto.
* openfile.c (overwrite_filep, make_outfile_name, open_outfile)
(open_sigfile): Use iobuf_is_pipe_filename to check for pipes so
that special filesnames are taken into account. This is bug 327.
* tdbdump.c (import_ownertrust): Ditto.
* sign.c (write_plaintext_packet): Ditto.
(sign_file, clearsign_file, sign_symencrypt_file):
* progress.c (handle_progress): Ditto.
* plaintext.c (handle_plaintext): Ditto.
(ask_for_detached_datafile, hash_datafiles):
* encode.c (encode_simple, encode_crypt): Ditto.
2004-10-12 Werner Koch <wk@g10code.com>
* keygen.c (read_parameter_file): Changed to use iobuf based file
reading to allow the special file name feature to work.
* keygen.c (read_parameter_file): New keyword "Handle". This is
bug 287.
(print_status_key_not_created): New.
(print_status_key_created): Add new arg HANDLE.
(do_generate_keypair): Print not created status.
* status.c, tatus.h (STATUS_KEY_NOT_CREATED): New.
2004-10-11 David Shaw <dshaw@jabberwocky.com>
* pkclist.c (do_edit_ownertrust): Use the same translated string
for showing the user ID as mainproc.c:print_pkenc_list.
* mainproc.c (print_pkenc_list): Allow translating the quotes
around the user ID.
* card-util.c, g10.c, photoid.c, trustdb.c: The last of the \"%s\"
-> `%s' quoting for things that aren't user IDs.
* keyserver.c (keyserver_spawn): If there is no keyserver host,
print the whole URI since it is self-contained.
2004-10-11 Werner Koch <wk@g10code.com>
* keyserver.c (keyserver_spawn): Print an empty string in log_info
if the host is not set (e.g. finger).
2004-10-10 David Shaw <dshaw@jabberwocky.com>
* card-util.c, keyedit.c, openfile.c, pkclist.c, delkey.c,
keygen.c, photoid.c, revoke.c: Some yes-or-no prompts end in
"(y/n)". Some don't. Consistently use y/n everywhere.
* keygen.c (ask_key_flags): New.
(ask_algo): Call it here in --expert mode so we don't need to
specify each possible variation of RSA capabilities.
* keygen.c (do_add_key_flags): The spec says that all primary keys
MUST be able to certify. Force the certify flag on for primaries
(and off for subkeys).
* keygen.c (generate_keypair): Fix generating keys with the auth
flag.
2004-10-08 David Shaw <dshaw@jabberwocky.com>
* encr-data.c (decrypt_data): Give a warning with a weak key, but
still allow to decrypt the message.
2004-10-07 David Shaw <dshaw@jabberwocky.com>
* pkclist.c (build_pk_list): Keystrify.
* mainproc.c (check_sig_and_print), pkclist.c
(do_edit_ownertrust): Improve translatability of user ID prompts.
2004-10-06 David Shaw <dshaw@jabberwocky.com>
* helptext.c, pkclist.c (do_we_trust): It is not possible to get
here with a revoked or expired key, so BUG() that case. Remove
question about overriding revoked/expired. Also
--keyid-format-ify.
(do_we_trust_pre): Use print_pubkey_info() instead of printing the
info ourselves.
* passphrase.c (passphrase_to_dek): Improve translatability of
user ID prompts.
* keylist.c (print_pubkey_info): Use the user ID the pk was
selected by, if any.
* keyedit.c (sign_uids, ask_revoke_sig): Improve translatability
of user ID prompts.
(ask_revoke_sig, menu_revsig): Try and use common strings for
these two functions so they don't need to be translated twice.
* keyedit.c, keylist.c, keyserver.c, mainproc.c: The
revoked/expired/expires string change of 2004-09-29 was too
simple. Use two styles for each tag.
2004-10-06 Werner Koch <wk@g10code.com>
* ccid-driver.c (ccid_open_reader): Store the vendor ID.
(ccid_transceive_secure): New.
(parse_ccid_descriptor): Workaround for an SCM reader problem.
(send_escape_cmd): New.
2004-10-05 David Shaw <dshaw@jabberwocky.com>
* passphrase.c (agent_get_passphrase): Use keystrs for agent
strings, and fix sprintf warnings.
* keyserver.c (keyserver_spawn): Fix BUG() with certain sets of
mixed regular and preferred keyserver refreshes. Noted by
Sebastian Wiesinger.
* keyedit.c (show_key_with_all_names): Show uid validity in menu.
2004-10-03 Timo Schulz <twoaday@g10code.de>
* apdu.c (apdu_open_remote_reader) [_WIN32]: Do not set ENOSYS.
2004-10-03 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (print_and_check_one_sig_colon): Fix bad keyids in
colon delsig output. Noted by Peter Palfrader.
(show_prefs): Do not reference missing selfsig. Noted by Alex
Moroz.
2004-10-01 Werner Koch <wk@g10code.com>
* gpgv.c (i18n_init): Always use LC_ALL.
2004-09-30 Werner Koch <wk@g10code.com>
* app-openpgp.c (verify_chv3) [GNUPG_MAJOR_VERSION!=1]: Typo fix.
2004-09-30 David Shaw <dshaw@jabberwocky.com>
* gpgv.c, keydb.c (keydb_add_resource): Factored keyring creation
out to ..
(maybe_create_keyring): .. new. Make sure that we do the checks
in a locked state. Problem reported by Stefan Haller. Try to
create the home directory before acquiring a lock for the keyring.
From Werner on stable branch.
* g10.c (main): Blow up if we didn't lose setuid. From Werner on
stable branch.
2004-09-29 David Shaw <dshaw@jabberwocky.com>
* keyedit.c, keylist.c, keyserver.c, mainproc.c: Reduce the many
variations of "revoked" ("revoked", "[revoked]", " [revoked]",
"[revoked] ") "and" expired down to two to simplify translation.
2004-09-28 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (print_and_check_one_sig): Account for the extra space
that show-sig-expire takes up so we do not wrap lines.
(show_key_with_all_names): No need to show subkey revocations as a
seperate line since we now show revocation date in the main subkey
line.
* signal.c (got_fatal_signal): HAVE_DECL_SYS_SIGLIST is defined,
but zero if not found. Noted by John Clizbe.
* keyserver.c (parse_keyrec): Fix problem with non-expiring keys
appearing expired in --search-keys results.
2004-09-27 Werner Koch <wk@g10code.com>
* card-util.c (card_edit): Take admin only status from the table.
* app-openpgp.c: Made all strings translatable.
(verify_chv3) [GNUPG_MAJOR_VERSION]: Make opt.allow_admin
available for use in gnupg 2.
(verify_chv3): Reimplemented countdown showing to use only
functions from this module. Flush the CVH status cache on a
successful read.
(get_one_do): Hack to bypass the cache for cards versions > 1.0.
(store_fpr): Store the creation date for card version > 1.0.
2004-09-25 David Shaw <dshaw@jabberwocky.com>
* main.h, g10.c (main), card-util.c (change_pin): If "admin" has
not been issued, skip right to the CHV1/CHV2 PIN change. No need
to show the unblock or admin PIN change option.
(card_edit): Add "admin" command to add admin commands to the
menu. Do not allow admin commands until "admin" is given.
* app-openpgp.c (verify_chv3): Show a countdown of how many wrong
admin PINs can be entered before the card is locked.
* options.h, g10.c (main), app-openpgp.c (verify_chv3): Remove
--allow-admin.
2004-09-24 David Shaw <dshaw@jabberwocky.com>
* main.h: Create S2K_DIGEST_ALGO macro so we do not need to always
set opt.s2k_digest_algo. This helps fix a problem with PGP 2.x
encrypted symmetric messages. Change all callers (encode.c,
g10.c, keyedit.c, keygen.c, passphrase.c, sign.c).
* armor.c, cardglue.c, getkey.c, import.c, keygen.c: Be consistent
in some more quoted strings. Always use 'user ID', not 'user id',
"quotes" for user IDs, etc.
* keyedit.c (keyedit_menu), gpgv.c (agent_scd_getattr (stub)),
keygen.c (copy_mpi, generate_raw_key): Fix a compile problem and a
few warnings when building without card support.
2004-09-23 Werner Koch <wk@g10code.com>
* card_util.c (generate_card_keys): ask whether backup should be
created.
(card_store_subkey): Factored some code out to ..
* keygen.c (save_unprotected_key_to_card): .. new function.
(gen_card_key_with_backup): New.
(generate_raw_key): New.
(generate_keypair): New arg BACKUP_ENCRYPTION_DIR. Changed all
callers.
(do_generate_keypair): Divert to gen_card_key_with_backup when
desired.
* apdu.c (open_pcsc_reader): Do not print empty reader string.
* keygen.c (ask_algo): Allow creation of AUTH keys.
* keyid.c (usagestr_from_pk): New.
* app-openpgp.c (app_openpgp_storekey): Call flush_cache.
(get_cached_data): Move local data initialization to ..
(app_select_openpgp): .. here. Read some flags for later use.
(do_getattr): New read-only attribute EXTCAP.
* keyedit.c (keyedit_menu): New command "keytocard"
(keyedit_menu): Bad hack for the not_with_sk element.
(show_key_with_all_names): Print the usage.
(find_pk_from_sknode): New.
* card-util.c (card_store_subkey): New.
(copy_mpi): New.
* cardglue.c (agent_openpgp_storekey): New.
2004-09-22 Werner Koch <wk@g10code.com>
* card-util.c (card_generate_subkey, generate_card_keys): Factored
common code out to ...
(get_info_for_key_operation, check_pin_for_key_operation)
(restore_forced_chv1, replace_existing_key_p)
(show_card_key_info): ... new functions.
2004-09-21 David Shaw <dshaw@jabberwocky.com>
* mainproc.c (check_sig_and_print), keyedit.c (show_prefs,
menu_set_keyserver_url): Make sure that keyserver URLs with
control characters inside are printed properly. In fact, handle
them as UTF8.
* keyedit.c (keyedit_menu): Don't show "addcardkey" in the menu if
we do not have card support.
* keydb.h, keyserver.c (print_keyrec, keyserver_spawn): fpr is an
array of unsigned bytes.
2004-09-20 Werner Koch <wk@g10code.com>
* g10.c: Make -K an alias for --list-secret-keys.
* keylist.c (print_card_serialno): New. Taken from gnupg 1.9.11.
(list_keyblock_print): Make use of it.
* keyedit.c (show_key_with_all_names): Print the card S/N.
* keyedit.c (keyedit_menu): New command ADDCARDKEY.
* card-util.c (card_generate_subkey): New.
* keygen.c (generate_card_subkeypair): New.
(gen_card_key): New arg IS_PRIMARY; changed all callers.
* cardglue.c (open_card): Use shutdown code if possible.
(check_card_serialno): Ditto.
* ccid-driver.c (do_close_reader): Factored some code out from ...
(ccid_close_reader): ..here.
(ccid_shutdown_reader): New.
* apdu.c (apdu_shutdown_reader): New.
(shutdown_ccid_reader): New.
2004-09-17 Werner Koch <wk@g10code.com>
* g10.c (list_config): New config option ccid-reader-id.
(gpgconf_list): Add "reader-port".
* apdu.c (open_ccid_reader): New arg PORTSTR. Pass it to
ccid_open_reader.
(apdu_open_reader): Pass portstr to open_ccid_reader.
(apdu_open_reader): No fallback if a full CCID reader id has been
given.
* ccid-driver.c (ccid_get_reader_list): New.
(ccid_open_reader): Changed API to take a string for the reader.
Removed al the cruft for the libusb development vesion which seems
not to be maintained anymore and there are no packages anyway.
The stable library works just fine.
(struct ccid_reader_id_s): Deleted and replaced everywhere by a
simple string.
(usb_get_string_simple): Removed.
(bulk_in): Do valgrind hack here and not just everywhere.
2004-09-16 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (show_key_with_all_names, show_prefs): Show preferred
keyserver(s) in "showpref" output.
* keygen.c (keygen_add_keyserver_url), keyedit.c
(menu_set_keyserver_url): Allow setting a keyserver URL of "none"
to remove an existing keyserver URL.
* keyedit.c (menu_set_keyserver_url): Confirm replacement of a
keyserver URL before overwriting the old one.
2004-09-15 David Shaw <dshaw@jabberwocky.com>
* gpgv.c (agent_scd_getattr): Stub.
* misc.c (get_signature_count): New. Get the signature count from
a smartcard.
(pct_expando): Call it here so the %c expando becomes the number
of signatures issued. This allows for notations or the like with
an automatic signature count.
* ccid-driver.c (usb_get_string_simple): Replacement function to
work with older libusb.
2004-09-15 Werner Koch <wk@g10code.com>
* g10.c [HAVE_LIBUSB]: New option --debug-ccid-driver.
* ccid-driver.c (read_device_info): Removed.
(make_reader_id, scan_or_find_devices): New.
(ccid_open_reader): Simplified by make use of the new functions.
(ccid_set_debug_level): New. Changed the macros to make use of
it. It has turned out that it is often useful to enable debugging
at runtime so I added this option.
2004-09-13 David Shaw <dshaw@jabberwocky.com>
* getkey.c (premerge_public_with_secret): Fix subkey<->binding sig
mismatch when some secret subkeys are missing. Discovered by
Michael Roth.
* main.h, keylist.c (print_subpackets_colon): Make a public
function.
* keyedit.c (print_and_check_one_sig_colon): New. Print a
with-colons version of the sig record.
(menu_delsig): Call it here for a with-colons delsig.
2004-09-12 David Shaw <dshaw@jabberwocky.com>
* options.h, keylist.c (print_one_subpacket,
print_subpackets_colon): Print a spk record for each request
subpacket.
(list_keyblock_colon): Call them here.
* g10.c (parse_subpacket_list, parse_list_options): New. Make the
list of subpackets we are going to print.
(main): Call them here.
2004-09-11 David Shaw <dshaw@jabberwocky.com>
* card-util.c (fetch_url, card_edit): Use the pubkey URL stored on
the card to fetch an updated copy. Works with either straight
URLs or HKP or LDAP keyservers.
* keyserver-internal.h, keyserver.c (keyserver_import_fprint),
import.c (revocation_present): Use a keyserver_spec so the caller
can pass in whatever keyserver they like.
2004-09-10 David Shaw <dshaw@jabberwocky.com>
* app-openpgp.c (get_cached_data): Avoid mallocing zero since it
breaks us when using --enable-m-guard.
* ccid-driver.c (read_device_info): Fix segfault when usb device
is not accessible.
(ccid_open_reader): Allow working with an even older version of
libusb (usb_busses global instead of usb_get_busses()).
2004-09-09 Werner Koch <wk@g10code.com>
* cardglue.h: Add members for CA fingerprints.
* cardglue.c (agent_release_card_info): Invalid them.
(learn_status_cb): Store them.
* app-common.h, app-openpgp.c, iso7816.c, iso7816.h
* apdu.c, apdu.h, ccid-driver.c, ccid-driver.h
* card-util.c: Updated from current gnupg-1.9.
Changes are:
* ccid-driver.h (CCID_DRIVER_ERR_ABORTED): New.
* ccid-driver.c (ccid_open_reader): Support the stable 0.1 version
of libusb.
(ccid_get_atr): Handle short messages.
* apdu.c (my_rapdu_get_status): Implemented.
* apdu.c: Include <signal.h>.
* apdu.c (reader_table_s): Add function pointers for the backends.
(apdu_close_reader, apdu_get_status, apdu_activate)
(send_apdu): Make use of them.
(new_reader_slot): Intialize them to NULL.
(dump_ccid_reader_status, ct_dump_reader_status): New.
(dump_pcsc_reader_status): New.
(open_ct_reader, open_pcsc_reader, open_ccid_reader)
(open_osc_reader, open_rapdu_reader): Intialize function pointers.
(ct_activate_card, ct_send_apdu, pcsc_send_apdu, osc_send_apdu)
(error_string): Removed. Replaced by apdu_strerror.
(get_ccid_error_string): Removed.
(ct_activate_card): Remove the unused loop.
(reset_ct_reader): Implemented.
(ct_send_apdu): Activate the card if not yet done.
(pcsc_send_apdu): Ditto.
* ccid-driver.h: Add error codes.
* ccid-driver.c: Implement more or less proper error codes all
over the place.
* apdu.c (apdu_send_direct): New.
(get_ccid_error_string): Add some error code mappings.
(send_apdu): Pass error codes along for drivers already supporting
them.
(host_sw_string): New.
(get_ccid_error_string): Use above.
(send_apdu_ccid): Reset the reader if it has not yet been done.
(open_ccid_reader): Don't care if the ATR can't be read.
(apdu_activate_card): New.
(apdu_strerror): New.
(dump_reader_status): Only enable it with opt.VERBOSE.
* iso7816.c (map_sw): Add mappings for the new error codes.
* apdu.c (open_ct_reader, open_pcsc_reader, open_ccid_reader)
(reset_ccid_reader, open_osc_reader): Call dump_reader_status only
in verbose mode.
* app-openpgp.c (do_getattr): Fix for sending CA-FPR.
* app-openpgp.c (app_openpgp_readkey): Fixed check for valid
exponent.
* app-openpgp.c (do_setattr): Sync FORCE_CHV1.
* card-util.c (change_login): Kludge to allow reading data from a
file.
(card_edit): Pass ARG_STRING to change_login.
(card_status): Print CA fingerprints.
(change_cafpr): New.
(card_edit): New command CAFPR.
2004-04-30 Werner Koch <wk@gnupg.org>
* g10.c (main) <gpgconf>: Use gpg.conf and not /dev/null as
default filename.
2004-04-28 Werner Koch <wk@gnupg.org>
* card-util.c (card_edit): Remove PIN verification.
(generate_card_keys): New arg SERIALNO. Do PIN verification here
after resetting forced_chv1.
2004-09-09 Werner Koch <wk@g10code.com>
* signal.c (got_fatal_signal): Do readline cleanup. Print signal
number if we can't print the name. Use new autoconf macro
HAVE_DECL_SYS_SIGLIST.
(get_signal_name): Removed.
* photoid.c: Include ttyio.h.
* parse-packet.c (skip_rest): Removed. Changed all callers to use
the new iobuf_skip_reset. Orginal patch by Florian Weimer.
2004-09-07 Werner Koch <wk@g10code.de>
* photoid.c (generate_photo_id): Use tty_printf and not just
printf. Put _() around one string.
2004-09-03 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (parse_keyrec): Force the 'e'xpired flag on as soon
as we know the key is definitely expired. Some translatable
string cleanup.
2004-08-27 David Shaw <dshaw@jabberwocky.com>
* encode.c, exec.c, g10.c, sign.c: Some translatable string
cleanup. Change some "this" to `this'.
2004-08-23 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_spawn): Show log line for what keyserver
action we are taking.
* keyid.c (keystr): If printing a keyid that lacks the high 4
bytes, print the low 4 alone.
(keystr_from_desc): Handle short keyids and warn on v3
fingerprints.
* keydb.h, getkey.c (get_user_id_printable,
get_user_id_string_printable): Rename to get_user_id_native and
get_user_id_string_native and remove the printable stuff since
we're print-ifying valid utf8 characters. Change all callers in
import.c, sign.c, keylist.c, and encode.c.
* keyserver.c (keyserver_search_prompt): Make sure the search
string is converted from UTF-8 before display.
2004-08-19 Werner Koch <wk@g10code.de>
* seskey.c (encode_session_key): Changed the zero random byte
substituting code to actually do clever things. Thanks to
Matthias Urlichs for noting the implementation problem.
2004-08-18 Marcus Brinkmann <marcus@g10code.de>
* passphrase.c (agent_get_passphrase): Fix detection of gpg-agent
cancellation.
2004-08-08 David Shaw <dshaw@jabberwocky.com>
* plaintext.c (handle_plaintext): Bigger buffer for extra safety.
* g10.c (main): New alias --throw-keyid for --throw-keyids, so
that it continues to work in old configuration files. Noted by
Jens Adam.
* pkclist.c (algo_available): --pgp8 now allows blowfish, zlib,
and bzip2.
* status.c (do_get_from_fd): Flush stdout if status isn't flushing
it for us. This guarantees that any menus that were displayed
before the prompt don't get stuck in a buffer. Noted by Peter
Palfrader. This is Debian bug #254072.
* sign.c (update_keysig_packet): Revert change of 2004-05-18. It
is not appropriate to strip policy and notations when remaking a
sig. That should only happen when specifically requested by the
user.
2004-08-05 David Shaw <dshaw@jabberwocky.com>
* armor.c (radix64_read): No armor CRC is legal according to the
spec (the CRC is a MAY).
2004-07-28 David Shaw <dshaw@jabberwocky.com>
* misc.c (argsplit): Properly split quoted args from the keyword
and trim whitespace afterwards.
2004-07-27 David Shaw <dshaw@jabberwocky.com>
* misc.c (optsep): Add the ability to understand keyword="quoted
arg with spaces" type options.
2004-07-16 David Shaw <dshaw@jabberwocky.com>
* keylist.c (list_keyblock_print): Always use the new listing
format where uids are always on a line for themselves. Mark
expired secret keys as expired.
* options.h, g10.c (main): Rename list show-validity to
show-uid-validity as it only shows for uids.
* armor.c (armor_filter): Do not use padding to get us to 8 bytes
of header. Rather, use 2+4 as two different chunks. This avoids
a fake filename of "is".
2004-07-15 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (sign_uids): Properly handle remaking a self-sig on
revoked or expired user IDs. Also, once we've established that a
given uid cannot or will not be signed, don't continue to ask
about each sig.
* mainproc.c (proc_symkey_enc), seckey-cert.c (do_check): Check
the S2K hash algorithm before we try to generate a passphrase
using it. This prevents hitting BUG() when generating a
passphrase using a hash that we don't have.
* sign.c (sign_symencrypt_file): Allow using --force-mdc in --sign
--symmetric messages.
* g10.c (main): Alias --charset as --display-charset to help avoid
the continuing confusion and make room for possible changes in
devel.
* parse-packet.c (parse_plaintext): Show the hex value for the
literal packet mode since it may not be printable.
* keygen.c (make_backsig): Make sure that the backsig was built
successfully before we try and use it.
* status.h, status.c (get_status_string), plaintext.c
(handle_plaintext): New status tags PLAINTEXT and
PLAINTEXT_LENGTH.
2004-06-16 Werner Koch <wk@gnupg.org>
* free-packet.c (copy_secret_key): Get last fix right.
2004-06-16 Werner Koch <wk@gnupg.org>
* free-packet.c (copy_secret_key): Fixed memory leak when D is not
NULL.
* passphrase.c (passphrase_to_dek): Added a few comments to the
code.
2004-05-26 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_refresh): Keep track of keys already
fetched so we don't do a regular keyserver fetch if the preferred
keyserver fetch has exhausted the list.
2004-05-23 David Shaw <dshaw@jabberwocky.com>
* verify.c (verify_signatures): Verify multiple files in the same
order in which we hashed them when issuing the signature. Noted
by Nicholas Cole.
* pkclist.c (do_edit_ownertrust): Fix a kbnode leak and do another
keyid-format conversion.
2004-05-22 Werner Koch <wk@gnupg.org>
* trustdb.c (check_regexp): s/EXP/EXPR/.
* keyedit.c (trustsig_prompt): Remoev useless range check.
* options.h: Renamed ctrl to glo_ctrl. Changed all users.
* ccid-driver.c (ccid_open_reader): Print a warning when CCID
can't be used.
2004-05-21 David Shaw <dshaw@jabberwocky.com>
* mainproc.c (check_sig_and_print): If we're honoring preferred
keyservers, and auto-key-retrieve is set, try and get a missing
key from the preferred keyserver subpacket when we verify the sig.
* gpgv.c (parse_preferred_keyserver, free_keyserver_spec): Stubs.
* keyserver.c (keyidlist): Use new parse_preferred_keyserver
function.
(keyserver_work): Use the passed-in keyserver spec rather than the
options global one.
* keyserver-internal.h, keyserver.c (parse_preferred_keyserver):
New function to take a sig and return a split out keyserver_spec.
(keyserver_import_keyid): Now takes a keyserver_spec.
* keyserver.c (keyidlist): Go back to the old fast keyid lister.
Only merge selfsigs if we have to for honor-keyserver-url.
(keyserver_refresh): Keyserver URL handler moved here.
(calculate_keyid_fpr): Removed.
* keydb.h, keyid.c (keystr_from_desc): Calculate a key string from
a KEYDB_SEARCH_DESC.
* keyserver.c (keyserver_spawn): Fix keyserver options on tempfile
only platforms. Noted by Roger Sondermann.
2004-05-20 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_work): Allow --refresh-keys with a
preferred keyserver to happen even if there is no global keyserver
set.
* sig-check.c (do_check_messages): No need to check for Elgamal
signatures any longer.
(do_check_messages, do_check, check_key_signature2):
--keyid-format conversion.
* pkclist.c (show_paths, edit_ownertrust): Remove some unused
code.
* options.h (ctrl): New for member IN_AUTO_KEY_RETRIEVE.
* mainproc.c (check_sig_and_print): track whether we are
retrieving a key.
* status.c (status_currently_allowed): New.
(write_status_text, write_status_text_and_buffer): Use it here.
* g10.c: New command --gpgconf-list.
(gpgconf_list): New. From Werner on stable branch.
2004-05-19 David Shaw <dshaw@jabberwocky.com>
* pubkey-enc.c (get_session_key, get_it), keyedit.c
(show_key_with_all_names, show_basic_key_info): --keyid-format
conversion.
2004-05-18 David Shaw <dshaw@jabberwocky.com>
* sign.c (update_keysig_packet): Policies and notations should be
stripped out when remaking a self-signature. Noted by Atom
Smasher.
* keyserver.c (parse_keyserver_uri): Fix compiler warnings.
2004-05-11 David Shaw <dshaw@jabberwocky.com>
* options.h, keyserver-internal.h, keyserver.c
(parse_keyserver_uri): Improved URI parser that keeps track of the
path information and doesn't modify the input string.
(keyserver_spawn): Tell keyserver plugins about the path.
2004-05-11 Werner Koch <wk@gnupg.org>
* keylist.c (show_policy_url, show_keyserver_url, show_notation)
(list_one): Use const char* for i18n string helpers.
* keygen.c (do_generate_keypair, read_parameter_file): Really
close the files.
(do_generate_keypair): Create the secret key file using safe
permissions. Noted by Atom Smasher.
2004-05-10 David Shaw <dshaw@jabberwocky.com>
* options.h, mainproc.c (symkey_decrypt_seskey), keyserver.c
(struct keyrec, parse_keyrec, keyserver_search_prompt), keyedit.c
(keyedit_menu), g10.c (add_keyserver_url, add_policy_url): Fix
some compiler warnings.
2004-05-08 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu, menu_set_keyserver_url): Allow passing
preferred keyserver on "keyserver" command line. Sanity check
keyserver URL before accepting it.
* keyserver-internal.h, g10.c (main), keyserver.c
(parse_keyserver_uri): Add an option to require the scheme:// and
change all callers.
(free_keyserver_spec): Make public.
2004-05-07 Werner Koch <wk@gnupg.org>
* sign.c (write_plaintext_packet): Fixed the detection of too
large files in the same way as in encode.c.
2004-05-04 David Shaw <dshaw@jabberwocky.com>
* keylist.c (show_notation): Use bits to select which sort of
notation to show. Don't allow a not-shown notation to prevent us
from issuing the proper --status-fd message.
* options.h, g10.c (main): Add show-std/standard-notations and
show-user-notations. show-notations is both. Default is to show
standard notations only during verify. Change all callers.
2004-04-28 David Shaw <dshaw@jabberwocky.com>
* main.h, keylist.c (show_notation): Add argument to show only
user notations, only standard notations, or both. Change all
callers.
* keyserver.c (keyserver_spawn): We still need EXEC_TEMPFILE_ONLY.
2004-04-28 Werner Koch <wk@gnupg.org>
* card-util.c (card_edit): Require PIN only for generate.
* app-openpgp.c (do_setattr): Sync FORCE_CHV1.
2004-04-27 Werner Koch <wk@gnupg.org>
* keyserver.c (keyserver_spawn) [EXEC_TEMPFILE_ONLY]: Removed
setting use_temp_file because this option has been removed.
* g10.c: New commands --allow-admin and --deny-admin.
* options.h (opt): Add member ALLOW_ADMIN.
* tlv.h, tlv.c: New. Copied from gnupg-1.9.
* cardglue.c (open_card): The serialno is now set internally by
app_select_openpgp; changed invocation.
* cardglue.h (app_t, ctrl_t): New.
(GPG_ERR_EBUSY, GPG_ERR_ENOENT, GPG_ERR_NOT_FOUND, GPG_ERR_BUG)
(GPG_ERR_NOT_IMPLEMENTED, GPG_ERR_EACCESS): New.
(gpg_err_code_from_errno): New.
* app-common.h, app-openpgp.c, iso7816.c, iso7816.h
* apdu.c, apdu.h, ccid-driver.c, ccid-driver.h
* card-util.c: Updated from current gnupg-1.9.
Changes are:
* app-common.h: New members FNC.DEINIT and APP_LOCAL.
* app-openpgp.c (do_deinit): New.
(get_cached_data, flush_cache_item, flush_cache_after_error)
(flush_cache): New.
(get_one_do): Replaced arg SLOT by APP. Make used of cached data.
(verify_chv2, verify_chv3): Flush some cache item after error.
(do_change_pin): Ditto.
(do_sign): Ditto.
(do_setattr): Flush cache item.
(do_genkey): Flush the entire cache.
(compare_fingerprint): Use cached data.
* apdu.c (apdu_send_le): Reinitialize RESULTLEN. Handle
SW_EOF_REACHED like SW_SUCCESS.
* ccid-driver.c (parse_ccid_descriptor): Store some of the reader
features away. New arg HANDLE
(read_device_info): New arg HANDLE. Changed caller.
(bulk_in): Handle time extension requests.
(ccid_get_atr): Setup parameters and the IFSD.
(compute_edc): New. Factored out code.
(ccid_transceive): Use default NADs when required.
* apdu.h: New pseudo stati SW_HOST_NOT_SUPPORTED,
SW_HOST_LOCKING_FAILED and SW_HOST_BUSY.
* iso7816.c (map_sw): Map it.
* ccid-driver.c (ccid_slot_status): Add arg STATUSBITS.
* apdu.c (apdu_get_status): New.
(ct_get_status, pcsc_get_status, ocsc_get_status): New stubs.
(get_status_ccid): New.
(apdu_reset): New.
(reset_ct_reader, reset_pcsc_reader, reset_osc_reader): New stubs.
(reset_ccid_reader): New.
(apdu_enum_reader): New.
* apdu.c (lock_slot, trylock_slot, unlock_slot): New helpers.
(new_reader_slot) [USE_GNU_PTH]: Init mutex.
(apdu_reset, apdu_get_status, apdu_send_le): Run functions
in locked mode.
2004-04-25 David Shaw <dshaw@jabberwocky.com>
* getkey.c (get_seckey_byname2): Significantly simplify this
function by using key_byname to do the heavy lifting. Note that
this also fixes an old problem when the first key on the secret
keyring has an unusable stub primary, but is still chosen.
* getkey.c (key_byname): If namelist is NULL, return the first key
in the keyring.
2004-04-22 David Shaw <dshaw@jabberwocky.com>
* keygen.c (make_backsig): If DO_BACKSIGS is not defined, do not
create backsigs.
* getkey.c (merge_selfsigs_subkey): Find 0x19 backsigs on subkey
selfsigs and verify they are valid. If DO_BACKSIGS is not
defined, fake this as always valid.
* packet.h, parse-packet.c (parse_signature): Make parse_signature
non-static so we can parse 0x19s in self-sigs.
* main.h, sig-check.c (check_backsig): Check a 0x19 signature.
(signature_check2): Give a backsig warning if there is no or a bad
0x19 with signatures from a subkey.
2004-04-21 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (dump_sig_subpkt, parse_one_sig_subpkt,
can_handle_critical): Parse and display 0x19 signatures.
2004-04-20 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (parse_keyserver_uri): Do not accept "http" as an
alias for "hkp". They are not the same thing.
2004-04-19 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main): Add keyserver-option
honor-keyserver-url. parse_keyserver_options now returns a
success code.
* keyserver.c (parse_keyserver_options): Return error on failure
to parse. Currently there is no way to fail as any unrecognized
options get saved to be sent to the keyserver plugins later.
Check length of keyserver option tokens since with =arguments we
must only match the prefix.
(free_keyserver_spec): Moved code from parse_keyserver_url.
(keyserver_work, keyserver_spawn): Pass in a struct keyserver_spec
rather than using the global keyserver option.
(calculate_keyid_fpr): New. Fills in a KEYDB_SEARCH_DESC for a
key.
(keyidlist): New implementation using get_pubkey_bynames rather
than searching the keydb directly. If honor-keyserver-url is set,
make up a keyserver_spec and try and fetch that key directly. Do
not include it in the returned keyidlist in that case.
2004-04-16 David Shaw <dshaw@jabberwocky.com>
* plaintext.c (handle_plaintext): Accept 'u' as a plaintext mode
that requires end of line conversion. This is being considered
for a UTF8 text packet. If this doesn't take place, no major harm
done. If it does take place, we'll get a jump on starting the
changeover.
* g10.c (main): --no-use-embedded-filename.
* build-packet.c (calc_plaintext, do_plaintext): Do not create
illegal (packet header indicates a size larger than the actual
packet) encrypted data packets when not compressing and using a
filename longer than 255 characters.
* keyedit.c (no_primary_warning): Cleanup. (menu_expire): Don't
give primary warning for subkey expiration changes. These cannot
reorder primaries.
* keygen.c (gen_elg, gen_dsa, gen_rsa, do_create,
do_generate_keypair, generate_subkeypair): New is_subkey argument
to set whether a generated key is a subkey. Do not overload the
ret_sk. This is some early cleanup to do backsigs for signing
subkeys.
* keygen.c (write_keybinding, do_generate_keypair,
generate_subkeypair): Keep track of the unprotected subkey secret
key so we can make a backsig with it.
* keygen.c (make_backsig): New function to add a backsig to a
binding sig of signing subkeys. Currently disabled.
(write_keybinding): Call it here, for signing subkeys only.
* sign.c (make_keysig_packet): Allow generating 0x19 signatures
(same as 0x18 or 0x28, but used for backsigs).
* packet.h, build-packet.c (build_sig_subpkt): Add new
SIGSUBPKT_SIGNATURE type for embedded signatures.
* main.h, misc.c (optsep, argsplit, optlen, parse_options):
Simplify code and properly handle a partial match against an
option with an argument.
* keyserver-internal.h, keyserver.c (parse_keyserver_options): Use
new optsep and argsplit functions.
2004-04-15 David Shaw <dshaw@jabberwocky.com>
* main.h, misc.c (argsplit): Refactor argsep into argsplit and
argsep so they can be called separately.
* options.h, keyserver.c (parse_keyserver_options): Remove
duplicate code from parse_keyserver_options by calling the generic
parse_options.
* keyserver.c (keyserver_spawn, keyserver_refresh), g10.c (main),
gpgv.c (main), mainproc.c (check_sig_and_print), import.c
(revocation_present): Change all callers.
2004-04-14 David Shaw <dshaw@jabberwocky.com>
* packet.h, getkey.c (fixup_uidnode, merge_selfsigs_subkey): Keep
track of which self-sig we actually chose.
* keyedit.c (menu_expire, menu_set_primary_uid,
menu_set_preferences): Use it here to avoid updating non-used
self-sigs and possibly promoting an old self-sig into
consideration again.
* options.h, import.c, keyserver-internal.h, g10.c, mainproc.c,
keyserver.c (parse_keyserver_uri): Parse keyserver URI into a
structure. Cleanup for new "guess my keyserver" functionality, as
well as refreshing via a preferred keyserver subpacket.
* options.h: Encapsulate keyserver details. Change all callers.
2004-04-05 Werner Koch <wk@gnupg.org>
* status.h (STATUS_NEWSIG): New.
* status.c (get_status_string): Add it.
2004-03-27 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu): Request a trustdb update when adding a
new user ID so the new ID gets validity set. Reported by Owen
Taylor.
2004-03-25 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), compress-bz2.c (init_uncompress):
Rename --bzip2-compress-lowmem to --bzip2-decompress-lowmem since
it applies to decompression, not compression.
2004-03-24 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (sign_uids, show_key_and_fingerprint, ask_revoke_sig,
menu_revsig, menu_showphoto): --keyid-format conversion.
(menu_addrevoker): Use print_pubkey_info() rather than duplicating
code.
2004-03-19 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (update_min_ownertrust, validate_keys): Do not use
keystr functions in log_debug.
* import.c (import_one): Try and collapse user IDs when importing
a key for the first time.
* keyedit.c (menu_addrevoker): Allow appointing a subkey as a
designated revoker if the user forces it via keyid!, so long as
the subkey can certify. Also use the proper date string when
prompting for confirmation.
* g10.c (main): Maintain ordering of multiple Comment lines.
Requested by Peter Hyman.
2004-03-17 David Shaw <dshaw@jabberwocky.com>
* mainproc.c (proc_pubkey_enc, print_pkenc_list, list_node):
--keyid-format conversion.
2004-03-16 David Shaw <dshaw@jabberwocky.com>
* getkey.c (skip_unusable, merge_selfsigs_main,
premerge_public_with_secret, lookup, get_user_id_string):
--keyid-format conversion.
2004-03-15 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (add_utk, verify_own_keys, update_min_ownertrust,
get_validity, ask_ownertrust, validate_keys): --keyid-format
conversion.
* import.c (check_prefs_warning, check_prefs): --keyid-format
conversion and a little better text.
(import_one, import_secret_one, import_revoke_cert, chk_self_sigs,
delete_inv_parts, merge_blocks): Still more --keyid-format
conversions.
2004-03-06 David Shaw <dshaw@jabberwocky.com>
* keylist.c (print_seckey_info, print_pubkey_info): --keyid-format
conversion.
(list_keyblock_print): 0xshort should not push us into the new
list format since it is not much longer than regular 8-character
short keyids.
* keydb.h, keyid.c (keystr_from_pk, keystr_from_sk): New functions
to pull a key string from a key in one step. This isn't faster
than before, but makes for neater code.
* keylist.c (list_keyblock_print): Use keystr_from_xx here.
(print_key_data): No need to pass a keyid in.
2004-03-05 David Shaw <dshaw@jabberwocky.com>
* keyid.c (keyid_from_sk): Minor performance boost by caching
secret key keyids so we don't have to calculate them each time.
* getkey.c (merge_selfsigs_subkey): Do not mark subkeys valid if
we do not support their pk algorithm. This allows for early
(during get_*) rejection of a subkey, and selection of another.
* passphrase.c (passphrase_to_dek): Give a little more information
when we have room to do so.
2004-03-04 David Shaw <dshaw@jabberwocky.com>
* revoke.c (export_minimal_pk), export.c (do_export_stream),
passphrase.c (passphrase_to_dek), keyserver.c (print_keyrec): A
few more places to use --keyid-format.
* options.h, g10.c (main), export.c (parse_export_options,
do_export_stream): Remove --export-all and the "include-non-rfc"
export-option as they are no longer meaningful with the removal of
v3 Elgamal keys.
* armor.c (fake_packet, armor_filter): Use the 2440 partial length
encoding for the faked plaintext packet.
2004-03-03 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), mainproc.c (check_sig_and_print):
Remove verify-option show-long-keyids and replace with
the more general keyid-format.
* build-packet.c (write_header2): Remove call to start old gpg
partial length mode and change all callers.
(do_plaintext): Turn off partial length encoding now that we're
done writing the packet.
(do_comment, do_user_id): Try for a headerlen of 2 since that's
the smallest and most likely encoding for these packets.
* parse-packet.c (parse): Remove call to start old gpg partial
length mode.
2004-03-02 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main): Add a more flexible --keyid-format
option to replace the list-option (and eventually verify-option)
show-long-keyids. The format can be short, long, 0xshort, and
0xlong.
* keydb.h, keyid.c (keystr, keystrlen): New functions to generate
a printable keyid.
* keyedit.c (print_and_check_one_sig, show_key_with_all_names),
keylist.c (list_keyblock_print): Use new keystr() function here to
print keyids.
* packet.h, free-packet.c (free_encrypted, free_plaintext),
parse-packet.c (copy_packet, skip_packet, skip_rest, read_rest,
parse_plaintext, parse_encrypted, parse_gpg_control): Use a flag
to indicate partial or indeterminate encoding. This is the first
step in some minor surgery to remove the old gpg partial length
encoding.
2004-03-01 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (parse): Only data-type packets are allowed to
use OpenPGP partial length encoding.
2004-02-25 David Shaw <dshaw@jabberwocky.com>
* delkey.c (do_delete_key): Allow deleting a public key with a
secret present if --expert is set.
* plaintext.c (handle_plaintext): Make bytecount static so it
works with multiple literal packets inside a message.
* encode.c, helptext.c (keygen.algo, keygen.algo.elg_se), keygen.c
(ask_algo), sig-check.c (do_check_messages), skclist.c
(build_sk_list): Rename "ElGamal" to "Elgamal" as that is the
proper spelling nowadays. Suggested by Jon Callas.
2004-02-24 David Shaw <dshaw@jabberwocky.com>
* plaintext.c: Copyright.
* encode.c (encode_simple): Show cipher with --verbose.
* options.h, g10.c (main), keyedit.c (sign_keys): Add
--ask-cert-level option to enable cert level prompts during
sigs. Defaults to on. Simplify --default-cert-check-level to
--default-cert-level. If ask-cert-level is off, or batch is on,
use the default-cert-level as the cert level.
* options.h, g10.c (main), trustdb.c (mark_usable_uid_certs):
Simplify --min-cert-check-level to --min-cert-level.
2004-02-22 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), trustdb.c (mark_usable_uid_certs): Add
--min-cert-check-level option to specify minimum cert check level.
Defaults to 2 (so 0x11 sigs are ignored). 0x10 sigs cannot be
ignored.
2004-02-21 David Shaw <dshaw@jabberwocky.com>
* plaintext.c (handle_plaintext): Properly handle a --max-output
of zero (do not limit output at all).
* keyserver.c (keyserver_spawn): Use the full 64-bit keyid in the
INFO header lines, and include "sig:" records for the benefit of
people who store their keys in LDAP servers. It makes it easy to
do queries for things like "all keys signed by Isabella".
* main.h, misc.c (hextobyte): Removed. It's in libutil.a now.
2004-02-20 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_export): Disallow user strings that
aren't key IDs.
(keyserver_import): Clarify error message.
(keyserver_spawn): Properly handle 8 bit characters in user IDs in
the info lines during SEND.
* mkdtemp.c: Removed.
* Makefile.am: We get mkdtemp.c from libutil.a now, so don't link
with @LIBOBJS@.
* keyserver.c (keyserver_spawn): Pass the scheme to the keyserver
helper.
2004-02-18 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), plaintext.c (handle_plaintext): Add
--max-output option to help people deal with decompression bombs.
2004-02-15 David Shaw <dshaw@jabberwocky.com>
* build-packet.c (do_user_id): Do not force a header for attribute
packets as they require a new CTB, and we don't support forced
headers for new CTBs yet.
2004-02-14 David Shaw <dshaw@jabberwocky.com>
* build-packet.c (write_header2): If a suggested header length is
provided along with a zero length, interpret this as an actual
zero length packet and not as an indeterminate length packet.
(do_comment, do_user_id): Use it here as these packets might be
naturally zero length.
* parse-packet.c (parse): Show packet type when failing due to an
indeterminate length packet.
* misc.c (parse_options): Only provide args for the true (i.e. not
"no-xxx") form of options.
2004-02-13 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (argsep): Move to misc.c.
* main.h, misc.c (parse_options), export.c (parse_export_options),
import.c (parse_import_options), g10.c (main): Use it here to
allow for options with optional arguments. Change all callers.
* import.c (check_prefs): Some language fixes.
(sec_to_pub_keyblock, import_secret_one): Without knowing the
number of MPIs there are, we cannot try and sk-to-pk-ize a key.
2004-02-12 David Shaw <dshaw@jabberwocky.com>
* import.c (check_prefs): New function to check preferences on a
public key to ensure that it does not advertise any that we cannot
fulfill. Use the keyedit command list function to optionally
rewrite the prefs.
(import_one, import_secret_one): Use it here when importing a
public key that we have the secret half of, or when importing a
secret key that we have the public half of.
* main.h, keyedit.c (keyedit_menu): Remove sign_mode and enhance
the more general command list functionality to replace it.
* g10.c (main): Use the general command functionality to implement
--sign-key, --lsign-key, --nrsign-key, and --nrlsign-key.
* import.c (import_one): Do the revocation check even in the case
when a key, a revocation key set in a direct key signature, and a
revocation from that revocation key, all arrive piecemeal.
Needless to say, this is pretty obscure.
2004-02-11 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), keylist.c (list_keyblock_print): Add
"show-unusable-subkeys" list-option to show revoked and/or expired
subkeys.
2004-02-10 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu): Prompt for subkey removal for both
secret and public subkeys.
* keylist.c (list_keyblock_print), keyedit.c
(show_key_with_all_names): Show the revocation date of a
key/subkey, and general formatting work.
* packet.h, getkey.c (merge_selfsigs_main, merge_selfsigs_subkey,
merge_selfsigs): Keep track of the revocation date of a key.
* keydb.h, keyid.c (revokestr_from_pk): New function to print the
revocation date of a key.
* keygen.c (keygen_set_std_prefs): Build the default preferences
list at runtime as it properly handles algorithms disabled at
build or run time.
* getkey.c (merge_selfsigs_main): Properly handle expired user IDs
when the expired self-sig is not the only self-sig.
* misc.c (compress_algo_to_string): Return NULL on failure like
all of the other xxxx_algo_to_string() functions.
* mainproc.c (list_node): Minor spacing tweak to match --list-keys
output.
* keylist.c (list_keyblock_print), mainproc.c (list_node): Mark
revoked subkeys as revoked. Requested by Matthew Wilcox. Revoked
overrides expiration when both apply.
* keyedit.c (show_prefs): Use compress algo constants.
(show_basic_key_info): Make revoked and expired tags translatable.
* g10.c (rm_group): Properly ungroup from a list of groups.
2004-01-30 David Shaw <dshaw@jabberwocky.com>
* g10.c (main, rm_group): Add --ungroup command to remove a
particular group.
(add_group): When adding a group with the same name as an already
existing group, merge the two groups.
(list_config): Show an error message when listing a config item
that doesn't exist.
(main): Replace -z0 trick for no compression.
* packet.h, keyedit.c (show_key_with_all_names_colon), keylist.c
(list_keyblock_colon), mainproc.c (list_node, proc_tree): Minor
cleanup to remove local_id, which is no longer used.
2004-01-27 David Shaw <dshaw@jabberwocky.com>
* getkey.c: Set MAX_PK_CACHE_ENTRIES and MAX_UID_CACHE_ENTRIES to
PK_UID_CACHE_SIZE (set in ./configure).
* getkey.c (get_pubkey): When reading key data into the cache,
properly handle keys that are partially (pk, no UIDs) cached
already. This is Debian bug #176425 and #229549.
* compress.c (init_compress, push_compress_filter2): Do the right
thing (i.e. nothing) with compress algo 0.
* main.h, decrypt.c (decrypt_messages): Accept filenames to
decrypt on stdin. This is bug #253.
2004-01-23 David Shaw <dshaw@jabberwocky.com>
* mainproc.c (list_node): Show sigs with --verbose.
* options.h, g10.c (set_screen_dimensions): New function to look
at COLUMNS and LINES.
* keyserver.c (parse_keyrec, keyserver_search_prompt), keyedit.c
(print_and_check_one_sig): Use new screen dimension variables.
2004-01-21 David Shaw <dshaw@jabberwocky.com>
* g10.c (list_config): New function to dump config options to
stdout. Currently requires --with-colons.
(collapse_args): New function to turn argc/argv into a single
string.
(main): Use it here to pass list_config() more than one argument
as a single string.
(print_algo_numbers): Helper to print algorithm number for
--list-config "pubkey", "cipher", "hash"/"digest", and "compress"
config options.
* packet.h, getkey.c (merge_selfsigs, merge_selfsigs_main),
pkclist.c (check_signatures_trust): Indicate who has revoked a key
(the owner or a designated revoker). If a key was revoked by
both, prefer the owner. If a subkey is already revoked by the
owner, don't allow a designated revokation of the whole key to
override this. We're still revoked either way, of course.
* keyedit.c (print_and_check_one_sig, keyedit_menu): Use the
COLUMNS environment variable (if any) to hint how wide the
terminal is. Disabled on _WIN32. Suggested by Janusz
A. Urbanowicz.
2004-01-20 David Shaw <dshaw@jabberwocky.com>
* keylist.c (set_attrib_fd): Open attribute fd in binary
mode. This isn't meaningful on POSIX systems, but the Mingw builds
aren't exactly POSIX.
* trustdb.c (reset_trust_records): New, faster, implementation
that doesn't involve a keyring scan.
(clear_validity): Removed.
* g10.c (main), keydb.h, keydb.c (keydb_rebuild_caches),
keyring.h, keyring.c (keyring_rebuild_cache): Add "noisy" flag so
cache rebuilds can remain noisy when called for itself, and quiet
when called as part of the trustdb rebuild.
* trustdb.c (validate_keys): Rebuild the sig caches before
building the trustdb. Note that this is going to require some
architectual re-thinking, as it is agonizingly slow.
2004-01-19 David Shaw <dshaw@jabberwocky.com>
* sig-check.c (check_key_signature2): Comments.
* keyring.c (keyring_rebuild_cache): Clear sig cache for any
signatures that we can no longer process (say, if the user removed
support for a necessary pubkey or digest algorithm).
2004-01-16 David Shaw <dshaw@jabberwocky.com>
* misc.c (print_cipher_algo_note): May as well call Rijndael AES
at this point.
* keygen.c (do_create), misc.c (openpgp_pk_algo_usage): Remove the
last bits of Elgamal type 20 support.
2004-01-03 Stefan Bellon <sbellon@sbellon.de>
* compress.c [__riscos__]: Only use RISC OS' own ZLib module if
configured to use it.
2003-12-30 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), import.c (parse_import_options,
import_one, import_secret_one), keyserver.c (keyserver_refresh):
Change --merge-only to --import-option merge-only. Deprecate
--merge-only.
2003-12-29 David Shaw <dshaw@jabberwocky.com>
* misc.c (pull_in_libs): Dead code. Removed.
* sig-check.c (check_revocation_keys): Comments.
* getkey.c (merge_selfsigs_main): Don't bother to check designated
revoker sigs if the key is already revoked.
* packet.h, getkey.c (merge_selfsigs_main): New "maybe_revoked"
flag on PKs. It is set when there is a revocation signature from
a valid revocation key, but the revocation key is not present to
verify the signature.
* pkclist.c (check_signatures_trust): Use it here to give a
warning when showing key trust.
* compress-bz2.c: Include stdio.h. Solaris 9 has a very old bzip2
library and we can at least guarantee that it won't fail because
of the lack of stdio.h.
* tdbio.c: Fixed format string bugs related to the use of DB_NAME.
Reported by Florian Weimer.
2003-12-28 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), keyserver.c (keyserver_opts,
parse_keyserver_uri): honor-http-proxy is no longer an option
since we can do the same thing with http-proxy with no arguments.
Also remove broken-http-proxy since it can be better handled in
the HTTP helper.
* keyserver.c (argsep): New variation on strsep that knows about
optional arguments.
(parse_keyserver_options): Use it here for optional arguments.
2003-12-28 Stefan Bellon <sbellon@sbellon.de>
* plaintext.c (handle_plaintext) [__riscos__]: Don't mangle
filename if the user specified it.
* g10.c, gpgv.c [__riscos__]: Removal of unnecessary #ifdef
__riscos__ sections.
2003-12-27 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (strip_leading_space, get_arg): New.
(parse_keyserver_options): Use them here to allow arguments to
keyserver-options. Since none of our options need arguments yet,
just pass them through whole to the keyserver helper.
* main.h, misc.c (parse_options): Add a "noisy" flag to enable and
disable the messages about which option didn't match or matched
ambiguously. Change all callers (g10.c, keyserver.c).
* main.h, import.c (import_options), export.c (export_options):
Pass the noisy flag through.
2003-12-17 David Shaw <dshaw@jabberwocky.com>
* build-packet.c (write_fake_data, do_secret_key), seckey-cert.c
(do_check): Use an unsigned length for mpi_get_opaque.
* options.h: It's impolite to assign -1 to an unsigned
opt.force_ownertrust.
* sig-check.c (cmp_help, do_check), sign.c (do_sign): Remove old
unused code.
* keyid.c (keyid_from_sk): Make sure lowbits is initialized.
2003-12-12 David Shaw <dshaw@jabberwocky.com>
* sig-check.c (do_check): Move the signing algo and hash checks
from here...
(signature_check2): ... to here.
(check_key_signature2): ... and here. This is a minor
optimization to avoid fetching a key (which can be expensive,
especially if it is not self-signed, and there are many key
signatures on it which need to be checked for ultimate trust) if
the signature would have failed anyway because of algorithm or
hash problems.
2003-12-10 David Shaw <dshaw@jabberwocky.com>
* packet.h, build-packet.c (hash_public_key): Remove function ...
* keydb.h, keyid.c (hash_public_key, do_fingerprint_md): ... and
make a new one here that shares code with the fingerprint
calculations. This removes some duplicated functionality, and is
also around 14% faster. (Every bit helps).
* import.c (import_one): No longer need the Elgamal import
warning.
* getkey.c (get_pubkey_fast): This one is sort of obscure.
get_pubkey_fast returns the primary key when requesting a subkey,
so if a user has a key signed by a subkey (we don't do this, but
used to), AND that key is not self-signed, AND the algorithm of
the subkey in question is not present in GnuPG, AND the algorithm
of the primary key that owns the subkey in question is present in
GnuPG, then we will try and verify the subkey signature using the
primary key algorithm and hit a BUG(). The fix is to not return a
hit if the keyid is not the primary. All other users of
get_pubkey_fast already expect a primary only.
2003-12-09 David Shaw <dshaw@jabberwocky.com>
* keyid.c (do_fingerprint_md): Remove the rules to hash the old v3
Elgamal keys. They are no longer needed.
* keyid.c (keyid_from_sk, keyid_from_pk, fingerprint_from_pk,
fingerprint_from_sk): Enforce the v3-is-only-RSA rule. Anything
that isn't RSA gets a zero keyid and fingerprint.
* keyid.c (do_fingerprint_md): Properly handle hashing of keys
that we don't know the structure of by using the opaque MPI.
(do_fingerprint_md_sk): We cannot calculate the fingerprint from a
secret key unless we know the structure (since we can't leave off
the secret key parts), so fail early.....
(keyid_from_sk, fingerprint_from_sk): .... and return all zeroes.
2003-12-03 David Shaw <dshaw@jabberwocky.com>
* g10.c (strusage, main): Show development version warning in
--version output.
(main): Set --bzip2-compress-level to the default value at
startup. Remove --emulate-checksum-bug noop.
* options.h, g10.c (main), main.h, seskey.c (do_encode_md,
encode_md_value), sig-check.c (do_check), sign.c (do_sign): Remove
--emulate-md-encode-bug as it only applied to Elgamal signatures,
which are going away.
2003-11-30 David Shaw <dshaw@jabberwocky.com>
* mainproc.c (proc_symkey_enc, proc_encrypted): Add ability to use
--override-session-key on --symmetric messages (new-style or
old-style).
(proc_pubkey_enc): Move code to show session key from here to
proc_encrypted() so it can work with any type of message.
Suggested by Michael Young.
2003-11-29 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (validate_keys): Reset the trustdb before checking if
we have any ultimately trusted keys. This ensures that if we lose
all our ultimately trusted keys, we don't leave behind the old
validity calculations. Noted by Peter Palfrader.
* revoke.c (gen_desig_revoke): Specify in the comment when a
designated revocation is generated.
* getkey.c (merge_selfsigs_main, merge_selfsigs_subkey,
get_seckey_byname2): Remove Elgamal check since we are removing
type 20 keys altogether.
2003-11-27 David Shaw <dshaw@jabberwocky.com>
* pkclist.c (build_pk_list): Do not allow an empty PK list in
interactive mode.
* keygen.c (ask_algo): Remove ability to generate Elgamal
sign+encrypt keys.
* getkey.c (merge_selfsigs_main, merge_selfsigs_subkey,
get_seckey_byname2): Disallow use of sign+encrypt Elgamal keys.
2003-11-20 David Shaw <dshaw@jabberwocky.com>
* seskey.c (do_encode_md): Comment about earlier (pre-PGP 2.3)
encodings.
* misc.c (compress_algo_to_string): Translate "Uncompressed".
Requested by Tommi Vainikainen.
(string_to_compress_algo): Include multi-string for
"uncompressed|none".
2003-11-17 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), compress-bz2.c (init_uncompress): Add
--bz2-compress-lowmem to set bzlib "small" flag for low memory
(but slow) decompression.
2003-11-15 David Shaw <dshaw@jabberwocky.com>
* compress.c (init_compress): Remove compress level 10 trick,
since it is no longer needed.
* g10.c: Fix typoed option name.
* compress-bz2.c (init_compress): Compression level 0 is not
meaningful for bzip2.
* options.h, g10.c (main), compress.c (init_compress),
compress-bz2.c (init_compress): Add --compress-level and
--bzip2-compress-level. -z sets them both. Change various
callers.
* encode.c (encode_simple), sign.c (sign_symencrypt_file):
Properly use default_compress_algo (--compress-algo, followed by
the highest --personal-compress-preference, followed by ZIP) to
get the algorithm.
2003-11-14 David Shaw <dshaw@jabberwocky.com>
* options.h, trustdb.c (trust_model_string, init_trustdb): Add
support for "external" trust model, where the user can provide a
pregenerated trustdb.
* keyedit.c (keyedit_menu): Do not allow editing ownertrust with
an external trust model trustdb.
2003-11-13 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c, keyedit.c, keylist.c, mainproc.c: Clarify the
plurarility (or not) of various list and verify options.
2003-11-12 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Add --symmetric --sign --encrypt.
* main.h, encode.c (setup_symkey): New. Prompt for a passphrase
and create a DEK for symmetric encryption.
(write_symkey_enc): New. Write out symmetrically encrypted
session keys.
(encode_crypt, encrypt_filter): Use them here here when creating a
message that can be decrypted with a passphrase or a pk.
* sign.c (sign_file): Call setup_symkey if we are doing a
--symmetric --sign --encrypt.
2003-11-09 David Shaw <dshaw@jabberwocky.com>
* mainproc.c (proc_symkey_enc): Don't show algorithm information
when --quiet is set. Suggested by Duncan Harris. Also don't fail
with BUG() when processing a --symmetric message with a cipher we
don't have.
* g10.c: Alias --personal-xxx-prefs to --personal-xxx-preferences.
* pkclist.c (build_pk_list): When adding recipients interactively,
allow the user to stop at any point.
2003-10-31 David Shaw <dshaw@jabberwocky.com>
* trustdb.h, trustdb.c (register_trusted_keyid): New. Adds a
keyid to the list of ultimately trusted keys.
* keygen.c (do_generate_keypair): Use it here so that the ultimate
ownertrust happens before the trustdb (might be) rebuilt. Also
fix an error where the newly generated pk is thought to be a
subkey by the trustdb.
* g10.c (main): Fix --export-all do actually do something
different than --export.
* pkclist.c (build_pk_list): Show all recipients rather than
showing each recipient as they are added.
* mainproc.c (proc_symkey_enc, proc_encrypted): Keep a count of
the number of passphrases that can decrypt a symmetric or mixed
symmetric/pk message and include it in the list of keys shown to
the user.
2003-10-30 David Shaw <dshaw@jabberwocky.com>
* misc.c (compress_algo_to_string, string_to_compress_algo,
check_compress_algo): Add bzip2.
* compress.c (compress_filter): Make static to help force the use
of push_compress_filter. Remove default algorithm setting since
that is done in push_compress_filter now.
* main.h: Use named algorithm.
* filter.h, compress.c (push_compress_filter,
push_compress_filter2): New. Figure out which is the appropriate
compression filter to use, and push it into place.
* compress.c (handle_compressed), encode.c (encode_simple,
encode_crypt), sign.c (sign_file, sign_symencrypt_file), import.c
(read_block), export.c (do_export): Use push_compress_filter
instead of pushing the compression filter ourselves.
* compress-bz2.c: New. Bzlib versions of the compression filter
routines.
* Makefile.am: Include compress-bz2.c if bz2lib is available.
2003-10-30 Werner Koch <wk@gnupg.org>
* apdu.c (close_ct_reader, close_pcsc_reader): Implemented.
(get_ccid_error_string): New. Not very useful messages, though.
2003-10-29 Werner Koch <wk@gnupg.org>
* cardglue.c (open_card): Ask for card insertion.
(check_card_serialno): New.
(agent_scd_pksign, agent_scd_pkdecrypt): Use it here.
* cardglue.c (open_card): Issue insertion status message.
* status.h, status.c (STATUS_CARDCTRL): New.
* status.c (cpr_get_answer_okay_cancel): New.
2003-10-28 Werner Koch <wk@gnupg.org>
* keylist.c (list_keyblock_print): Denote secrets keys stored on a
card with an '>'. Print the '#' also for subkeys.
(list_keyblock_colon): Introduce new field 15 for sec/ssb to print
the serial number.
2003-10-26 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Enhance the version-specific config file code to
try for more specific matches before giving up (e.g. 1.3.3-cvs,
1.3.3, 1.3, 1).
2003-10-25 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Add --symmetric --encrypt command. This generates
a message that can be decrypted via a passphrase or public key
system.
* main.h, encode.c (encode_seskey): Allow passing in an
already-created session key dek.
(encode_simple): Use the actual symmetric cipher when encrypting a
session key for a symmetric message.
(encode_crypt): Add a flag to trigger a hybrid mode that can be
decrypted via a passphrase or a pk. Change all callers.
* mainproc.c (symkey_decrypt_sesskey): There is no way to tell the
difference here between a bad passphrase and a cipher algorithm
that we don't have, so use a error message that makes that clear.
Use the actual list of ciphers when checking whether a cipher is
invalid. Return error if the decrypted cipher algorithm is
invalid.
(proc_symkey_enc): In a mixed passphrase/pk message, if a valid
dek already exists from decrypting via pk, do not try to process
the passphrase.
(proc_symkey_enc): Indicate when we're decrypting a session key as
opposed to decrypting data. If a passphrase is invalid, discard
the dek so we'll keep trying.
2003-10-25 Werner Koch <wk@gnupg.org>
* ccid-driver.c (ccid_open_reader): Return an error if no USB
devices are found.
* Makefile.am: Replaced INTLLIBS by LIBINTL.
* g10.c (main) [ENABLE_CARD_SUPPORT]: Add a default for
--pcsc-driver.
* cardglue.c (learn_status_cb): Fixed faulty use of !space.
2003-10-24 Werner Koch <wk@gnupg.org>
* apdu.c (apdu_open_reader): Hacks for PC/SC under Windows.
2003-10-21 Werner Koch <wk@gnupg.org>
* passphrase.c (ask_passphrase): Add optional promptid arg.
Changed all callers.
* cardglue.c (pin_cb): Use it here, so the machine interface can
tell whether the Admin PIN is requested.
* cardglue.c (agent_scd_checkpin): New.
* misc.c (openpgp_pk_algo_usage): Added AUTH usage.
* app-openpgp.c (check_against_given_fingerprint): New. Factored
out that code elsewhere.
(do_check_pin): New.
* card-util.c (card_edit): New command "passwd". Add logic to
check the PIN in advance.
(card_status): Add new args to return the serial number. Changed
all callers.
2003-10-14 David Shaw <dshaw@jabberwocky.com>
* import.c (import_one): Show the keyid when giving the Elgamal
slow import warning.
* g10.c (main): Older versions used --comment "" to indicate no
comment. Don't add an empty comment.
2003-10-13 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (show_key_with_all_names): Ownertrust is only
meaningful for the PGP or classic trust models. Both validity and
ownertrust are not meaningful for the always trust model.
2003-10-11 Werner Koch <wk@gnupg.org>
* keygen.c: Always enable the gen_card_key prototype.
2003-10-10 Werner Koch <wk@gnupg.org>
* cardglue.c (card_close): New.
(agent_scd_change_pin): Implemented.
* ccid-driver.c (ccid_close_reader): New.
* apdu.c (close_ccid_reader, close_ct_reader, close_csc_reader)
(close_osc_reader, apdu_close_reader): New. Not all are properly
implemented yet.
* g10.c (g10_exit): Use close_card.
2003-10-09 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Give a deprecated option warning for
--show-keyring, --show-photos, --show-policy-url, --show-notation,
and their respective no- forms.
* options.skel: Remove show-photos and replace with
list/verify-options show-photos. Remove no-mangle-dos-filenames.
* misc.c (parse_options): Allow for incomplete (but unambiguous)
options.
2003-10-09 Werner Koch <wk@gnupg.org>
* ccid-driver.c (ccid_transceive): Add T=1 chaining for sending.
* sign.c (do_sign) [!ENABLE_CARD_SUPPORT]: Return an error for
card keys.
* cardglue.c (agent_scd_pkdecrypt): Implemented.
* pubkey-enc.c (get_it) [ENABLE_CARD_SUPPORT]: Divert decryption
to card
2003-10-08 Werner Koch <wk@gnupg.org>
* cardglue.c (pin_cb): Detect whether an admin or regular PIN is
requested.
(genkey_status_cb): New.
(agent_scd_genkey): Implemented.
* keygen.c (generate_keypair): New arg CARD_SERIALNO and prepare
parameters for on card key generation. Changed all callers.
(do_generate_keypair): Add new arg card and merged casrd specific
changes from 1.9.
(proc_parameter_file): New arg card, apss it down to
do_generate_keypair and changed all callers.
(gen_card_key): New.
* g10.c: Include cardclue.h.
(main): s/app_set_default_reader_port/card_set_reader_port/.
* cardglue.c (card_set_reader_port): New to address include file
issues.
2003-10-02 Werner Koch <wk@gnupg.org>
* cardglue.c (learn_status_cb): Release values before assignment
so that it can be used by getattr to update the structure.
(agent_scd_getattr): New.
* keylist.c (print_pubkey_info): Add FP arg for optional printing
to a stream. Changed all callers.
2003-10-01 Werner Koch <wk@gnupg.org>
* app-common.h, app-openpgp.c, iso7816.c, iso7816.h, apdu.c
* apdu.h, ccid-driver.c, ccid-driver.h, card-util.c: Updated
from current GnuPG 1.9. Changes over there are:
* card-util.c: Tweaked to use this source also under 1.3.
(card_edit): New command "forcesig".
* card-util.c (print_name, print_isoname): Use 0 and not LF fro
the max_n arg of tty_print_utf8_string2.
* card-util.c (change_pin): Simplified. We now have only a PIN
and an Admin PIN.
* ccid-driver.c: Detect GnuPG 1.3 and include appropriate files.
* apdu.c: Ditto.
* app-openpgp.c: Ditto.
* iso7816.c: Ditto.
(generate_keypair): Renamed to ..
(do_generate_keypair): .. this.
* app-common.h [GNUPG_MAJOR_VERSION]: New.
* iso7816.h [GNUPG_MAJOR_VERSION]: Include cardglue.h
* app-openpgp.c (do_change_pin): Make sure CVH1 and CHV2 are
always synced.
(verify_chv2, verify_chv3): New. Factored out common code.
(do_setattr, do_sign, do_auth, do_decipher): Change the names of
the prompts to match that we have only 2 different PINs.
(app_select_openpgp): Check whether the card enforced CHV1.
(convert_sig_counter_value): New. Factor out code from
get_sig_counter.
* Makefile.am (card_support_source): Depend on new AM conditional
to get the ifdef ENABLE_CARD_SUPPORT off the way from source
copied files.
(update-source-from-gnupg-2): Maintainer helper.
2003-10-01 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Add --no-groups to zero --group list.
* encode.c (encode_simple): Allow for 32 bytes (256 bits) of
symmetrically encrypted session key. Use --s2k-cipher-algo to
choose cipher, rather than the default cipher.
* parse-packet.c (parse_subkeyenc): Give a warning if an
symmetrically encrypted session key is seen without salt. Show in
--list-packets if a symetrically encrypted session key is present.
* pubkey-enc.c (get_it): Always show cipher-not-in-prefs warning
unless --quiet is set. Use text name of cipher in warning.
2003-09-30 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), mainproc.c (check_sig_and_print): Add
--verify-option show-unusable-uids.
* gpgv.c (check_trustdb_stale): Stub.
* trustdb.c (get_validity): Move the up-to-date check to
check_trustdb_stale (new), so that it can be called before
validity is checked.
* keylist.c (list_keyblock_print): Disable the overall key
validity display until it can be thought about more. Use
check_trustdb_stale here to avoid putting the check warning in the
middle of a listed key.
* trustdb.c (init_trustdb): Only verify_own_keys() for those trust
models that it applies to (i.e. classic and OpenPGP).
2003-09-29 Werner Koch <wk@gnupg.org>
* keygen.c (do_add_key_flags, parse_parameter_usage): Add support
the proposed AUTH key flag.
* getkey.c (fixup_uidnode, merge_selfsigs_main)
(merge_selfsigs_subkey, premerge_public_with_secret): Ditto.
* keylist.c (print_capabilities): Ditto.
* parse-packet.c (parse_key): Allow to parse the divert-to-card
S2K mode.
* build-packet.c (do_secret_key): Handle divert-to-card S2K
* seckey-cert.c (is_secret_key_protected): Ditto.
(check_secret_key): Ditto.
* keygen.c (do_ask_passphrase): Renamed from ask_passphrase.
* passphrase.c (ask_passphrase): New.
2003-09-28 Werner Koch <wk@gnupg.org>
* g10.c (main): New commands --card-edit, --card-status and
--change-pin. New options --ctapi-driver, --pcsc-driver and
--disable-ccid
* options.h (DBG_CARD_IO): New.
* cardglue.c, cardclue.h: Enhanced.
* card-util.c: New. Taken from current the gnupg 1.9 branch.
* app-common.h, app-openpgp.c, iso7816.c, iso7816.h, apdu.c
* apdu.h, ccid-driver.c, ccid-driver.h: New. Takem from the current
gnupg 1.9 branch withy minor changes to include directives.
* Makefile.am: Added these files.
2003-09-27 Werner Koch <wk@gnupg.org>
* sign.c (do_sign) [ENABLE_CARD_SUPPORT]: Divert to card.
* cardglue.c, cardglue.h: New.
* Makefile.am (gpg_LDADD): Added.
(card_support_sources): New.
2003-09-25 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), keylist.c (list_keyblock_print): Add
"show-unusable-uids" list-option to show revoked and/or expired
user IDs.
2003-09-24 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (show_key_with_all_names): Show names a little neater
by putting the [revoked] or [expired] in the space used for the
[validity]. There is also no point in showing "[unknown]
[revoked]".
2003-09-23 David Shaw <dshaw@jabberwocky.com>
* sign.c (mk_notation_policy_etc): Capitalize "URL".
* trustdb.c (validate_keys): Give a little more information while
rebuilding trustdb.
* pkclist.c (do_edit_ownertrust): Clarify "don't know".
* g10.c (main): Default to --no-mangle-dos-filenames.
* keydb.h, keyring.c (keyring_search), trustdb.c (search_skipfnc):
Expand the skipfnc to include a pointer to the user ID that
matched.
* getkey.c (skip_disabled): Rename to skip_unusable, and add
checks for expired or revoked user IDs.
2003-09-22 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Deprecate --default-comment in favor of
--no-comments.
* options.h, g10.c (main), armor.c (armor_filter): Allow using
--comment multiple times to get multiple Comment: header lines.
--no-comments resets list.
2003-09-11 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Trim --help to commonly used options. Remove -f.
2003-09-08 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Error out if --multifile is used with the commands
that don't support it yet (--sign, --clearsign, --detach-sign,
--symmetric, and --store).
* g10.c (main): Add --multifile as an alias to turn --encrypt into
--encrypt-files (plus --verify-files, --decrypt-files).
* encode.c (use_mdc), g10.c (main): Use RFC1991 and RFC2440
directly to check for MDC usability. Do not set the force_mdc or
disable_mdc flags since there is no point any longer.
2003-09-04 David Shaw <dshaw@jabberwocky.com>
* armor.c (parse_hash_header, armor_filter), g10.c (print_hex,
print_mds), pkclist.c (algo_available): Drop TIGER/192 support.
2003-09-03 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (show_key_with_all_names): Fix assertion failure when
using toggle to see a secret key. Reported by Maxim Britov.
2003-08-31 David Shaw <dshaw@jabberwocky.com>
* g10.c (add_keyserver_url), keyedit.c (keyedit_menu), sign.c
(mk_notation_policy_etc): Clarify a few strings. It's a
"preferred keyserver URL".
* g10.c (main): Use "keyserver-url" instead of
"preferred-keyserver" for the sake of short and simple commands.
2003-08-30 David Shaw <dshaw@jabberwocky.com>
* main.h, keygen.c (keygen_add_keyserver_url): Signature callback
for adding a keyserver URL.
* keyedit.c (keyedit_menu, menu_set_keyserver_url): New command to
set preferred keyserver to specified (or all) user IDs.
* build-packet.c (build_sig_subpkt): Set preferred keyserver flag
while building a preferred keyserver subpacket.
* keylist.c (show_policy_url, show_keyserver_url): URLs might be
UTF8.
* keyedit.c (menu_addrevoker): Fix leaking a few bytes.
2003-08-29 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (show_key_with_all_names): Use list-option
show-long-keyid in main --edit-key display.
* keyedit.c (print_and_check_one_sig): Use list-option
show-long-keyid in --edit-key "check" function.
2003-08-28 David Shaw <dshaw@jabberwocky.com>
* passphrase.c (agent_send_all_options): Make use of $GPG_TTY.
* g10.c (main): Disable use-agent if passphrase-fd is given
later. Suggested by Kurt Garloff.
* exec.c, g10.c, gpgv.c, passphrase.c, photoid.c:
s/__MINGW32__/_WIN32/ to help building on native Windows
compilers. Requested by Brian Gladman. From Werner on stable
branch.
2003-08-25 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main): Add list-option
list-preferred-keyserver.
* keyedit.c (change_passphrase): When responding 'no' to the blank
passphrase question, re-prompt for a new passphrase. This is bug
#202.
* mainproc.c (check_sig_and_print): Use two different preferred
keyserver displays - one if the key is not present (to tell the
user where to get the key), the other if it is present (to tell
the user where the key can be refreshed).
* packet.h, parse-packet.c (parse_signature): Set flag if a
preferred keyserver is present.
* keylist.c (list_keyblock_print): Show keyserver url in listings
with list-option show-keyserver-url.
2003-08-24 David Shaw <dshaw@jabberwocky.com>
* Makefile.am: Use NETLIBS instead of EGDLIBS.
* mainproc.c (check_sig_and_print): Get the uid validity before
printing any sig results to avoid munging the output with trustdb
warnings.
* g10.c (main): Don't include --show-keyring in --help as it is
deprecated.
2003-08-21 David Shaw <dshaw@jabberwocky.com>
* gpgv.c: Remove extra semicolon (typo).
* options.skel: Note that keyserver.pgp.com isn't synchronized,
and explain the roundrobin a bit better.
* sig-check.c (check_key_signature2), import.c (import_one,
import_revoke_cert, chk_self_sigs, delete_inv_parts,
collapse_uids, merge_blocks): Make much quieter during import of
slightly munged, but recoverable, keys. Use log_error for
unrecoverable import failures.
* keyring.c (keyring_rebuild_cache): Comment.
* sign.c (mk_notation_and_policy): Making a v3 signature with
notations or policy urls is an error, not an info (i.e. increment
the errorcount). Don't print the notation or policy url to stdout
since it can be mixed into the output stream when piping and munge
the stream.
2003-08-12 David Shaw <dshaw@jabberwocky.com>
* packet.h, sig-check.c (signature_check2, do_check,
do_check_messages): Provide a signing-key-is-revoked flag. Change
all callers.
* status.h, status.c (get_status_string): New REVKEYSIG status tag
for a good signature from a revoked key.
* mainproc.c (do_check_sig, check_sig_and_print): Use it here.
* import.c (import_revoke_cert, merge_blocks, merge_sigs): Compare
actual signatures on import rather than using keyid or class
matching. This does not change actual behavior with a key, but
does mean that all sigs are imported whether they will be used or
not.
* parse-packet.c (parse_signature): Don't give "signature packet
without xxxx" warnings for experimental pk algorithms. An
experimental algorithm may not have a notion of (for example) a
keyid (i.e. PGP's x.509 stuff).
2003-08-02 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), keylist.c (list_keyblock_print),
keyedit.c (print_and_check_one_sig): New "show-sig-expire"
list-option to show signature expiration dates (if any).
2003-07-24 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main, add_keyserver_url): Add
--sig-preferred-keyserver to implant a "where to get my key"
subpacket into a signature.
* sign.c (mk_notation_and_policy): Rename to
mk_notation_policy_etc and add preferred keyserver support for
signatures.
2003-07-21 David Shaw <dshaw@jabberwocky.com>
* keygen.c (do_add_key_flags): Don't set the certify flag for
subkeys.
(ask_algo): Provide key flags for DSA, Elgamal_e, and Elgamal
subkeys.
(generate_keypair): Provide key flags for the default DSA/Elgamal
keys.
* sig-check.c (signature_check, signature_check2,
check_key_signature, check_key_signature2): Allow passing NULLs
for unused parameters in the x2 form of each function to avoid the
need for dummy variables. getkey.c, mainproc.c: Change all
callers.
* trustdb.h, trustdb.c (read_trust_options): New. Returns items
from the trustdb version record.
* keylist.c (public_key_list): Use it here for the new "tru"
record.
* gpgv.c (read_trust_options): Stub.
2003-07-20 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (show_key_with_all_names): Use list-option
show-validity in --edit-key interface as well.
2003-07-19 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), mainproc.c (check_sig_and_print): Add
verify-options "show-validity" and "show-long-keyid" to show
trustdb validity and long keyids during (file) signature
verification.
* packet.h, main.h, sig-check.c (signature_check2,
check_key_signature2, do_check): If ret_pk is set, fill in the pk
used to verify the signature. Change all callers in getkey.c,
mainproc.c, and sig-check.c.
* keylist.c (list_keyblock_colon): Use the ret_pk from above to
put the fingerprint of the signing key in "sig" records during a
--with-colons --check-sigs. This requires --no-sig-cache as well
since we don't cache fingerprints.
2003-07-10 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (parse_signature): No need to reserve 8 bytes for
the unhashed signature cache any longer.
* misc.c (pct_expando): Add two new expandos - signer's
fingerprint (%g), and signer's primary fingerprint (%p).
* Makefile.am: Include W32LIBS where appropriate.
* g10.c (main): Add --rfc2440 alias for --openpgp since in a few
months, they won't be the same thing.
* keyserver.c (parse_keyserver_uri): Accept "http" as an alias for
"hkp", since it is occasionally written that way.
(keyserver_spawn): Use ascii_isspace to avoid locale issues.
* keygen.c (ask_user_id): Make --allow-freeform-uid apply to the
email field as well as the name field, and allow mixing fields
when it is set.
* options.skel: Use subkeys.pgp.net as the default keyserver.
* trustdb.c (validate_one_keyblock): Certifications on revoked or
expired uids do not count in the web of trust.
* signal.c (init_one_signal, pause_on_sigusr, do_block): Only use
sigprocmask() if we have sigset_t, and only use sigaction() if we
have struct sigaction. This is for Forte c89 on Solaris which
seems to define only the function call half of the two pairs by
default.
(pause_on_sigusr): Typo.
(do_block): If we can't use sigprocmask() and sigset_t, try to get
the number of signals from NSIG as well as MAXSIG, and if we
can't, fail with an explanation.
* signal.c, tdbio.c: Comment out the transaction code. It was not
used in this version, and was causing some build problems on
quasi-posix platforms (Solaris and Forte c89).
* keylist.c (list_keyblock_colon): Don't include validity values
when listing secret keys since they can be incorrect and/or
misleading. This is a temporary kludge, and will be handled
properly in 1.9/2.0.
* mainproc.c (check_sig_and_print): Only show the "key available
from" preferred keyserver line if the key is not currently
present.
* keyedit.c (sign_uids): Do not sign expired uids without --expert
(same behavior as revoked uids). Do not allow signing a user ID
without a self-signature. --expert overrides. Add additional
prompt to the signature level question.
(menu_expire): When changing expiration dates, don't replace
selfsigs on revoked uids since this would effectively unrevoke
them. There is also no point in replacing expired selfsigs. This
is bug #181
2003-07-10 David Shaw <dshaw@jabberwocky.com> (from Werner on stable branch)
* g10.c (add_notation_data): Make sure that only ascii is passed
to iscntrl. Noted by Christian Biere.
* getkey.c (classify_user_id2): Replaced isspace by spacep
* keygen.c (ask_user_id): Ditto.
(get_parameter_algo): Ditto.
* keyedit.c (keyedit_menu): Ditto.
* tdbdump.c (import_ownertrust): Ditto. s/isxdigit/hexdigitp/.
* revoke.c (ask_revocation_reason):
* keyserver.c (keyserver_spawn): Dito.
2003-06-10 Werner Koch <wk@gnupg.org>
* parse-packet.c (parse): Disallow old style partial length for
all key material packets to avoid possible corruption of keyrings.
2003-06-08 Werner Koch <wk@gnupg.org>
* import.c (import_keys_internal): Invalidate the cache so that
the file descriptor gets closed. Fixes bug reported by Juan
F. Codagnone.
2003-06-04 David Shaw <dshaw@jabberwocky.com>
* options.skel: Use new hkp://subkeys.pgp.net as sample keyserver
since they at least handle subkeys correctly.
* options.h, g10.c (main), main.h, keylist.c (show_keyserver_url),
mainproc.c (check_sig_and_print), parse-packet.c (dump_sig_subpkt,
parse_one_sig_subpkt, can_handle_critical): Add read-only support
for preferred keyserver subpackets. They're basically policy URLs
with a different name. Add a verify-option
"show-preferred-keyserver" to turn them on and off (on by default,
as per stable branch).
* g10.c (main): Add "--set-notation" as alias to "--notation-data"
this is to make things consistent with --set-policy-url meaning
both sigs and certs.
2003-06-03 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), keylist.c (list_keyblock_print): Add
"show-validity" and "show-long-keyid" list-options.
* gpgv.c (get_validity, trust_value_to_string): Stubs.
* g10.c (main): Use SAFE_VERSION instead of VERSION in the
version-specific gpg.conf file so it can be overridden on RISCOS.
2003-06-01 David Shaw <dshaw@jabberwocky.com>
* g10.c (main), keylist.c (show_policy_url, show_notation),
mainproc.c (check_sig_and_print): Emulate the old policy and
notation behavior (display by default). Send to status-fd whether
it is displayed on the screen or not.
* g10.c (main): Since we now have some options in devel that won't
work in a stable branch gpg.conf file, try for a version-specific
gpg.conf-VERSION file before falling back to gpg.conf.
* main.h, options.h: Move various option flags to options.h.
2003-05-31 David Shaw <dshaw@jabberwocky.com>
* mainproc.c (check_sig_and_print), main.h, keylist.c
(show_policy, show_notation): Collapse the old print_notation_data
into show_policy() and show_notation() so there is only one
function to print notations and policy URLs.
* options.h, main.h, g10.c (main), keyedit.c
(print_and_check_one_sig), keylist.c (list_one,
list_keyblock_print), pkclist.c (do_edit_ownertrust), sign.c
(mk_notation_and_policy): New "list-options" and "verify-options"
commands. These replace the existing
--show-photos/--no-show-photos,
--show-notation/--no-show-notation,
--show-policy-url/--no-show-policy-url, and --show-keyring
options. The new method is more flexible since a user can specify
(for example) showing photos during sig verification, but not in
key listings. The old options are emulated.
* main.h, misc.c (parse_options): New general option line
parser. Fix the bug in the old version that did not handle report
syntax errors after a valid entry.
* import.c (parse_import_options), export.c
(parse_export_options): Call it here instead of duplicating the
code.
2003-05-30 David Shaw <dshaw@jabberwocky.com>
* keylist.c (list_one): Don't show the keyring filename when in
--with-colons mode. Actually translate "Keyring" string.
* mainproc.c (proc_tree): We can't currently handle multiple
signatures of different classes or digests (we'd pretty much have
to run a different hash context for each), but if they are all the
same, make an exception. This is Debian bug #194292.
* sig-check.c (check_key_signature2): Make string translatable.
* packet.h, getkey.c (fixup_uidnode): Mark real primary uids
differently than assumed primaries.
* keyedit.c (no_primary_warning): Use the differently marked
primaries here in a new function to warn when an --edit-key
command might rearrange the self-sig dates enough to change which
uid is primary.
(menu_expire, menu_set_preferences): Use no_primary_warning()
here.
* Makefile.am: Use @DLLIBS@ for -ldl.
2003-05-26 David Shaw <dshaw@jabberwocky.com>
* getkey.c (premerge_public_with_secret): Made "no secret subkey
for" warning a verbose item and translatable. (From wk on stable
branch)
* sig-check.c (check_key_signature2): Made "no subkey for subkey
binding packet" a verbose item instead of a !quiet one. There are
too many garbled keys out in the wild. (From wk on stable branch)
* filter.h: Remove const from WHAT. (From wk on stable branch)
* progress.c (handle_progress): Store a copy of
NAME. (progress_filter): Release WHAT, make sure not to print a
NULL WHAT. (From wk on stable branch)
* openfile.c (open_sigfile): Adjust free for new progress
semantics. (From wk on stable branch)
* plaintext.c (ask_for_detached_datafile): Don't dealloc
pfx->WHAT. (From wk on stable branch)
* seckey-cert.c (do_check): Issue the RSA_OR_IDEA status when the
cipher algo is IDEA to make it easier to track down the
problem. (From twoaday on stable branch)
2003-05-24 David Shaw <dshaw@jabberwocky.com>
* armor.c, g10.c, kbnode.c, misc.c, pkclist.c, sign.c,
build-packet.c, getkey.c, keydb.c, openfile.c, plaintext.c,
status.c, gpgv.c, keygen.c, options.h, sig-check.c, tdbio.h,
encode.c, mainproc.c, parse-packet.c, signal.c, textfilter.c: Edit
all preprocessor instructions to remove whitespace before the '#'.
This is not required by C89, but there are some compilers out
there that don't like it.
2003-05-21 David Shaw <dshaw@jabberwocky.com>
* trustdb.h, trustdb.c (is_disabled), gpgv.c (is_disabled): Rename
is_disabled to cache_disabled_value, which now takes a pk and not
just the keyid. This is for speed since there is no need to
re-fetch a key when we already have that key handy. Cache the
result of the check so we don't need to hit the trustdb more than
once.
* getkey.c (skip_disabled): New function to get a pk and call
is_disabled on it. (key_byname): Use it here.
* packet.h, getkey.c (skip_disabled), keylist.c
(print_capabilities): New "pk_is_disabled" macro to retrieve the
cached disabled value if available, and fill it in via
cache_disabled_value if not available.
* trustdb.c (get_validity): Cache the disabled value since we have
it handy and it might be useful later.
* parse-packet.c (parse_key): Clear disabled flag when parsing a
new key. Just in case someone forgets to clear the whole key.
* getkey.c (merge_selfsigs_main): Add an "if all else fails" path
for setting a single user ID primary when there are multiple set
primaries all at the same second, or no primaries set and the most
recent user IDs are at the same second, or no signed user IDs at
all. This is arbitrary, but deterministic.
* exec.h, photoid.h: Add copyright message.
* keylist.c (list_keyblock_print): Don't dump attribs for
revoked/expired/etc uids for non-colon key listings. This is for
consistency with --show-photos.
* main.h, keylist.c (dump_attribs), mainproc.c
(check_sig_and_print): Dump attribs if --attrib-fd is set when
verifying signatures.
* g10.c (main): New --gnupg option to disable the various
--openpgp, --pgpX, etc. options. This is the same as --no-XXXX
for those options.
* revoke.c (ask_revocation_reason): Clear old reason if user
elects to repeat question. This is bug 153.
* keyedit.c (sign_uids): Show keyid of the key making the
signature.
2003-05-21 Werner Koch <wk@gnupg.org>
* progress.c (handle_progress)
* sign.c (write_plaintext_packet)
* encode.c (encode_simple,encode_crypt): Make sure that a filename
of "-" is considered to be stdin so that iobuf_get_filelength
won't get called. This fixes bug 156 reported by Gregery Barton.
2003-05-02 David Shaw <dshaw@jabberwocky.com>
* packet.h, build-packet.c (build_sig_subpkt), export.c
(do_export_stream), import.c (remove_bad_stuff, import),
parse-packet.c (dump_sig_subpkt, parse_one_sig_subpkt): Remove
vestigal code for the old sig cache subpacket. This wasn't
completely harmless as it caused subpacket 101 to disappear on
import and export.
* options.h, armor.c, cipher.c, g10.c, keyedit.c, pkclist.c,
sign.c, encode.c, getkey.c, revoke.c: The current flags for
different levels of PGP-ness are massively complex. This is step
one in simplifying them. No functional change yet, just use a
macro to check for compliance level.
* sign.c (sign_file): Fix bug that causes spurious compression
preference warning.
* sign.c (clearsign_file): Fix bug that prevents proper warning
message from appearing when clearsigning in --pgp2 mode with a
non-v3 RSA key.
* main.h, misc.c (compliance_option_string, compliance_string,
compliance_failure), pkclist.c (build_pk_list), sign.c (sign_file,
clearsign_file), encode.c (encode_crypt,
write_pubkey_enc_from_list): New functions to put the "this
message may not be usable...." warning in one place.
* options.h, g10.c (main): Part two of the simplification. Use a
single enum to indicate what we are compliant to (1991, 2440,
PGPx, etc.)
* g10.c (main): Show errors for failure in export, send-keys,
recv-keys, and refresh-keys.
* options.h, g10.c (main): Give algorithm warnings for algorithms
chosen against the --pgpX and --openpgp rules.
* keydb.h, pkclist.c (algo_available): Make TIGER192 invalid in
--openpgp mode.
* sign.c (sign_file), pkclist.c (algo_available): Allow passing a
hint of 0.
2003-05-01 David Shaw <dshaw@jabberwocky.com>
* tdbio.c (create_version_record): Only create new trustdbs with
TM_CLASSIC or TM_PGP.
* trustdb.h, trustdb.c (trust_string, get_ownertrust_string,
get_validity_string, ask_ownertrust, validate_keys), pkclist.c
(do_edit_ownertrust): Rename trust_string to trust_value_to_string
for naming consistency.
* trustdb.h, trustdb.c (string_to_trust_value): New function to
translate a string to a trust value.
* g10.c (main): Use string_to_trust_value here for
--force-ownertrust.
* options.h, g10.c (main), trustdb.c (trust_model_string,
init_trustdb, check_trustdb, update_trustdb, get_validity,
validate_one_keyblock): An "OpenPGP" trust model is misleading
since there is no official OpenPGP trust model. Use "PGP"
instead.
2003-04-30 David Shaw <dshaw@jabberwocky.com>
* build-packet.c (build_sig_subpkt): Comments.
* exec.c (exec_write): Cast NULL to void* to properly terminate
varargs list.
* keyedit.c (show_key_with_all_names): Just for safety, catch an
invalid pk algorithm.
* sign.c (make_keysig_packet): Crucial that the call to mksubpkt
comes LAST before the calls to finalize the sig as that makes it
possible for the mksubpkt function to get a reliable pointer to
the subpacket area.
* pkclist.c (do_we_trust_pre): If an untrusted key was chosen by a
particular user ID, use that ID as the one to ask about when
prompting whether to use the key anyway.
(build_pk_list): Similar change here when adding keys to the
recipient list.
* trustdb.c (update_validity): Fix bug that prevented more than
one validity record per trust record.
(get_validity): When retrieving validity for a (user) supplied
user ID, return the validity for that user ID only, and do not
fall back to the general key validity.
(validate_one_keyblock): Some commentary on whether
non-self-signed user IDs belong in the web of trust (arguably,
they do).
2003-04-27 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Add --no-textmode.
* export.c (do_export_stream), keyedit.c (show_key_with_all_names,
menu_addrevoker), mainproc.c (check_sig_and_print), photoid.c
(show_photos), sign.c (mk_notation_and_policy), trustdb.c
(get_validity, reset_trust_records, validate_keys): Make some
strings translatable.
* mainproc.c (check_sig_and_print): Show digest algorithm and sig
class when verifying a sig with --verbose on, and add version, pk
and hash algorithms and sig class to VALIDSIG.
* parse-packet.c (enum_sig_subpkt): Make a warning message a
--verbose warning message since we don't need to warn every time
we see an unknown critical (we only need to invalidate the
signature).
* trustdb.c (init_trustdb): Check the trustdb options even with
TM_AUTO since the auto may become TM_CLASSIC or TM_OPENPGP.
2003-04-26 David Shaw <dshaw@jabberwocky.com>
* sign.c (do_sign): Show the hash used when making a signature in
verbose mode.
* tdbio.h, tdbio.c (tdbio_read_model): New function to return the
trust model used in a given trustdb.
* options.h, g10.c (main), trustdb.c (init_trustdb, check_trustdb,
update_trustdb): Use tdbio_read_model to implement an "auto" trust
model which is set via the trustdb.
2003-04-23 David Shaw <dshaw@jabberwocky.com>
* import.c (import_revoke_cert): Remove ultimate trust when
revoking an ultimately trusted key.
* keyedit.c (sign_uids): Allow replacing expired signatures.
Allow duplicate signatures with --expert.
* pkclist.c (check_signatures_trust): Don't display a null
fingerprint when checking a signature with --always-trust enabled.
* filter.h (progress_filter_context_t), progress.c
(handle_progress), plaintext.c (ask_for_detached_datafile,
hash_datafiles): Fix compiler warnings. Make "what" constant.
* build-packet.c (do_plaintext): Do not create invalid literal
packets with >255-byte names.
2003-04-15 Werner Koch <wk@gnupg.org>
* Makefile.am (AM_CFLAGS): Make use of AM_CFLAGS and AM_LDFLAGS.
* g10.c, options.h: New option --enable-progress-filter.
* progress.c (handle_progress): Make use of it.
2003-04-15 Marcus Brinkmann <marcus@g10code.de>
* progress.c: New file.
* Makefile.am (common_source): Add progress.c.
* filter.h (progress_filter_context_t): New type.
(progress_filter, handle_progress): New prototypes.
* main.h (open_sigfile): New argument for prototype.
* openfile.c (open_sigfile): New argument to install progress
filter.
* encode.c (encode_simple): New variable PFX. Register
progress filter. Install text_filter after that.
(encode_crypt): Likewise.
* sign.c (sign_file): Likewise.
(clearsign_file): Likewise.
* decrypt.c (decrypt_message): Likewise.
(decrypt_messages): Likewise.
* verify.c (verify_signatures): Likewise.
(verify_one_file): Likewise.
* plaintext.c (hash_datafiles): Likewise.
(ask_for_detached_datafile): Likewise.
2003-04-10 Werner Koch <wk@gnupg.org>
* passphrase.c (read_passphrase_from_fd): Do a dummy read if the
agent is to be used. Noted by Ingo Klöcker.
(agent_get_passphrase): Inhibit caching when we have no
fingerprint. This is required for key generation as well as for
symmetric only encryption.
* passphrase .c (agent_get_passphrase): New arg CANCELED.
(passphrase_to_dek): Ditto. Passed to above. Changed all
callers to pass NULL.
* seckey-cert.c (do_check): New arg CANCELED.
(check_secret_key): Terminate loop when canceled.
* keyedit.c (change_passphrase): Pass ERRTEXT untranslated to
passphrase_to_dek and translate where appropriate.
* seckey-cert.c (check_secret_key): Ditto.
* keygen.c (ask_passphrase): Ditto.
* passphrase.c (agent_get_passphrase): Translate the TRYAGAIN_TEXT.
Switch the codeset to utf-8.
2003-04-09 Werner Koch <wk@gnupg.org>
* decrypt.c (decrypt_messages): Fixed error handling; the function
used to re-loop with same file after an error. Reported by Joseph
Walton.
2003-04-08 David Shaw <dshaw@jabberwocky.com>
* main.h, g10.c (main), import.c (parse_import_options,
fix_pks_corruption): It's really PKS corruption, not HKP
corruption. Keep the old repair-hkp-subkey-bug command as an
alias.
* g10.c (main): Rename --no-version to --no-emit-version for
consistency. Keep --no-version as an alias.
2003-04-04 David Shaw <dshaw@jabberwocky.com>
* pkclist.c (algo_available): PGP 8 can use the SHA-256 hash.
* sign.c (sign_file, clearsign_file, sign_symencrypt_file): Remove
unused code.
2003-04-01 Werner Koch <wk@gnupg.org>
* mainproc.c (check_sig_and_print): Add primary key fpr to VALIDSIG
status.
2003-03-24 David Shaw <dshaw@jabberwocky.com>
* keydb.h: Err on the side of making an unknown signature a SIG
rather than a CERT.
* import.c (delete_inv_parts): Discard any key signatures that
aren't key types (i.e. 0x00, 0x01, etc.)
* g10.c (main): Add deprecated option warning for
--list-ownertrust. Add --compression-algo alias for
--compress-algo. Change --version output strings to match
"showpref" strings, and make translatable.
* status.c (do_get_from_fd): Accept 'y' as well as 'Y' for
--command-fd boolean input.
* trustdb.c: Fix typo (DISABLE_REGEXP -> DISABLE_REGEX)
* keyedit.c (show_key_with_all_names_colon): Show no-ks-modify
flag.
2003-03-11 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), keyserver.c (kopts): Add "try-dns-srv"
keyserver option. Defaults to on.
* passphrase.c (agent_get_passphrase): Fix memory leak with
symmetric messages. Fix segfault with symmetric messages. Fix
incorrect prompt with symmetric messages.
2003-03-10 Werner Koch <wk@gnupg.org>
* compress.c (init_uncompress): Use a 15 bit window size so that
the output of implementations which don't run for PGP 2
compatibility won't get garbled.
2003-03-04 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (validate_keys): Mask the ownertrust when building the
list of fully valid keys so that disabled keys are still counted
in the web of trust.
(get_ownertrust_with_min): Do the same for the minimum ownertrust
calculation.
* parse-packet.c (dump_sig_subpkt): Show the notation names for
not-human-readable notations. Fix cosmetic off-by-one length
counter.
* options.skel: Add explantion and commented-out
"no-mangle-dos-filenames".
* mainproc.c (proc_encrypted): Make string translatable.
* keyserver.c (keyserver_spawn): Quote ':', '%', and any 8-bit
characters in the uid strings sent to the keyserver helper.
* keyring.c (keyring_rebuild_cache): Lock the keyring while
rebuilding the signature caches to prevent another gpg from
tampering with the temporary copy.
* keygen.c (keygen_set_std_prefs): Include AES192 and AES256 in
default prefs.
* keyedit.c (show_prefs): Make strings translatable.
* keydb.c: Double the maximum number of keyrings to 40.
* gpgv.c (main): Fix bug #113 - gpgv should accept the
--ignore-time-conflict option.
* g10.c (main): --openpgp disables --pgpX. Double the amount of
secure memory to 32k (keys are getting bigger these days).
* Makefile.am: Makefile.am: Use @CAPLIBS@ to link in -lcap if we
are using capabilities.
2003-02-26 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_spawn): Include various pieces of
information about the key in the data sent to the keyserver
helper. This allows the helper to use it in instructing a remote
server which may not have any actual OpenPGP smarts in parsing
keys.
* main.h, export.c (export_pubkeys_stream, do_export_stream): Add
ability to return only the first match in an exported keyblock for
keyserver usage. This should be replaced at some point with a
more flexible solution where each key can be armored seperately.
2003-02-22 David Shaw <dshaw@jabberwocky.com>
* sign.c (sign_file): Do not push textmode filter onto an unopened
IOBUF (segfault). Noted by Marcus Brinkmann. Push and
reinitialize textmode filter for each file in a multiple file
list.
* packet.h, getkey.c (fixup_uidnode), keyedit.c (show_prefs): Set
and show the keyserver no-modify flag.
* keygen.c (add_keyserver_modify): New.
(keygen_upd_std_prefs): Call it here.
(keygen_set_std_prefs): Accept "ks-modify" and "no-ks-modify" as
prefs to set and unset keyserver modify flag.
* g10.c (main): Accept "s1" in addition to "idea" to match the
other ciphers.
* main.h, misc.c (idea_cipher_warn): We don't need this if IDEA
has been disabled.
2003-02-21 David Shaw <dshaw@jabberwocky.com>
* keygen.c (keygen_set_std_prefs): Don't put AES or CAST5 in
default prefs if they are disabled.
* g10.c (main): Use 3DES instead of CAST5 if we don't have CAST5
support. Use 3DES for the s2k cipher in --openpgp mode.
(print_mds): #ifdef all of the optional digest algorithms.
2003-02-12 David Shaw <dshaw@jabberwocky.com>
* keydb.h, getkey.c (classify_user_id, classify_user_id2): Make
'exact' a per-desc item. Merge into one function since
'force_exact' is no longer needed.
(key_byname): Use new classify_user_id function, and new exact
flag in KEYDB_SEARCH_DESC.
* keyring.h, keyring.c (keyring_search): Return an optional index
to show which KEYDB_SEARCH_DESC was the matching one.
* keydb.h, keydb.c (keydb_search): Rename to keydb_search2, and
pass the optional index to keyring_search. Add a macro version of
keydb_search that calls this new function.
* export.c (do_export_stream): If the keyid! syntax is used,
export only that specified key. If the key in question is a
subkey, export the primary plus that subkey only.
2003-02-11 David Shaw <dshaw@jabberwocky.com>
* exec.c (set_exec_path): Add debugging line.
* g10.c (print_hex, print_mds): Print long hash strings a lot
neater. This assumes at least an 80-character display, as there
are a few other similar assumptions here and there. Users who
need unformatted hashes can still use with-colons. Check that
SHA384 and 512 are available before using them as they are no
longer always available.
* Makefile.am: Use a local copy of libexecdir along with @PACKAGE@
as GNUPG_LIBEXECDIR so it can be easily overridden at make time.
2003-02-04 David Shaw <dshaw@jabberwocky.com>
* armor.c (parse_hash_header, armor_filter): Accept the new SHAs
in the armor Hash: header.
* g10.c (print_hex): Print long hash strings a little neater.
(print_mds): Add the new SHAs to the hash list.
2003-02-02 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_revuid): Properly handle a nonselfsigned uid on
a v4 key (treat as a v4 revocation).
* import.c (print_import_check): Do not re-utf8 convert user IDs.
2003-01-27 David Shaw <dshaw@jabberwocky.com>
* mainproc.c (list_node): Show signature expiration date in
with-colons sig records.
* keylist.c (list_keyblock_colon), mainproc.c (list_node): Show
trust sig information in with-colons sig records.
2003-01-16 David Shaw <dshaw@jabberwocky.com>
* g10.c (add_group): Trim whitespace after a group name so it does
not matter where the user puts the = sign.
* options.skel: Comment out the first three lines in case someone
manually copies the skel file to their homedir.
* sign.c (clearsign_file): Only use pgp2mode with v3 keys and
MD5. This matches what we do when decoding such messages and
prevents creating a message (v3+RIPEMD/160) that we can't verify.
* sig-check.c (signature_check2): Use G10ERR_GENERAL as the error
for signature digest conflict. BAD_SIGN implies that a signature
was checked and we may try and print out a user ID for a key that
doesn't exist.
2003-01-15 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (init_trustdb, get_validity): Don't use a changed
trust model to indicate a dirty trustdb, and never auto-rebuild a
dirty trustdb with the "always" trust model.
* g10.c (add_group): Last commit missed the \t ;)
2003-01-14 David Shaw <dshaw@jabberwocky.com>
* packet.h, parse-packet.c (setup_user_id), free-packet.c
(free_user_id), keydb.h, keyid.c (namehash_from_uid): New function
to rmd160-hash the contents of a user ID packet and cache it in
the uid object.
* keylist.c (list_keyblock_colon): Use namehash in field 8 of
uids. Show dates for creation (selfsig date), and expiration in
fields 6 and 7.
* trustdb.c (get_validity, get_validity_counts, update_validity):
Use new namehash function rather than hashing it locally.
2003-01-14 Werner Koch <wk@gnupg.org>
* g10.c (add_group): Fixed group parsing to allow more than one
delimiter in a row and also allow tab as delimiter.
2003-01-12 David Shaw <dshaw@jabberwocky.com>
* tdbio.c (tdbio_set_dbname): Fix assertion failure with
non-fully-qualified trustdb names.
2003-01-11 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (get_validity_info, get_ownertrust_info,
trust_letter): Simplify by returning a ? for error directly.
* keyedit.c (show_key_with_all_names): Use get_validity_string and
get_ownertrust_string to show full word versions of trust
(i.e. "full" instead of 'f').
* trustdb.h, trustdb.c (get_ownertrust_string,
get_validity_string): Same as get_ownertrust_info, and
get_validity_info, except returns a full string.
* trustdb.c (get_ownertrust_with_min): New. Same as
'get_ownertrust' but takes the min_ownertrust value into account.
2003-01-10 David Shaw <dshaw@jabberwocky.com>
* armor.c (armor_filter): Comment about PGP's end of line tab
problem.
* trustdb.h, trustdb.c (trust_letter): Make
static. (get_ownertrust_info, get_validity_info): Don't mask the
trust level twice.
* trustdb.h, gpgv.c, trustdb.c (get_validity, get_validity_info),
keylist.c (list_keyblock_colon), keyedit.c
(show_key_with_all_names_colon, menu_revuid): Pass a user ID in
rather than a namehash, so we only have to do the hashing in one
place.
* packet.h, pkclist.c (build_pk_list), free-packet.c
(release_public_key_parts): Remove unused namehash element for
public keys.
2003-01-07 David Shaw <dshaw@jabberwocky.com>
* keygen.c (keygen_set_std_prefs): Warn when setting an IDEA
preference when IDEA is not available.
2003-01-06 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (get_validity_info): 'd' for disabled is not a
validity value any more.
* packet.h, tdbio.h, tdbio.c (tdbio_read_record,
tdbio_write_record), trustdb.c (update_validity): Store temporary
full & marginal counts in the trustdb.
(clear_validity, get_validity_counts): Return and clear temp
counts.
(store_validation_status): Keep track of which keyids have been
stored.
(validate_one_keyblock, validate_key_list): Use per-uid copies of
the full & marginal counts so they can be recalled for multiple
levels.
(validate_keys): Only use unused keys for each new round.
(reset_unconnected_keys): Rename to reset_trust_records, and only
skip specifically excluded records.
* keylist.c (print_capabilities): Show 'D' for disabled keys in
capabilities section.
* trustdb.c (is_disabled): Remove incorrect comment.
2003-01-03 David Shaw <dshaw@jabberwocky.com>
* import.c (import_one): Only do the work to create the status
display for interactive import if status is enabled.
* keyring.c (keyring_search): skipfnc didn't work properly with
non-keyid searches. Noted by Stefan Bellon.
* getkey.c (merge_selfsigs_main): Remove some unused code and make
sure that the pk selfsigversion member accounts for 1F direct
sigs.
2003-01-02 Werner Koch <wk@gnupg.org>
* keydb.c (keydb_add_resource): Don't assume that try_make_homedir
terminates but check again for the existence of the directory and
continue then.
* openfile.c (copy_options_file): Print a warning if the skeleton
file has active options.
2002-12-29 David Shaw <dshaw@jabberwocky.com>
* getkey.c (merge_selfsigs_main), main.h, sig-check.c
(check_key_signature2): Pass the ultimately trusted pk directly to
check_key_signature2 to avoid going through the key selection
mechanism. This prevents a deadly embrace when two keys without
selfsigs each sign the other.
2002-12-27 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_refresh): Don't print the "refreshing..."
line if there are no keys to refresh or if there is no keyserver
set.
* getkey.c (merge_selfsigs_main): Any valid user ID should make a
key valid, not just the last one. This also fixes Debian bug
#174276.
2002-12-27 Stefan Bellon <sbellon@sbellon.de>
* import.c (print_import_check): Changed int to size_t.
2002-12-27 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu, menu_revuid): Add "revuid" feature to
revoke a user ID. This is the same as issuing a revocation for
the self-signature, but a much simpler interface to do it.
2002-12-26 David Shaw <dshaw@jabberwocky.com>
* keydb.h, getkey.c (key_byname): Flag to enable or disable
including disabled keys. Keys specified via keyid (i.e. 0x...)
are always included.
* getkey.c (get_pubkey_byname, get_seckey_byname2,
get_seckey_bynames), keyedit.c (keyedit_menu, menu_addrevoker):
Include disabled keys in these functions.
* pkclist.c (build_pk_list): Do not include disabled keys for -r
or the key prompt. Do include disabled keys for the default key
and --encrypt-to.
* trustdb.h, trustdb.c (is_disabled): New skipfnc for skipping
disabled keys.
* gpgv.c (is_disabled): Stub.
* keygen.c (keygen_add_key_expire): Properly handle updating a key
expiration to a no-expiration value.
* keyedit.c (enable_disable_key): Comment.
* import.c (import_one): When in interactive mode and --verbose,
don't repeat some key information twice.
2002-12-22 Timo Schulz <ts@winpt.org>
* import.c (print_import_check): New.
(import_one): Use it here.
Use merge_keys_and_selfsig in the interactive mode to avoid
wrong key information.
* status.h: Add new status code.
* status.c: Ditto.
2002-12-13 David Shaw <dshaw@jabberwocky.com>
* pkclist.c (do_we_trust): Tweak language to refer to the "named
user" rather than "owner". Noted by Stefan Bellon.
* trustdb.h, trustdb.c (trustdb_pending_check): New function to
check if the trustdb needs a check.
* import.c (import_keys_internal): Used here so we don't rebuild
the trustdb if it is still clean.
(import_one, chk_self_sigs): Only mark trustdb dirty if the key
that is being imported has any sigs other than self-sigs.
Suggested by Adrian von Bidder.
* options.skel: Include the required '=' sign in the sample
'group' option. Noted by Stefan Bellon.
* import.c (chk_self_sigs): Don't try and check a subkey as if it
was a signature.
2002-12-11 David Shaw <dshaw@jabberwocky.com>
* tdbio.c (tdbio_read_record, tdbio_write_record): Compact the
RECTYPE_TRUST records a bit.
* g10.c (main): Comment out --list-trust-path until it can be
implemented.
* import.c (import_one): Warn when importing an Elgamal primary
that this may take some time (to verify self-sigs).
(chk_self_sigs): Try and cache all self-sigs so the keyblock is
written to the keyring with a good rich cache.
* keygen.c (ask_algo): Make the Elgamal sign+encrypt warning
stronger, and remove the RSA sign+encrypt warning.
2002-12-06 Stefan Bellon <sbellon@sbellon.de>
* options.h: Fixed typo (mangle_dos_names instead of
mangle_dos_filenames).
2002-12-05 Werner Koch <wk@gnupg.org>
* g10.c: New options --[no-]mangle-dos-filenames.
* options.h (opt): Added mangle-dos-filenames.
* openfile.c (open_outfile) [USE_ONLY_8DOT3]: Truncate the
filename only when this option is set; this is the default.
2002-12-04 David Shaw <dshaw@jabberwocky.com>
* main.h, keyedit.c, keygen.c: Back out previous (2002-12-01)
change. Minimal isn't always best.
* sign.c (update_keysig_packet): Use the current time rather then
a modification of the original signature time. Make sure that
this doesn't cause a time warp.
* keygen.c (keygen_add_key_expire): Properly handle a key
expiration date in the past (use a duration of 0).
* keyedit.c (menu_expire): Use update_keysig_packet so any sig
subpackets are maintained during the update.
* build-packet.c (build_sig_subpkt): Mark sig expired or unexpired
when the sig expiration subpacket is added.
(build_sig_subpkt_from_sig): Handle making an expiration subpacket
from a sig that has already expired (use a duration of 0).
* packet.h, sign.c (update_keysig_packet), keyedit.c
(menu_set_primary_uid, menu_set_preferences): Add ability to issue
0x18 subkey binding sigs to update_keysig_packet and change all
callers.
* trustdb.c (validate_keys): Show trust parameters when building
the trustdb, and make sure that the version record update was
successful.
(init_trustdb): If the current parameters aren't what was used for
building the trustdb, the trustdb is invalid.
* tbio.c (tdbio_db_matches_options): Update to work with new
trustdbs.
2002-12-03 David Shaw <dshaw@jabberwocky.com>
* tdbio.h, tdbio.c (tdbio_read_record, tdbio_write_record): Store
trust model in the trustdb version record.
(tdbio_update_version_record): New function to update version
record values during a trustdb check or update.
(tdbio_dump_record): Show trust model in dump.
* trustdb.c (validate_keys): Call tdbio_update_version_record on
success so that the correct options are stored in the trustdb.
* options.h: rearrange trust models so that CLASSIC is 0 and
OPENPGP is 1.
* options.h, g10.c (main), encode.c (write_pubkey_enc_from_list),
pkclist.c (algo_available), revoke.c (gen_revoke): Add --pgp8
mode. This is basically identical to --pgp7 in all ways except
that signing subkeys, v4 data sigs (including expiration), and SK
comments are allowed.
* getkey.c (finish_lookup): Comment.
* main.h, keylist.c (reorder_keyblock), keyedit.c (keyedit_menu):
Reorder user ID display in the --edit-key menu to match that of
the --list-keys display.
* g10.c (add_notation_data): Fix initialization.
2002-12-01 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_expire): Don't lose key flags when changing the
expiration date of a subkey. This is not the most optimal
solution, but it is minimal change on the stable branch.
* main.h, keygen.c (do_copy_key_flags): New function to copy key
flags, if any, from one sig to another.
(do_add_key_expire): New function to add key expiration to a sig.
(keygen_copy_flags_add_expire): New version of
keygen_add_key_expire that also copies key flags.
(keygen_add_key_flags_and_expire): Use do_add_key_expire.
* import.c (fix_hkp_corruption): Comment.
2002-11-25 Stefan Bellon <sbellon@sbellon.de>
* plaintext.c (handle_plaintext) [__riscos__]: If nooutput is set,
no filetype is needed obviously.
2002-11-24 David Shaw <dshaw@jabberwocky.com>
* main.h, misc.c (default_cipher_algo, default_compress_algo):
New. Return the default algorithm by trying
--cipher-algo/--compress-algo, then the first item in the pref
list, then s2k-cipher-algo or ZIP.
* sign.c (sign_file, sign_symencrypt_file), encode.c
(encode_simple, encode_crypt): Call default_cipher_algo and
default_compress_algo to get algorithms.
* g10.c (main): Allow pref selection for compress algo with
--openpgp.
* mainproc.c (proc_encrypted): Use --s2k-digest-algo for
passphrase mangling rather than --digest-algo.
* sign.c (hash_for): If --digest-algo is not set, but
--personal-digest-preferences is, then use the first hash
algorithm in the personal list. If the signing algorithm is DSA,
then use the first 160-bit hash algorithm in the personal list.
If --pgp2 is set and it's a v3 RSA key, use MD5.
* g10.c (main), keydb.c (keydb_add_resource,
keydb_locate_writable): Rename --default-keyring as
--primary-keyring. Stefan wins the naming contest.
2002-11-23 David Shaw <dshaw@jabberwocky.com>
* g10.c (add_notation_data): Disallow notation names that do not
contain a '@', unless --expert is set. This is to help prevent
people from polluting the (as yet unused) IETF namespace.
* main.h: Comments about default algorithms.
* photoid.c (image_type_to_string): Comments about 3-letter file
extensions.
* encode.c (encode_simple), passphrase.c (passphrase_to_dek),
sign.c (sign_symencrypt_file): Use --s2k-digest-algo for
passphrase mangling rather than --digest-algo.
2002-11-21 David Shaw <dshaw@jabberwocky.com>
* keygen.c (keygen_set_std_prefs): Properly handle an empty
preference string.
* misc.c (string_to_compress_algo): "none" is a bad choice since
it conflicts with the "none" in setpref.
2002-11-14 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Allow compression algorithm names as the argument
to --compress-algo. The old algorithm names still work for
backwards compatibility.
* misc.c (string_to_compress_algo): Allow "none" as an alias for
"uncompressed".
2002-11-13 Stefan Bellon <sbellon@sbellon.de>
* getkey.c (get_pubkey_byfprint_fast): Fixed type incompatibility,
was unsigned char instead of byte.
2002-11-13 David Shaw <dshaw@jabberwocky.com>
* encode.c (encode_simple): Make sure that files larger than about
4G use partial length encoding. This is required because OpenPGP
allows only for 32 bit length fields. From Werner on stable
branch.
* getkey.c (get_pubkey_direct): Renamed to...
(get_pubkey_fast): this and made extern.
(get_pubkey_byfprint_fast): New. From Werner on stable branch.
* keydb.h, import.c (import_one): Use get_pubkey_fast instead of
get_pubkey. We don't need a merged key and actually this might
lead to recursions.
(revocation_present): Likewise for search by fingerprint. From
Werner on stable branch.
* g10.c (main): Try to create the trustdb even for non-colon-mode
list-key operations. This is required because getkey needs to
know whether a a key is ultimately trusted. From Werner on stable
branch.
* exec.c [__CYGWIN32__]: Keep cygwin separate from Mingw32;
we don't need it here as it behaves more like a Posix system.
From Werner on stable branch.
* passphrase.c (agent_get_passphrase): Ditto. From Werner on
stable branch.
* tdbio.c (MY_O_BINARY): Need binary mode with Cygwin. From
Werner on stable branch.
* g10.c, gpgv.c (main) [__CYGWIN32__]: Don't get the homedir from
the registry. From Werner on stable branch.
* keyedit.c (show_key_with_all_names_colon): Make --with-colons
--edit display match the validity and trust of --with-colons
--list-keys.
* passphrase.c (agent_send_all_options): Fix compile warning.
* keylist.c (list_keyblock_colon): Validity for subkeys should
match that of the primary key, and not that of the last user ID.
* getkey.c (merge_selfsigs): Revoked/expired/invalid primary keys
carry these facts onto all their subkeys, but only after the
subkey has a chance to be marked valid. This is to fix an
incorrect "invalid public key" error verifying a signature made by
a revoked signing subkey, with a valid unrevoked primary key.
2002-11-09 Werner Koch <wk@gnupg.org>
* passphrase.c (agent_send_all_options): Use tty_get_ttyname to
get the default ttyname.
2002-11-07 David Shaw <dshaw@jabberwocky.com>
* keyring.h, keyring.c (keyring_register_filename): Return the
pointer if a given keyring is registered twice.
* keydb.h, keydb.c (keydb_add_resource): Use flags to indicate a
default keyring.
(keydb_locate_writable): Prefer the default keyring if possible.
* g10.c (main): Add --default-keyring option.
2002-11-06 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), trustdb.c (ask_ownertrust): Add
--force-ownertrust option for debugging purposes. This allows
setting a whole keyring to a given trust during an
--update-trustdb. Not for normal use - it's just easier than
hitting "4" all the time to test a large trustdb.
* pubkey-enc.c (get_session_key): With hidden recipients or try a
given passphrase against all secret keys rather than trying all
secret keys in turn. Don't if --try-all-secrets or --status-fd is
enabled.
* passphrase.c (passphrase_to_dek): Mode 1 means do a regular
passphrase query, but don't prompt with the key info.
* seckey-cert.c (do_check, check_secret_key): A negative ask count
means to enable passphrase mode 1.
* keydb.h, getkey.c (enum_secret_keys): Add flag to include
secret-parts-missing keys (or not) in the list.
2002-11-05 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_search_prompt): When --with-colons is
enabled, don't try and fit the search output to the screen size -
just dump the whole list.
2002-11-04 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_search_prompt): When --with-colons is
enabled, just dump the raw keyserver protocol to stdout and don't
print the menu.
* keyserver.c (show_prompt): Don't show a prompt when command-fd
is being used.
* trustdb.c (trust_model_string, check_trustdb, update_trustdb,
validate_one_keyblock): It's not clear what a trustdb rebuild or
check means with a trust model other than "classic" or "openpgp",
so disallow this.
2002-11-03 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main): Add --trust-model option. Current
models are "openpgp" which is classic+trustsigs, "classic" which
is classic only, and "always" which is the same as the current
option --always-trust (which still works). Default is "openpgp".
* trustdb.c (validate_one_keyblock): Use "openpgp" trust model to
enable trust sigs.
* gpgv.c (main), mainproc.c (check_sig_and_print), pkclist.c
(do_we_trust, do_we_trust_pre, check_signatures_trust): Use new
--trust-model option in place of --always-trust.
* keyedit.c (sign_mk_attrib, trustsig_prompt, sign_uids,
keyedit_menu): Prompt for and create a trust signature with
"tsign". This is functional, but needs better UI text.
* build-packet.c (build_sig_subpkt): Able to build trust and
regexp subpackets.
* pkclist.c (do_edit_ownertrust): Comment.
2002-11-02 David Shaw <dshaw@jabberwocky.com>
* keygen.c (set_one_pref, keygen_set_std_prefs): Allow using the
full algorithm name (CAST5, SHA1) rather than the short form (S3,
H2).
* main.h, keygen.c (keygen_get_std_prefs), keyedit.c
(keyedit_menu): Return and use a fake uid packet rather than a
string since we already have a nice parser/printer in
keyedit.c:show_prefs.
* main.h, misc.c (string_to_compress_algo): New.
2002-11-01 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Add --no-throw-keyid.
* keydb.h, encode.c (write_pubkey_enc_from_list), g10.c (main),
pkclist.c (build_pk_list): Add --hidden-recipient (-R) and
--hidden-encrypt-to, which do a single-user variation on
--throw-keyid. The "hide this key" flag is carried in bit 0 of
the pk_list flags field.
* keyserver.c (parse_keyrec): Fix shadowing warning.
2002-10-31 Stefan Bellon <sbellon@sbellon.de>
* compress.c (init_compress) [__riscos__]: Use
riscos_load_module() to load ZLib module.
* g10.c (main) [__riscos__]: Renames due to changes in riscos.c
(e.g. prefixes all RISC OS specific functions with riscos_*).
* photoid.c (show_photos) [__riscos__]: Likewise.
* signal.c (got_fatal_signal) [__riscos__]: Likewise.
* trustdb.c (check_regexp) [__riscos__]: Branch to RISC OS RegEx
handling.
2002-10-31 David Shaw <dshaw@jabberwocky.com>
* build-packet.c (do_plaintext), encode.c (encode_sesskey,
encode_simple, encode_crypt), sign.c (write_plaintext_packet): Use
wipememory() instead of memset() to wipe sensitive memory as the
memset() might be optimized away.
2002-10-30 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (check_regexp): Modern regexps require REG_EXTENDED.
2002-10-29 David Shaw <dshaw@jabberwocky.com>
* packet.h, trustdb.h, trustdb.c (trust_string): New. Return a
string like "fully trusted", "marginally trusted", etc.
(get_min_ownertrust): New. Return minimum ownertrust.
(update_min_ownertrust): New. Set minimum ownertrust.
(check_regexp): New. Check a regular epression against a user ID.
(ask_ownertrust): Allow specifying a minimum value.
(get_ownertrust_info): Follow the minimum ownertrust when
returning a letter.
(clear_validity): Remove minimum ownertrust when a key becomes
invalid.
(release_key_items): Release regexp along with the rest of the
info.
(validate_one_keyblock, validate_keys): Build a trust sig chain
while validating. Call check_regexp for regexps. Use the minimum
ownertrust if the user does not specify a genuine ownertrust.
* pkclist.c (do_edit_ownertrust): Only allow user to select a
trust level greater than the minimum value.
* parse-packet.c (can_handle_critical): Can handle critical trust
and regexp subpackets.
* trustdb.h, trustdb.c (clear_ownertrusts), delkey.c
(do_delete_key), import.c (import_one): Rename clear_ownertrust to
clear_ownertrusts and have it clear the min_ownertrust value as
well.
* keylist.c (list_keyblock_print): Indent uid to match pub and
sig.
* keyedit.c (print_and_check_one_sig, show_key_and_fingerprint,
menu_addrevoker), keylist.c (list_keyblock_print,
print_fingerprint): Show "T" or the trust depth for trust
signatures, and add spaces to some strings to make room for it.
* packet.h, parse-packet.c (dump_sig_subpkt, parse_one_sig_subpkt,
parse_signature): Parse trust signature values.
* tdbio.h, tdbio.c (tdbio_read_record, tdbio_write_record):
Reserve a byte for the minimum ownertrust value (for use with
trust signatures).
2002-10-29 Stefan Bellon <sbellon@sbellon.de>
* build-packet.c (calc_plaintext, do_plaintext): Removed RISC OS
specific filetype parts (it's now done in make_basename()).
* plaintext.c (handle_plaintext): Tidied up RISC OS specific
filetype parts.
* encode.c (encode_simple, encode_crypt): Added argument to
make_basename() call.
* sign.c (write_plaintext_packet): Added argument to
make_basename() call.
2002-10-28 Stefan Bellon <sbellon@sbellon.de>
* build-packet.c (calc_plaintext, do_plaintext): Added filetype
handling for RISC OS' file types.
* plaintext.c (handle_plaintext) [__riscos__]: Added filetype
handling for RISC OS' file types.
2002-10-23 David Shaw <dshaw@jabberwocky.com>
* main.h, import.c (sec_to_pub_keyblock, import_secret_one,
parse_import_options), g10.c (main): New import-option
"convert-sk-to-pk" to convert a secret key into a public key
during import. It is on by default.
2002-10-23 Werner Koch <wk@gnupg.org>
* pubkey-enc.c (get_it): Fix segv, test for revoked only when PK
has been assigned.
2002-10-18 Timo Schulz <ts@winpt.org>
* keylist.c: (print_pubkey_info): New.
(print_seckey_info): New.
* main.h: Prototypes for the new functions.
* delkey.c (do_delete_key): Use it here.
* revoke.c (gen_desig_revoke): Ditto.
2002-10-17 Werner Koch <wk@gnupg.org>
* pkclist.c (do_edit_ownertrust): Show all user IDs. This should
be enhanced to also show the current trust level. Suggested by
Florian Weimer.
2002-10-17 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Handle --strict and --no-strict from the command
line before the options file is loaded.
2002-10-15 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Disable --textmode when encrypting (symmetric or
pk) in --pgp2 mode as PGP 2 can't handle the unknown length
literal packet. Reported by Michael Richardson.
2002-10-14 David Shaw <dshaw@jabberwocky.com>
* keyserver-internal.h, keyserver.c (print_keyrec, parse_keyrec,
show_prompt, keyserver_search_prompt, keyserver_spawn): Go to
version 1 of the keyserver protocol. This is a better design,
similar to --with-colons, that allows for keys with multiple user
IDs rather than using multiple keys. It also matches the machine
readable pksd format. Also use a prettier --search-keys listing
format that can fill different size windows (currently set at 24
lines).
2002-10-12 Werner Koch <wk@gnupg.org>
* keygen.c (print_status_key_created): New.
(do_generate_keypair): Use it to print the fingerprint.
(generate_subkeypair): Likewise.
2002-10-11 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_addrevoker): Properly back out if the signature
fails. Also, do not allow appointing the same revoker twice, and
report ALREADY_SIGNED if the user tries it.
2002-10-07 David Shaw <dshaw@jabberwocky.com>
* import.c (import_keys_internal): Missed one s/inp/inp2/.
* keylist.c (print_capabilities): Properly indicate per-key
capabilities of sign&encrypt primary keys that have
secret-parts-missing (i.e. no capabilities at all)
* mainproc.c (symkey_decrypt_sesskey): Fix compiler warning.
2002-10-04 David Shaw <dshaw@jabberwocky.com>
* getkey.c (get_pubkey_direct): Don't cache keys retrieved via
this function as they may not have all their fields filled in.
* sig-check.c (signature_check2): Use new is_primary flag to check
rather than comparing main_keyid with keyid as this still works in
the case of a not fully filled in pk.
2002-10-04 Werner Koch <wk@gnupg.org>
* import.c (import_keys_internal): s/inp/inp2/ to avoid shadowing
warning.
* passphrase.c (agent_get_passphrase): Fixed signed/unsigned char
problem in %-escaping. Noted by Ingo Klöcker.
2002-10-03 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main): Add --strict and --no-strict to switch
the log_warning severity level from info to error.
* keylist.c (print_capabilities): Secret-parts-missing keys should
show that fact in the capabilities, and only primary signing keys
can certify other keys.
* packet.h, parse_packet.c (parse_key): Add is_primary flag for
public keys (it already exists for secret keys).
2002-10-02 David Shaw <dshaw@jabberwocky.com>
* import.c (import_secret_one): Check for an illegal (>110)
protection cipher when importing a secret key.
* keylist.c (list_keyblock_print): Show a '#' for a
secret-parts-missing key.
* parse_packet.c (parse_key): Some comments.
* revoke.c (gen_revoke): Remove some debugging code.
* trustdb.c (verify_own_keys): Make trusted-key a non-deprecated
option again.
* seckey-cert.c (do_check): Don't give the IDEA warning unless the
cipher in question is in fact IDEA.
2002-10-01 David Shaw <dshaw@jabberwocky.com>
* import.c (import_one): Make sure that a newly imported key
starts with a clean ownertrust.
2002-10-01 Werner Koch <wk@gnupg.org>
* getkey.c (get_pubkey_direct): New.
(merge_selfsigs_main): Use it here to look for an ultimately
trusted key. Using the full get_pubkey might lead to an
infinitive recursion.
2002-09-29 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (parse_keyserver_uri): Force the keyserver URI
scheme to lowercase to be case-insensitive.
2002-09-28 David Shaw <dshaw@jabberwocky.com>
* export.c (do_export_stream): Comment.
* sig-check.c (check_key_signature2): Properly handle a
non-designated revocation import.
2002-09-26 Werner Koch <wk@gnupg.org>
* g10.c (set_homedir): New. Changed all direct assignments to use
this.
* gpgv.c (set_homedir): Ditto.
2002-09-25 David Shaw <dshaw@jabberwocky.com>
* Makefile.am: Link gpg with EGDLIBS (i.e. NETLIBS) as EGD uses
sockets. Remove the old NETLIBS variable since the keyserver
stuff is no longer internal.
2002-09-24 David Shaw <dshaw@jabberwocky.com>
* import.c (import_keys_stream): Fix compiler type warning.
* keyring.c (keyring_rebuild_cache), sig-check.c
(check_key_signature2), import.c (import, chk_self_sigs): Minor
language cleanups.
2002-09-23 Stefan Bellon <sbellon@sbellon.de>
* main.h: Introduced fast-import as import option. Removed
fast as separate option from prototypes.
* import.c (parse_import_options): Added fast-import option.
(import_*): Removed fast as separate option.
* g10.c (main): Added option fast-import, removed old fast
as separate argument.
* keyserver.c (keyserver_spawn): Removed old fast as separate
argument.
2002-09-22 Stefan Bellon <sbellon@sbellon.de>
* import.c (import_keys, import_keys_stream,
import_keys_internal): Added trustdb update/check to key import if
not fast-import and interactive set/no-auto-check-trustdb unset.
Avoided function clone by introducing import_keys_internal.
2002-09-19 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_spawn): Properly handle line truncation.
Don't leak memory (~10-20 bytes) on searches.
(keyserver_search_prompt): Cleanup.
* keylist.c (list_keyblock_colon): Show 1F direct key signatures
in --with-colons listing.
2002-09-16 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_addrevoker): The direct key signature for
revocation keys must be at least v4 to carry the revocation key
subpacket. Add a PGP 2.x warning for revocation keys.
2002-09-14 David Shaw <dshaw@jabberwocky.com>
* g10.c (check_permissions): Rearrange strings to make translating
easier (don't incorporate string parts).
* keyedit.c (sign_uids): Make strings translatable.
* sig-check.c (check_key_signature2): Make string translatable.
2002-09-13 David Shaw <dshaw@jabberwocky.com>
* getkey.c (check_revocation_keys): Move....
* main.h, sig-check.c (check_revocation_keys): to here. Also
return the signature_check error code rather than 0/1 and cache
the sig result.
* sig-check.c (check_key_signature2): Divert to
check_revocation_keys if a revocation sig is made by someone other
than the pk owner.
* getkey.c (merge_selfsigs_main): Tidy.
2002-09-13 Werner Koch <wk@gnupg.org>
* g10.c (main) [__MINGW32__]: Activate oLoadExtension.
2002-09-12 David Shaw <dshaw@jabberwocky.com>
* Makefile.am, hkp.c, hkp.h, keyserver.c (keyserver_work): Remove
internal HKP support.
* keyserver.c (keyserver_spawn): Remove whitespace after keyserver
commands.
2002-09-10 David Shaw <dshaw@jabberwocky.com>
* exec.c (expand_args): Remove loop left over from earlier
implementation.
(exec_write): Missed one tick.
2002-09-10 Werner Koch <wk@gnupg.org>
* g10.c, options.h: Removed option --emulate-checksum-bug.
* misc.c (checksum_u16_nobug): Removed.
(checksum_u16): Removed the bug emulation.
(checksum_mpi): Ditto.
(checksum_mpi_counted_nbits): Removed and replaced all calls
with checksum_mpi.
* parse-packet.c (read_protected_v3_mpi): New.
(parse_key): Use it here to store it as an opaque MPI.
* seckey-cert.c (do_check): Changed the v3 unprotection to the new
why to store these keys.
(protect_secret_key): Likewise.
* build-packet.c (do_secret_key): And changed the writing.
* tdbio.c (tdbio_set_dbname, open_db): Use new macro MY_O_BINARY
to avoid silly ifdefs.
(open_db): Fallback to RDONLY so that gpg may be used from a
RO-medium.
* encode.c (encode_simple): Make sure we don't use an ESK packet
when we don't have a salt in the S2K.
* misc.c (pct_expando) <case f>: Make sure that LEN is initialized.
* exec.c (exec_finish): Use ticks to denote filenames in messages.
(make_tempdir, exec_write): Changed format of messages.
* keyserver.c (print_keyinfo): Release USERID in on error.
(keyserver_work) [!DISABLE_KEYSERVER_HELPERS]: Exclude the unused
code.
2002-09-09 Werner Koch <wk@gnupg.org>
* parse-packet.c (make_attribute_uidname): Add new ar MAX_NAMELEN
for sanity checks. Changed both callers. Limit the size of an %s.
* options.skel: Comment lock-once out, so that this file does not
change anything when copied to a new home directory.
* openfile.c (try_make_homedir): Don't exit after copying the
option skeleton.
* options.h: Don't use a comma when declaring variables over more
than one line.
* mainproc.c (symkey_decrypt_sesskey): Check length of the session
key.
* hkp.c (dehtmlize): Use ascii_tolower to protect against weird
locales. Cast the argument for isspace for the sake of broken
HP/UXes.
(parse_hkp_index): s/ascii_memcasecmp/ascii_strncasecmp/.
* g10.c: Removed option --emulate-3des-s2k-bug.
* passphrase.c (hash_passphrase): Was used here.
* export.c (parse_export_options)
* keyserver.c (parse_keyserver_options)
* import.c (parse_import_options)
* g10.c (check_permissions): s/ascii_memcasecmp/ascii_strncasecmp/.
2002-09-09 David Shaw <dshaw@jabberwocky.com>
* g10.c (add_group): Use '=' to separate group name from group
members. Use a better error message for when no = is found.
* hkp.c (hkp_export): Use CRLF in headers.
2002-09-03 David Shaw <dshaw@jabberwocky.com>
* mainproc.c (print_pkenc_list): Don't increment the error counter
when printing the list of keys a message was encrypted to. This
would make gpg give a non-zero exit code even for completely valid
messages if the message was encrypted to more than one key that
the user owned.
2002-09-02 Werner Koch <wk@gnupg.org>
* g10.c (main): Try to set a default character set. Print the
used one in verbosity level 3.
* gpgv.c (main): Try to set a default character set.
* status.c, status.h (STATUS_IMPORT_OK): New.
* import.c (import_one,import_secret_one): Print new status.
2002-08-30 David Shaw <dshaw@jabberwocky.com>
* pkclist.c (build_pk_list): Add new status code to indicate an
untrusted user. This (or a disabled key) fail with "unavailable
pubkey" (G10ERR_UNU_PUBKEY).
* pkclist.c (build_pk_list): Fail if any recipient keys are
unusable.
* options.skel: The PGP LDAP keyserver is back. Use MIT keyserver
as a sample rather than cryptnet as cryptnet does not support
searching yet.
* keyedit.c (show_key_with_all_names): Fix error message
(preferences are userid/selfsig and not key specific).
2002-08-30 Werner Koch <wk@gnupg.org>
* pkclist.c (do_we_trust_pre): Changed the wording of a warning.
* encode.c (encode_simple,encode_crypt): Use new style CTB for
compressssed packets when using MDC. We need to do this so that
concatenated messages are properly decrypted. Old style
compression assumes that it is the last packet; given that we
can't determine the length in advance, the uncompressor does not
know where to start. Actually we should use the new CTB always
but this would break PGP 2 compatibility.
* parse-packet.c (parse): Special treatment for new style CTB
compressed packets.
* build-packet.c (do_mdc): Removed. Was not used.
(do_encrypted_mdc): Count in the version number and the MDC packet.
2002-08-28 David Shaw <dshaw@jabberwocky.com>
* sig-check.c (do_check_messages, do_check): Show keyid in error
messages.
* keyserver.c (print_keyinfo): More readable key listings for
--search-keys responses.
2002-08-26 David Shaw <dshaw@jabberwocky.com>
* hkp.c (parse_hkp_index, dehtmlize): Move HTML functionality into
new "dehtmlize" function. Remove HTML before trying to parse each
line from the keyserver. If the keyserver provides key type
information in the listing, use it.
2002-08-23 David Shaw <dshaw@jabberwocky.com>
* sig-check.c (do_check, do_check_messages): Emit the usual sig
warnings even for cached sigs. This also serves to protect
against missing a sig expiring while cached.
* getkey.c (merge_selfsigs_main): Don't check UID self-sigs twice.
2002-08-22 David Shaw <dshaw@jabberwocky.com>
* import.c (clean_subkeys, chk_self_sigs): Merge clean_subkeys
into chk_self_sigs. This improves efficiency as the same
signatures are not checked multiple times. Clarify when a subkey
is revoked (any revocation signature, even if it is dated before
the binding signature).
* getkey.c (merge_selfsigs_subkey): Subkey revocation comments.
* keylist.c (list_one): Stats are only for public key listings.
* g10.c (main), options.skel: Default should be include-revoked
for keyserver operations.
2002-08-21 Werner Koch <wk@gnupg.org>
* import.c (import_print_stats): Print new non_imported counter
which is currently not used because we terminate on errors.
2002-08-20 David Shaw <dshaw@jabberwocky.com>
* options.skel: Document no-include-attributes for
keyserver-options.
* keylist.c, keyedit.c, keyserver.c, sign.c: Some TODOs and
comments.
* export.c (do_export_stream): Fix noop bug in exporting sensitive
revocation keys.
* pkclist.c (do_edit_ownertrust): Comment out the option for
showing trust paths until it can be implemented.
2002-08-19 Werner Koch <wk@gnupg.org>
* getkey.c (get_user_id_native): Renamed to ..
(get_user_id_printable): this. Filter out all dangerous
characters. Checked all usages.
(get_user_id_string_native): Renamed to..
(get_user_id_string_printable): this. Filter out all dangerous
characters. Checked all usages.
* keyedit.c (show_basic_key_info): New.
* keylist.c (print_fingerprint): New mode 3.
* import.c (import_one): Use new function to display the user ID.
2002-08-16 Timo Schulz <ts@winpt.org>
* g10.c (main): Enable opt.interactive.
* import.c (import_one): Ask the user if the key shall be
imported when the interactive mode is used. Useful to extract
selected keys from a file.
2002-08-16 Werner Koch <wk@gnupg.org>
* seckey-cert.c: Workaround to allow decryption of v3 keys created
with a bug in the mpi_get_secure_buffer.
2002-08-14 David Shaw <dshaw@jabberwocky.com>
* hkp.c (parse_hkp_index): Properly handle really large keys
(5 digit key length) in HKP searches.
2002-08-13 David Shaw <dshaw@jabberwocky.com>
* encode.c (encode_simple): Fix problem with using compression
algo 2 and symmetric compressed files.
* encode.c (encode_simple, encode_crypt): If we are not using a
MDC, compress even if a file is already compressed. This is to
help against the chosen ciphertext attack.
* pkclist.c (select_algo_from_prefs): Fix requested algorithm bug
so the request succeeds even if the requested algorithm is not the
first found.
* cipher.c (write_header), encode.c (use_mdc, encode_simple,
encode_crypt, encrypt_filter), g10.c (main): Be more eager to use
a MDC. We use a MDC if the keys directly support it, if the keys
list AES (any) or TWOFISH anywhere in the prefs, or if the cipher
chosen does not have a 64 bit blocksize.
2002-08-08 David Shaw <dshaw@jabberwocky.com>
* options.skel: Some language tweaks, and remove the
load-extension section for random gatherers.
* keyring.c (create_tmp_file, rename_tmp_file): Create tmp files
with user-only permissions, but restore the original permissions
if the user has something special set.
* openfile.c (copy_options_file): Create new options file
(gpg.conf) with user-only permissions.
* keydb.c (keydb_add_resource): Create new keyrings with user-only
permissions.
* tdbio.c (tdbio_set_dbname): Create new trustdbs with user-only
permissions.
2002-08-07 David Shaw <dshaw@jabberwocky.com>
* sig-check.c (signature_check2): Sanity check that the md has a
context for the hash that the sig is expecting. This can happen
if a onepass sig header does not match the actual sig, and also if
the clearsign "Hash:" header is missing or does not match the
actual sig.
* keyedit.c (menu_revsig): Properly show a uid is revoked without
restarting gpg. This is Debian bug 124219, though their supplied
patch will not do the right thing.
* main.h, tdbio.c (tdbio_set_dbname), misc.c (removed
check_permissions), keydb.c (keydb_add_resource), g10.c (main,
check_permissions): Significant reworking of the permission check
mechanism. The new behavior is to check everything in the homedir
by checking the homedir itself. If the user wants to put
(possibly shared) keyrings outside the homedir, they are not
checked. The options file and any extension files are checked
wherever they are, as well as their enclosing directories. This
is Debian bug 147760.
2002-08-06 Stefan Bellon <sbellon@sbellon.de>
* g10.c (main): Use of EXTSEP_S in new gpg.conf string.
* openfile.c (copy_options_file): Ditto.
2002-08-06 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), mainproc.c (proc_encrypted):
--ignore-mdc-error option to turn a MDC check error into a
warning.
* encode.c (encode_crypt), g10.c (main), sign.c (sign_file,
clearsign_file): Use the same --pgpX warning string everywhere to
ease translations.
* encode.c (write_pubkey_enc_from_list): Warn when using
--throw-keyid with --pgpX. Noted by Vedaal Nistar.
* revoke.c (export_minimal_pk, gen_desig_revoke, gen_revoke):
Export a minimal pk along with the revocation cert when in --pgpX
mode so that PGP can import it.
2002-08-06 Werner Koch <wk@gnupg.org>
* options.skel: Changed comments.
* g10.c (main): Try to use "gpg.conf" as default option file.
* openfile.c (copy_options_file): Changed name of created file.
2002-08-02 Werner Koch <wk@gnupg.org>
* Makefile.am (LDFLAGS): Removed DYNLINK_LDFLAGS.
2002-07-30 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), mainproc.c (proc_encrypted): Return a
decryption failed error if a MDC does not verify. Warn if a MDC
is not present (can disable via --no-mdc-warning).
* exec.c (exec_write), g10.c (main), keyserver.c
(keyserver_spawn): Use new DISABLE_KEYSERVER_PATH rather than
FIXED_EXEC_PATH.
2002-07-28 David Shaw <dshaw@jabberwocky.com>
* sig-check.c (do_check): Properly validate v4 sigs with no hashed
section at all.
2002-07-25 Werner Koch <wk@gnupg.org>
* delkey.c (do_delete_key): Always allow to delete a key in batch mode
when specified by fingerprint. Suggested by Enzo Michelangeli.
2002-07-25 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_revsig): Change "revsig" to honor selected uids
so the user can revoke sigs from particular uids only.
* keylist.c (list_keyblock_print): Don't display expired uids in
--list-keys unless -v and not --list-sigs (just like revoked
uids).
* exec.c, export.c, import.c, keyedit.c, keyserver.c, misc.c:
"Warning" -> "WARNING"
2002-07-24 David Shaw <dshaw@jabberwocky.com>
* main.h, import.c (parse_import_options, fix_hkp_corruption,
import_one, delete_inv_parts), g10.c (main): New import-option
"repair-hkp-subkey-bug", which repairs as much as possible the HKP
mangling multiple subkeys bug. It is on by default for keyserver
receives, and off by default for regular --import.
* main.h, import.c (import, import_one, delete_inv_parts), hkp.c
(hkp_ask_import), keyserver.c (keyserver_spawn): Use keyserver
import options when doing keyserver receives.
* options.h, exec.h, exec.c (set_exec_path, exec_write), g10.c
(main), keyserver.c (keyserver_spawn): If the user does not use
"exec-path", completely replace $PATH with GNUPG_LIBEXECDIR before
calling the keyserver helper. If the user does use "exec-path",
append GNUPG_LIBEXECDIR after the specified path.
2002-07-23 David Shaw <dshaw@jabberwocky.com>
* import.c (parse_import_options), export.c
(parse_export_options): Fix offset problem with reversed ("no-")
meanings.
* import.c (delete_inv_parts): Discard subkey signatures (0x18 and
0x28) if found in the userid section of the key.
* sig-check.c (signature_check2): Signatures made by invalid
subkeys (bad/missing binding sig) are also invalid.
* keylist.c (print_fingerprint): Show the primary as well as the
secondary key fingerprint in modes 1 & 2.
2002-07-22 David Shaw <dshaw@jabberwocky.com>
* options.h, main.h, g10.c (main), import.c
(parse_import_options, delete_inv_parts), keyserver.c
(parse_keyserver_options): add new --import-options option. The
only current flag is "allow-local-sigs".
* g10.c (main): Don't disable MDC in pgp7 mode.
* options.h, g10.c (main), keyserver.c (parse_keyserver_options):
Remove old keyserver-option include-attributes now that there is
an export-option for the same thing.
* options.h, main.h, export.c (parse_export_options,
do_export_stream), g10.c (main): add new --export-options option.
Current flags are "include-non-rfc", "include-local-sigs",
"include-attributes", and "include-sensitive-revkeys".
* options.h, hkp.c (hkp_export), keyserver.c
(parse_keyserver_options, keyserver_spawn): try passing unknown
keyserver options to export options, and if successful, use them
when doing a keyserver --send-key.
* build-packet.c (build_sig_subpkt): We do not generate
SIGSUBPKT_PRIV_VERIFY_CACHE anymore.
* revoke.c (gen_desig_revoke): Lots more comments about including
sensitive revkeys along with the revocation sig itself.
* keyserver.c (parse_keyserver_options): Simpler implementation
that can skip one pass over the options.
2002-07-18 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu, menu_addrevoker): Allow specifying
"sensitive" as an argument to an addrevoker command. This sets
the 0x40 sensitive revoker flag.
* revoke.c (gen_desig_revoke): When generating a designated
revocation, include the direct key sig that contains the
designated revoker subpacket. This allows sensitive designated
revocation subpackets to be exported. Also indicate which
revokers are sensitive in the first place.
2002-07-17 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (show_key_with_all_names_colon): The 0x40 class bit in
a designated revoker means "sensitive", not "local". It's
exportable under the right circumstances.
* main.h, options.h, export.c (do_export_stream), g10.c (main),
hkp.c (hkp_export), keyserver.c (keyserver_spawn: Add a flag to
skip attribute packets and their signatures while exporting. This
is to accomodate keyservers (pksd again) that choke on attributes.
Use keyserver-option "include-attributes" to control it. This
defaults to ON (i.e. don't skip).
2002-07-09 David Shaw <dshaw@jabberwocky.com>
* options.h, keyserver.c (parse_keyserver_uri, keyserver_spawn,
keyserver_work), hkp.c (hkp_ask_import, hkp_export, hkp_search):
Use a much more strict reading of RFC-2396 for the keyserver URIs.
Specifically, don't try and be smart about checking the value of
":port" so long as it is all digits, and properly handle opaque
data (those scheme specific parts that do not start with "//").
2002-07-04 David Shaw <dshaw@jabberwocky.com>
* photoid.c (get_default_photo_command, show_photos): Honor
FIXED_PHOTO_VIEWER and DISABLE_PHOTO_VIEWER.
* mainproc.c (check_sig_and_print): Use --show-photos to show
photos when verifying a sig made by a key with a photo.
* keyserver.c (parse_keyserver_uri): Properly parse a URI with no
:port section and an empty file path, but with a terminating '/'.
(keyserver_work): Honor DISABLE_KEYSERVER_HELPERS.
* hkp.c (hkp_ask_import): Display keyserver URI as a URI, but only
if verbose.
* exec.c, g10.c: USE_EXEC_PATH -> FIXED_EXEC_PATH
2002-07-03 David Shaw <dshaw@jabberwocky.com>
* exec.h, exec.c (set_exec_path, exec_write), g10.c (main): If
USE_EXEC_PATH is defined at compile time, use it to lock the
exec-path and not allow the user to change it.
2002-07-02 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), keyserver.c (keyserver_refresh):
Maintain and use the original keyserver URI for cosmetics rather
than trying to recreate it when needed.
* mainproc.c (check_sig_and_print): Properly disregard expired
uids. Make sure that the first uid listed is a real uid and not
an attribute (attributes should only be listed in the "aka"
section). When there are no valid textual userids, try for an
invalid textual userid before using any attribute uid.
2002-07-01 David Shaw <dshaw@jabberwocky.com>
* options.skel: Fix a few typos, clarify "group", and remove
sample photo viewers for Win32 since they are the defaults now.
* parse-packet.c (make_attribute_uidname), keylist.c
(dump_attribs): Fix two typecast warnings.
* packet.h, build-packet.c (build_attribute_subpkt), exec.c
(expand_args), mkdtemp.c (mkdtemp), photoid.c
(parse_image_header): Fix some signedness compiler warnings.
2002-07-01 Werner Koch <wk@gnupg.org>
* photoid.c (get_default_photo_command): Also use __MINGW32__
instead of HAVE_DOSISH_SYSTEM.
* encode.c (encode_symmetric): Do not use the new encryption code.
2002-06-30 Werner Koch <wk@gnupg.org>
* photoid.c: Use __MINGW32__ to include windows because
HAVE_DOSISH_SYSTEM is also set for OS/2 and plain DOS. Provide
constant missing in older mingw installations.
2002-06-21 Stefan Bellon <sbellon@sbellon.de>
* g10.c [__riscos__]: Moved RISC OS specific stuff to util/riscos.c
and include/util.h.
* gpgv.c [__riscos__]: Likewise.
2002-06-20 David Shaw <dshaw@jabberwocky.com>
* keydb.h, pkclist.c (select_algo_from_prefs): Allow passing a
suggested algorithm which will be used if available.
* encode.c (encode_crypt, encrypt_filter), sign.c (sign_file): Use
new select_algo_from_prefs feature to check if forcing an
algorithm would violate the recipient preferences.
* photoid.c (get_default_photo_command, show_photos): Use
different default viewers on different platforms. Currently we
have Win 9x, Win NT (2k, xp), Mac OSX, RISC OS, and "everybody
else". These are #ifdefs as much as possible to avoid clutter.
* g10.c (strusage, build_list), keyedit.c (show_prefs), main.h,
misc.c (compress_algo_to_string, check_compress_algo), pkclist.c
(algo_available), keygen.c (keygen_set_std_prefs): New
algo_to_string and check functions for compress algorithms.
2002-06-20 Werner Koch <wk@gnupg.org>
* misc.c (setsysinfo): Removed a #warning for Alpha's uniligedn
trap disabling - it is quite possible that this is a debug relict.
2002-06-20 Stefan Bellon <sbellon@sbellon.de>
* g10.c [__riscos__]: Added image file system feature.
* gpgv.c [__riscos__]: Added image file system feature.
* photoid.c (show_photos) [__riscos__]: Set RISC OS filetype of
photo id according to MIME type.
2002-06-19 David Shaw <dshaw@jabberwocky.com>
* hkp.c (parse_hkp_index): Don't leak memory when failing out of a
bad HKP keyserver.
* g10.c (add_notation_data): Relax slightly the rules as to what
can go into a notation name - 2440 allows "@", for example.
2002-06-17 David Shaw <dshaw@jabberwocky.com>
* import.c (clean_subkeys, import_one): Only allow at most 1
binding sig and at most 1 revocation sig on a subkey, as per
2440:11.1.
* hkp.c (parse_hkp_index, hkp_search): Error if the keyserver
returns an unparseable HKP response.
2002-06-15 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (show_key_with_all_names), keylist.c
(list_keyblock_print): Show "[expired]" before expired uids.
* keyedit.c (show_key_with_all_names_colon), mainproc.c
(list_node), keylist.c (list_keyblock_colon): Show flag 'e' for
expired user ids. Use "uat" for user attribute packets instead of
"uid". Also use '<count> <length>' rather than the fake user id
string on attributes.
* keygen.c (keygen_add_revkey): Remove unused code.
* misc.c (check_permissions): Check directory permissions
properly - they are not special files.
* pkclist.c (expand_id, expand_group, build_pk_list): When
expanding groups before building a pk list, inherit flags from the
original pre-expanded string.
* pubkey-enc.c (is_algo_in_prefs): Don't use prefs from expired
uids.
2002-06-14 David Shaw <dshaw@jabberwocky.com>
* free-packet.c (copy_signature): Properly copy a signature that
carries a revocation key on it.
* pkclist.c (expand_id, expand_group, build_pk_list): Groups now
work properly when used in the "Enter the user ID" prompt.
2002-06-14 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (show_key_with_all_names): Display warning if a user
tries to show prefs on a v3 key with a v3 selfsig.
* kbnode.c (dump_kbnode): Show if a uid is expired.
* import.c (merge_blocks, import_revoke_cert): Show user ID
receiving a revocation certificate.
* free-packet.c (cmp_user_ids): Properly compare attribute ids.
* pkclist.c (expand_groups): Maintain the strlist flags while
expanding. Members of an expansion inherit their flags from the
expansion key.
* options.h, cipher.c (write_header), g10.c (main), keygen.c
(keygen_set_std_prefs): remove the personal_mdc flag. It no
longer serves a purpose now that the personal preference lists are
split into cipher/digest/zip.
2002-06-14 Timo Schulz <ts@winpt.org>
* skclist.c (is_insecure): Implemented.
2002-06-12 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_spawn): Properly handle PROGRAM responses
when they have a CRLF ending. Noted by Keith Ray.
* keyserver.c (keyserver_spawn): Handle CRLF endings from
keyserver helpers. Also don't leak the last line worth of memory
from the keyserver response.
* main.h, misc.c (deprecated_warning): New function to warn about
deprecated options and commands.
* g10.c (main), keyserver-internal.h, keyserver.c
(parse_keyserver_uri): Use new deprecated function to warn about
honor-http-proxy, auto-key-retrieve, and x-broken-hkp.
2002-06-11 David Shaw <dshaw@jabberwocky.com>
* Makefile.am: link gpg with NETLIBS for the built-in HKP access.
2002-06-10 David Shaw <dshaw@jabberwocky.com>
* options.h, keyserver.c (keyserver_opts), g10.c (main): New
keyserver option "include-subkeys". This feature already existed,
but now can be turned off. It defaults to on.
* options.h, keyserver.c (parse_keyserver_options,
keyserver_spawn): There are now enough options to justify making a
structure for the keyserver options rather than a page of
if-then-else-if-then-etc.
* getkey.c (merge_keys_and_selfsig, merge_selfsigs_main): Fix bug
in calculating key expiration dates.
2002-06-09 David Shaw <dshaw@jabberwocky.com>
* keydb.h, getkey.c (get_user_id_native), import.c (import_one):
Display user ID while importing a key. Note this applies to both
--import and keyserver --recv-keys.
* exec.c (exec_finish): Log unnatural exit (core dump, killed
manually, etc) for fork/exec/pipe child processes.
2002-06-08 Timo Schulz <ts@winpt.org>
* encode.c (encode_symmetric): Disable the compat flag
when the expert mode is enabled.
2002-06-07 David Shaw <dshaw@jabberwocky.com>
* options.skel, options.h, main.h, keydb.h, pkclist.c
(build_pk_list, expand_groups), g10.c (main, add_group): Add new
"group" command to allow one name to expand into multiple keys.
For simplicity, and to avoid potential loops, we only expand once
- you can't make an alias that points to an alias.
* main.h, g10.c (main), keygen.c (build_personal_digest_list):
Simplify the default digest list - there is really no need for the
other hashes since they will never be used after SHA-1 in the
list.
* options.skel, options.h, g10.c (main), hkp.c (hkp_ask_import,
hkp_export, hkp_search), keyserver.c (parse_keyserver_options,
parse_keyserver_uri, keyserver_work, keyserver_refresh): Make the
"x-broken-hkp" keyserver scheme into keyserver-option
"broken-http-proxy". Move honor_http_proxy into
keyserver_options. Canonicalize the three variations of "hkp",
"x-hkp", and "x-broken-hkp" into "hkp".
2002-06-07 Stefan Bellon <sbellon@sbellon.de>
* g10.c [__riscos__]: Added --attribute-file to do the same as
--attribute-fd, but with a filename not a fd as argument.
Added magic symbol for RISC OS to use different memory management.
* gpgv.c [__riscos__]: Added magic symbol for RISC OS to use
different memory management.
2002-06-06 David Shaw <dshaw@jabberwocky.com>
* main.h, g10.c (main), keygen.c (build_personal_digest_list): Put
in a default digest preference list consisting of SHA-1, followed
by every other installed digest except MD5. Note this is the same
as having no digest preference at all except for SHA-1 being
favored.
* options.h, g10.c (main), keygen.c (keygen_set_std_prefs),
pkclist.c (select_algo_from_prefs): Split
--personal-preference-list into three:
--personal-{cipher|digest|compress}-preferences. This allows a
user to set one without affecting another (i.e. setting only a
digest pref doesn't imply an empty cipher pref).
* exec.c (exec_read): This is a safer way of guessing the return
value of system(). Noted by Stefan Bellon.
2002-06-05 David Shaw <dshaw@jabberwocky.com>
* hkp.c (parse_hkp_index): Be more robust with keyservers
returning very unparseable responses.
* exec.c (exec_read): Catch and display an error when the remote
process exits unnaturally (i.e. segfault) so the user knows what
happened. Also fix exec_write stub which has a different number
of arguments now.
2002-06-05 Timo Schulz <ts@winpt.org>
* encode.c (encode_simple): Ignore the new mode for RFC1991.
* mainproc.c (symkey_decrypt_sesskey): Better check for weird
keysizes.
2002-06-05 Timo Schulz <ts@winpt.org>
* encode.c (encode_sesskey): New.
(encode_simple): Use it here. But by default we use the compat
mode which supress to generate encrypted session keys.
2002-06-05 Timo Schulz <ts@winpt.org>
* mainproc.c (symkey_decrypt_sesskey): New.
(proc_symkey_enc): Support for encrypted session keys.
2002-06-04 David Shaw <dshaw@jabberwocky.com>
* sign.c (hash_for, sign_file): When encrypting and signing at the
same time, consult the various hash prefs to pick a hash algorithm
to use. Pass in a 160-bit hint if any of the signing keys are
DSA.
* keydb.h, pkclist.c (select_algo_from_prefs, algo_available):
Pass a "hints" opaque pointer in to let the caller give hints as
to what algorithms would be acceptable. The only current hint is
for PREFTYPE_HASH to require a 160-bit hash for DSA. Change all
callers in encode.c (encode_crypt, encrypt_filter) and sign.c
(sign_file). If we settle on MD5 as the best algorithm based
solely on recepient keys and SHA1 is also a possibility, use SHA1
unless the user intentionally chose MD5. This is as per 2440:13.
* exec.c (make_tempdir): Fix duplicated filename problem.
2002-06-03 David Shaw <dshaw@jabberwocky.com>
* packet.h, parse-packet.c (enum_sig_subpkt): Report back from
enum_sig_subpkt when a subpacket is critical and change all
callers in keylist.c (show_policy_url, show_notation), mainproc.c
(print_notation_data), and pkclist.c (do_show_revocation_reason).
* keylist.c (show_policy_url, show_notation): Display if the
policy or notation is critical.
2002-06-03 David Shaw <dshaw@jabberwocky.com>
* main.h, g10.c (main), keylist.c (dump_attribs, set_attrib_fd,
list_keyblock_print, list_keyblock_colon), status.h, status.c
(get_status_string): New --attribute-fd feature to dump the
contents of attribute subpackets for frontends. If --status-fd is
also used, then a new status tag ATTRIBUTE is provided for each
subpacket.
* packet.h, getkey.c (fixup_uidnode, merge_selfsigs_main,
merge_selfsigs_subkey), parse-packet.c (setup_user_id): Keep track
of the expiration time of a user ID, and while we're at it, use
the expired flag from the selfsig rather than reparsing the
SIG_EXPIRE subpacket.
* photoid.c (generate_photo_id): When adding a new photo ID,
showing the photo for confirmation is not safe when noninteractive
since the "user" may not be able to dismiss a viewer window.
Noted by Timo Schulz.
2002-06-03 David Shaw <dshaw@jabberwocky.com>
* options.skel: Sample photo viewers for Win32.
* misc.c (pct_expando): Use the seckey for %k/%K if the pubkey is
not available.
* photoid.h, photoid.c (show_photos): Include the seckey in case a
user tries to view a photo on a secret key, and change all callers
in keyedit.c (menu_showphoto), keylist.c (list_keyblock_print),
and photoid.c (generate_photo_id).
2002-06-02 David Shaw <dshaw@jabberwocky.com>
* photoid.c (show_photos): Work properly when not called with a
public key.
2002-05-31 David Shaw <dshaw@jabberwocky.com>
* sign.c (mk_notation_and_policy): Free unneeded buffer.
* hkp.c (parse_hkp_index): Properly handle the '&' character
(i.e. "&amp;") in HKP responses.
* getkey.c (merge_selfsigs_main): Fix reversed expiration time
check with self-sigs.
* keyedit.c (sign_uids): When making a new self-sig on a v3 key,
make a v3 self-sig unless it is currently a v3 self-sig being
promoted to v4.
2002-05-31 Timo Schulz <ts@winpt.org>
* pkclist.c (do_show_revocation_reason): Don't use capital
letters for non-interactive output.
(show_revocation_reason): Now it is global.
* pubkey-enc.c (get_it): Show if the key has been revoked.
2002-05-30 David Shaw <dshaw@jabberwocky.com>
* sign.c (write_signature_packets, sign_file, clearsign_file,
sign_symencrypt_file): Make a v4 signature if a policy URL or
notation is set, unless v3 sigs are forced via rfc1991 or
force-v3-sigs. Also remove some doubled code and clarify an error
message (we don't sign in PGP2 mode - just detach-sign).
* parse-packet.c (parse_one_sig_subpkt): Add KS_FLAGS to the "any
size" section.
2002-05-29 David Shaw <dshaw@jabberwocky.com>
* keygen.c (keygen_set_std_prefs, add_feature_mdc): Use "mdc" and
"no-mdc" in the prefs string to allow switching on and off the MDC
feature. This is needed to properly export a key from GnuPG for
use on PGP which does not support MDC - without this, MDC-capable
implementations will still try and generate MDCs which will break
PGP.
* keygen.c (keygen_get_std_prefs): Show "[mdc]" in prefs string if
it is enabled.
* options.h, g10.c (main), cipher.c (write_header), keygen.c
(keygen_set_std_prefs): For consistency, allow the user to specify
mdc/no-mdc in the --personal-preference-list. If disabled, it
acts just like --disable-mdc.
2002-05-29 David Shaw <dshaw@jabberwocky.com>
* options.h, exec.c: Add some debugging info, using the 1024 debug
flag.
* exec.c (win_system): New system()-like function for win32 that
does not return until the child process terminates. Of course,
this doesn't help if the process itself exits before it is
finished.
2002-05-29 Werner Koch <wk@gnupg.org>
* encode.c (encode_simple): Intialize PKT when --no-literal is used.
* keyedit.c (show_key_with_all_names_colon): Renamed the record
for revocation keys to "rvk".
2002-05-27 Werner Koch <wk@gnupg.org>
* keyedit.c (show_key_with_all_names_colon): New.
(show_key_with_all_names): Divert to new function when required.
Sanitize printing of revoker name.
2002-05-27 David Shaw <dshaw@jabberwocky.com>
* build-packet.c (build_sig_subpkt): Handle setting sig flags for
certain subpacket types (notation, policy url, exportable,
revocable). keyedit.c (sign_mk_attrib): Flags no longer need to
be set here.
* packet.h, parse-packet.c (parse_one_sig_subpkt), build-packet.c
(build_sig_subpkt): Call parse_one_sig_subpkt to sanity check
buffer lengths before building a sig subpacket.
2002-05-26 David Shaw <dshaw@jabberwocky.com>
* sign.c (mk_notation_and_policy): Include secret key to enable %s
expandos, and pass notations through pct_expando as well.
* main.h, misc.c (pct_expando): Add %s and %S expandos for
signer's keyid.
2002-05-25 David Shaw <dshaw@jabberwocky.com>
* g10.c (strusage, build_list): Add compress algorithms to
--version list. Show algorithm numbers when --verbose --version
is done.
2002-05-22 David Shaw <dshaw@jabberwocky.com>
* options.h, main.h, keygen.c (keygen_set_set_prefs,
keygen_get_std_prefs, keygen_upd_std_prefs), keyedit.c
(keyedit_menu), g10.c (main), pkclist.c (select_algo_from_prefs):
Add --personal-preference-list which allows the user to factor in
their own preferred algorithms when the preference lists are
consulted. Obviously, this does not let the user violate a
recepient's preferences (and the RFC) - this only influences the
ranking of the agreed-on (and available) algorithms from the
recepients. Suggested by David Hollenberg.
* options.h, keygen.c (keygen_set_std_prefs), g10.c (main): Rename
--preference-list to --default-preference-list (as that is what it
really is), and make it a true default in that if the user selects
"default" they get this list and not the compiled-in list.
2002-05-22 Werner Koch <wk@gnupg.org>
* g10.c (main): Add missing LF in a info printout and made it
translatable. Noted by Michael Tokarev.
2002-05-21 Werner Koch <wk@gnupg.org>
* g10.c (main): Removed the undef of USE_SHM_COPROCESSING which
was erroneously introduced on 2002-01-09.
* signal.c (got_fatal_signal): Don't write the Nul to stderr.
Reported by David Hollenberg.
2002-05-18 David Shaw <dshaw@jabberwocky.com>
* main.h, g10.c (main), revoke.c (gen_desig_revoke): Generate a
designated revocation via --desig-revoke
* keyedit.c (keyedit_menu, menu_addrevoker): New "addrevoker"
command to add a designated revoker to a key.
2002-05-17 David Shaw <dshaw@jabberwocky.com>
* gpgv.c: Add stub for get_ownertrust().
* g10.c (main): --allow-freeform-uid should be implied by
OpenPGP. Add --no-allow-freeform-uid.
* keyedit.c (sign_uids): Issue a warning when signing a
non-selfsigned uid.
* getkey.c (merge_selfsigs_main): If a key has no selfsigs, and
allow-non-selfsigned-uid is not set, still try and make the key
valid by checking all uids for a signature from an ultimately
trusted key.
2002-05-16 David Shaw <dshaw@jabberwocky.com>
* main.h, keygen.c (keygen_add_revkey): Add revocation key
subpackets to a signature (callable by
make_keysig_packet). (write_direct_sig): Write a 1F direct key
signature. (parse_revocation_key): Parse a string in
algo:fpr:sensitive format into a revocation
key. (get_parameter_revkey, do_generate_keypair): Call above
functions when prompted from a batch key generation file.
* build-packet.c (build_sig_subpkt): Allow multiple revocation key
subpackets in a single sig.
* keydb.h, getkey.c (get_seckey_byfprint): Same as
get_pubkey_byfprint, except for secret keys. We only know the
fingerprint of a revocation key, so this is needed to retrieve the
secret key needed to issue a revokation.
* packet.h, parse-packet.c (parse_signature, parse_revkeys): Split
revkey parsing off into a new function that can be used to reparse
after manipulating the revkey list.
* sign.c (make_keysig_packet): Ability to make 1F direct key
signatures.
2002-05-15 David Shaw <dshaw@jabberwocky.com>
* options.skel: keyserver.pgp.com is gone, so list pgp.surfnet.nl
as a sample LDAP server instead.
* getkey.c (merge_selfsigs_main): Properly handle multiple
revocation keys in a single packet. Properly handle revocation
keys that are in out-of-order packets. Remove duplicates in
revocation key list.
2002-05-14 Timo Schulz <ts@winpt.org>
* exec.c (make_tempdir) [MINGW32]: Added missing '\'.
2002-05-14 Stefan Bellon <sbellon@sbellon.de>
* exec.c (make_tempdir): Make use of EXTSEP_S instead of hardcoded
dot as extension separator.
2002-05-13 David Shaw <dshaw@jabberwocky.com>
* photoid.c (show_photos): Use the long keyid as the filename for
the photo. Use the short keyid as the filename on 8.3 systems.
* exec.h, exec.c (make_tempdir, exec_write, exec_finish): Allow
caller to specify filename. This should make things easier on
windows and macs where the file extension is required, but a whole
filename is even better.
* keyedit.c (show_key_with_all_names, show_prefs): Show proper
prefs for a v4 key uid with no selfsig at all.
* misc.c (check_permissions): Don't check permissions on
non-normal files (pipes, character devices, etc.)
2002-05-11 Werner Koch <wk@gnupg.org>
* mainproc.c (proc_symkey_enc): Avoid segv in case the parser
encountered an invalid packet.
* keyserver.c (keyserver_export): Get confirmation before sending
all keys.
2002-05-10 Stefan Bellon <sbellon@sbellon.de>
* g10.c, hkp.c, keyedit.c, keyserver.c: Replaced all occurrances
of strcasecmp with ascii_strcasecmp and all occurrances of
strncasecmp with ascii_memcasecmp.
2002-05-10 David Shaw <dshaw@jabberwocky.com>
* packet.h, getkey.c (fixup_uidnode), keyedit.c (show_prefs): Show
assumed prefs for hash and compression as well as the cipher pref.
Show assumed prefs if there are no prefs at all on a v4
self-signed key.
* options.h, g10.c (main), sign.c (make_keysig_packet): New
--cert-digest-algo function to override the default key signing
hash algorithm.
2002-05-09 David Shaw <dshaw@jabberwocky.com>
* getkey.c (merge_selfsigs_main): Make sure the revocation key
list starts clean as this function may be called more than once
(e.g. from functions in --edit).
* g10.c, encode.c (encode_crypt), sign.c (sign_file,
sign_symencrypt_file): Make --compress-algo work like the
documentation says. It should be like --cipher-algo and
--digest-algo in that it can override the preferences calculation
and impose the setting the user wants. No --compress-algo setting
allows the usual preferences calculation to take place.
* main.h, compress.c (compress_filter): use new
DEFAULT_COMPRESS_ALGO define, and add a sanity check for compress
algo value.
2002-05-08 David Shaw <dshaw@jabberwocky.com>
* pkclist.c (select_algo_from_prefs): There is an assumed
compression preference for uncompressed data.
2002-05-07 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), getkey.c (finish_lookup), pkclist.c
(algo_available): --pgp7, identical to --pgp6 except that it
permits a few algorithms that PGP 7 added: AES128, AES192, AES256,
and TWOFISH. Any more of these --pgpX flags, and it'll be time to
start looking at a generic --emulate-pgp X option.
* export.c (do_export_stream): Warn the user when exporting a
secret key if it or any of its secret subkeys are protected with
SHA1 while simple_sk_checksum is set.
* parse-packet.c (parse_key): Show when the SHA1 protection is
used in --list-packets.
* options.h, build-packet.c (do_comment), g10.c (main): Rename
--no-comment as --sk-comments/--no-sk-comments (--no-comment still
works) and make the default be --no-sk-comments.
2002-05-07 Werner Koch <wk@gnupg.org>
* keygen.c (get_parameter_algo): Never allow generation of the
deprecated RSA-E or RSA-S flavors of PGP RSA.
(ask_algo): Allow generation of RSA sign and encrypt in expert
mode. Don't allow ElGamal S+E unless in expert mode.
* helptext.c: Added entry keygen.algo.rsa_se.
2002-05-07 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (sign_uids): If --expert is set, allow re-signing a
uid to promote a v3 self-sig to a v4 one. This essentially
deletes the old v3 self-sig and replaces it with a v4 one.
* packet.h, parse-packet.c (parse_key), getkey.c
(merge_keys_and_selfsig, merge_selfsigs_main): a v3 key with a v4
self-sig must never let the v4 self-sig express a key expiration
time that extends beyond the original v3 expiration time.
2002-05-06 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (sign_uids): When making a self-signature via "sign"
don't ask about sig level or expiration, and include the usual
preferences and such for v4 self-sigs. (menu_set_preferences):
Convert uids from UTF8 to native before printing.
* keyedit.c (sign_uids): Convert uids from UTF8 to native before
printing. (menu_set_primary_uid): Show error if the user tries to
make a uid with a v3 self-sig primary.
2002-05-05 David Shaw <dshaw@jabberwocky.com>
* import.c (import_one): When merging with a key we already have,
don't let a key conflict (same keyid but different key) stop the
import: just skip the bad key and continue.
* exec.c (make_tempdir): Under Win32, don't try environment
variables for temp directories - GetTempDir tries environment
variables internally, and it's better not to second-guess it in
case MS adds some sort of temp dir handling to Windows at some
point.
2002-05-05 Timo Schulz <ts@winpt.org>
* mainproc.c (proc_symkey_enc): Don't ask for a passphrase
in the list only mode.
2002-05-05 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_refresh): --refresh-keys implies
--merge-only so as not to import keys with keyids that match the
ones being refreshed. Noted by Florian Weimer.
2002-05-04 Stefan Bellon <sbellon@sbellon.de>
* free-packet.c (copy_public_key): Don't call m_alloc(0), therefore
added consistency check for revkey and numrefkeys.
* getkey.c (check_revocation_keys): Added consistency check for
revkey and numrefkeys.
* keyedit.c (show_key_with_all_names): Likewise.
2002-05-03 David Shaw <dshaw@jabberwocky.com>
* photoid.c: Provide default image viewer for Win32.
* misc.c (pct_expando): %t means extension, not name ("jpg", not
"jpeg").
* keyserver.c (keyserver_spawn), photoid.c (show_photos), exec.h,
exec.c: Allow the caller to determine the temp file extension when
starting an exec_write and change all callers.
* keyedit.c (sign_uids): Nonrevocable key signatures cause an
automatic promotion to v4.
* exec.c: Provide stubs for exec_ functions when NO_EXEC is
defined.
2002-05-02 David Shaw <dshaw@jabberwocky.com>
* photoid.h, photoid.c (parse_image_header, image_type_to_string):
Useful functions to return data about an image.
* packet.h, parse-packet.c (make_attribute_uidname,
parse_attribute_subpkts, parse_attribute), photoid.h, photoid.c
(show_photos): Handle multiple images in a single attribute
packet.
* main.h, misc.c (pct_expando), sign.c (mk_notation_and_policy),
photoid.c (show_photos): Simpler expando code that does not
require using compile-time string sizes. Call
image_type_to_string to get image strings (i.e. "jpg",
"image/jpeg"). Change all callers.
* keyedit.c (menu_showphoto), keylist.c (list_keyblock_print):
Allow viewing multiple images within a single attribute packet.
* gpgv.c: Various stubs for link happiness.
2002-05-02 David Shaw <dshaw@jabberwocky.com>
* build-packet.c (build_sig_subpkt), keyedit.c (sign_uids),
options.h, sign.c (mk_notation_and_policy), g10.c (main,
add_notation_data, add_policy_url (new), check_policy_url
(removed)): Allow multiple policy URLs on a given signature.
Split "--notation-data" into "--cert-notation" and
"--sig-notation" so the user can set different policies for key
and data signing. For backwards compatibility, "--notation-data"
sets both, as before.
2002-05-02 Werner Koch <wk@gnupg.org>
* options.skel: Removed the comment on trusted-keys because this
option is now deprecated.
2002-05-01 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_adduid): 2440bis04 says that multiple attribute
packets on a given key are legal.
* keyserver.c (keyserver_refresh): the fake v3 keyid hack applies
to "mailto" URLs as well since they are also served by pksd.
2002-04-29 Werner Koch <wk@gnupg.org>
Added a copyright year for files changed this year.
2002-04-25 Werner Koch <wk@gnupg.org>
* g10.c, options.h: New options --display, --ttyname, --ttytype,
--lc-ctype, --lc-messages to be used with future versions of the
gpg-agent.
* passphrase.c (agent_send_option,agent_send_all_options): New.
(agent_open): Send options to the agent.
* trustdb.c (update_ownertrust, clear_ownertrust): Do an explicit
do_sync because revalidation_mark does it only if when the
timestamp actually changes.
2002-04-23 David Shaw <dshaw@jabberwocky.com>
* main.h, keygen.c (do_generate_keypair), keylist.c
(print_signature_stats, list_all, list_one, list_keyblock,
list_keyblock_print, list_keyblock_colon): After generating a new
key, show the key information (name, keyid, fingerprint, etc.)
Also do not print uncheckable signatures (missing key..) in
--check-sigs. Print statistics (N missing keys, etc.) after
--check-sigs.
* keyedit.c (sign_uids): When signing a key with an expiration
date on it, the "Do you want your signature to expire at the same
time?" question should default to YES.
2002-04-22 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (parse_plaintext), packet.h, plaintext.c
(handle_plaintext): Fix bug in handling literal packets with
zero-length data (no data was being confused with partial body
length).
* misc.c (pct_expando), options.skel: %t means extension ("jpg").
%T means MIME type ("image/jpeg").
* import.c (import_one): Only trigger trust update if the keyring
is actually changed.
* export.c (do_export_stream): Missing a m_free.
2002-04-22 Stefan Bellon <sbellon@sbellon.de>
* keyid.c (expirestr_from_sk, expirestr_from_sig): Added _() to
string constant.
* exec.c (make_tempdir) [__riscos__]: Better placement of
temporary file.
2002-04-20 David Shaw <dshaw@jabberwocky.com>
* keygen.c (generate_subkeypair): 2440bis04 adds that creating
subkeys on v3 keys is a MUST NOT.
* getkey.c (finish_lookup): The --pgp6 "use the primary key"
behavior should only apply while data signing and not encryption.
Noted by Roger Sondermann.
2002-04-19 Werner Koch <wk@gnupg.org>
* keygen.c (keygen_set_std_prefs): Put back 3DES because the RFC
says it is good form to do so.
2002-04-19 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_deluid): Only cause a trust update if we delete
a non-revoked user id.
* hkp.c (hkp_ask_import), keyserver.c (parse_keyserver_options,
keyserver_spawn), options.h: Remove fast-import keyserver option
(no longer meaningful).
* g10.c (main), keyedit.c (sign_uids), options.h: Change
--default-check-level to --default-cert-check-level as it makes
clear what it operates on.
* g10.c (main): --pgp6 also implies --no-ask-sig-expire.
* delkey.c (do_delete_key): Comment.
* keyedit.c (sign_uids, keyedit_menu, menu_deluid, menu_delsig,
menu_expire, menu_revsig, menu_revkey): Only force a trustdb check
if we did something that changes it.
* g10.c: add "--auto-check-trustdb" to override a
"--no-auto-check-trustdb"
2002-04-19 Werner Koch <wk@gnupg.org>
* tdbio.c (tdbio_write_nextcheck): Return a status whether the
stamp was actually changed.
* trustdb.c (revalidation_mark): Sync the changes. Removed the
sync operation done by its callers.
(get_validity): Add logic for maintaining a pending_check flag.
(clear_ownertrust): New.
* keyedit.c (sign_uids): Don't call revalidation_mark depending on
primary_pk.
(keyedit_menu): Call revalidation_mark after "trust".
(show_key_with_all_names): Print a warning on the wrong listed key
validity.
* delkey.c (do_delete_key): Clear the owenertrust information when
deleting a public key.
2002-04-18 Werner Koch <wk@gnupg.org>
* seskey.c (encode_md_value): Print an error message if a wrong
digest algorithm is used with DSA. Changed all callers to cope
with a NULL return. Problem noted by Imad R. Faiad.
2002-04-18 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (mark_usable_uid_certs): Properly handle nonrevocable
signatures that can expire. In short, the only thing that can
override an unexpired nonrevocable signature is another unexpired
nonrevocable signature.
* getkey.c (finish_lookup): Always use primary signing key for
signatures when --pgp6 is on since pgp6 and 7 do not understand
signatures made by signing subkeys.
2002-04-18 Werner Koch <wk@gnupg.org>
* trustdb.c (validate_keys): Never schedule a nextcheck into the
past.
(validate_key_list): New arg curtime use it to set next_expire.
(validate_one_keyblock): Take the current time from the caller.
(clear_validity, reset_unconnected_keys): New.
(validate_keys): Reset all unconnected keys.
* getkey.c (premerge_public_with_secret): Fixed 0x12345678! syntax
for use with secret keys.
(lookup): Advance the searchmode after a search FIRST.
* seckey-cert.c (do_check): Always calculate the old checksum for
use after unprotection.
* g10.c, options.skel: New option --no-escape-from. Made
--escape-from and --force-v3-sigs the default and removed them
from the options skeleton.
2002-04-16 Werner Koch <wk@gnupg.org>
* parse-packet.c (parse_key): Support a SHA1 checksum as per
draft-rfc2440-bis04.
* packet.h (PKT_secret_key): Add field sha1chk.
* seckey-cert.c (do_check): Check the SHA1 checksum
(protect_secret_key): And create it.
* build-packet.c (do_secret_key): Mark it as sha-1 protected.
* g10.c, options.h: New option --simple-sk-checksum.
2002-04-13 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (parse_signature): Minor fix - signatures should
expire at their expiration time and not one second later.
* keygen.c (proc_parameter_file): Allow specifying preferences
string (i.e. "s5 s2 z1 z2", etc) in a batchmode key generation
file.
* keyedit.c (keyedit_menu): Print standard error message when
signing a revoked key (no new translation).
* getkey.c (merge_selfsigs): Get the default set of key prefs from
the real (not attribute) primary uid.
2002-04-12 David Shaw <dshaw@jabberwocky.com>
* pkclist.c (build_pk_list): Fix bug that allowed a key to be
selected twice in batch mode if one instance was the default
recipient and the other was an encrypt-to. Noted by Stefan
Bellon.
* parse-packet.c (dump_sig_subpkt): Show data in trust and regexp
sig subpackets.
* keyedit.c (keyedit_menu): Use new function real_uids_left to
prevent deleting the last real (i.e. non-attribute) uid. Again,
according to the attribute draft. (menu_showphoto): Make another
string translatable.
2002-04-11 David Shaw <dshaw@jabberwocky.com>
* build-packet.c (build_sig_subpkt): Delete subpackets from both
hashed and unhashed area on update. (find_subpkt): No longer
needed.
* keyedit.c (sign_uids): With --pgp2 on, refuse to sign a v3 key
with a v4 signature. As usual, --expert overrides. Try to tweak
some strings to a closer match so they can all be translated in
one place. Use different helptext keys to allow different help
text for different questions.
* keygen.c (keygen_upd_std_prefs): Remove preferences from both
hashed and unhashed areas if they are not going to be used.
2002-04-10 David Shaw <dshaw@jabberwocky.com>
* misc.c (pct_expando), options.skel: Use %t to indicate type of a
photo ID (in this version, it's always "jpeg"). Also tweak string
expansion loop to minimize reallocs.
* mainproc.c (do_check_sig): Variable type fix.
* keyedit.c (menu_set_primary_uid): Differentiate between true
user IDs and attribute user IDs when making one of them primary.
That is, if we are making a user ID primary, we alter user IDs.
If we are making an attribute packet primary, we alter attribute
packets. This matches the language in the latest attribute packet
draft.
* keyedit.c (sign_uids): No need for the empty string hack.
* getkey.c (fixup_uidnode): Only accept preferences from the
hashed segment of the self-sig.
2002-04-10 Werner Koch <wk@gnupg.org>
* tdbio.c (migrate_from_v2): Fixed the offset to read the old
ownertrust value and only add entries to the table if we really
have a value.
2002-04-08 David Shaw <dshaw@jabberwocky.com>
* status.h, status.c (get_status_string): Add KEYEXPIRED, EXPSIG,
and EXPKEYSIG. Add "deprecated-use-keyexpired-instead" to
SIGEXPIRED.
* sig-check.c (do_check): Start transition from SIGEXPIRED to
KEYEXPIRED, since the actual event is signature verification by an
expired key and not an expired signature. (do_signature_check,
packet.h): Rename as signature_check2, make public, and change all
callers.
* mainproc.c (check_sig_and_print, do_check_sig): Use status
EXPSIG for an expired, but good, signature. Add the expiration
time (or 0) to the VALIDSIG status line. Use status KEYEXPSIG for
a good signature from an expired key.
* g10.c (main): remove checks for no arguments now that argparse
does it.
2002-04-06 Werner Koch <wk@gnupg.org>
* keyring.c (keyring_get_keyblock): Disable the keylist mode here.
* encode.c (encode_simple, encode_crypt): Only test on compressed
files if a compress level was not explicity set.
* keygen.c (keygen_set_std_prefs): Removed Blowfish and Twofish
from the list of default preferences, swapped the preferences of
RMD160 and SHA1. Don't include a preference to 3DES unless the
IDEA kludge gets used.
* free-packet.c (free_packet): call free_encrypted also for
PKT_ENCRYPTED_MDC.
* compress.c (release_context): New.
(handle_compressed): Allocate the context and setup a closure to
release the context. This is required because there is no
guarabntee that the filter gets popped from the chain at the end
of the function. Problem noted by Timo and probably also the
cause for a couple of other reports.
(compress_filter): Use the release function if set.
* tdbio.c [__CYGWIN32__]: Don't rename ftruncate. Noted by
Disastry.
* parse-packet.c (parse_signature): Put parens around a bit test.
* exec.c (make_tempdir): Double backslash for TMP directory
creation under Windows. Better strlen the DIRSEP_S constants for
allocation measurements.
* decrypt.c (decrypt_messages): Release the passphrase aquired
by get_last_passphrase.
2002-04-02 Werner Koch <wk@gnupg.org>
* Makefile.am (EXTRA_DIST): Removed OPTIONS an pubring.asc - they
are no longer of any use.
2002-04-03 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (parse_keyserver_options): fix auto-key-retrieve to
actually work as a keyserver-option (noted by Roger Sondermann).
* keylist.c (reorder_keyblock): do not reorder the primary
attribute packet - the first user ID must be a genuine one.
2002-03-31 David Shaw <dshaw@jabberwocky.com>
* keylist.c (list_keyblock_colon): Fix ownertrust display with
--with-colons.
* keygen.c (generate_user_id), photoid.c (generate_photo_id):
Properly initialize the user ID refcount. A few more "y/n" ->
"y/N" in photoid.c.
* keyedit.c (ask_revoke_sig): Warn the user if they are about to
revoke an expired sig (not a problem, but they should know). Also
tweak a few prompts to change "y/n" to "y/N", which is how most
other prompts are written.
* keyserver.c (keyserver_search_prompt): Control-d escapes the
keyserver search prompt.
* pkclist.c (show_revocation_reason & callers): If a subkey is
considered revoked solely because the parent key is revoked, print
the revocation reason from the parent key.
* trustdb.c (get_validity): Allow revocation/expiration to apply
to a uid/key with no entry in the trustdb.
2002-03-29 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (printunquoted): unquote backslashes from keyserver
searches
* hkp.c (write_quoted): quote backslashes from keyserver searches
2002-03-26 Werner Koch <wk@gnupg.org>
* keygen.c (ask_keysize): Removed the warning for key sizes > 1536.
2002-03-25 Werner Koch <wk@gnupg.org>
* keyedit.c (sign_uids): Use 2 strings and not a %s so that
translations can be done the right way.
* helptext.c: Fixed small typo.
2002-03-23 David Shaw <dshaw@jabberwocky.com>
* import.c (append_uid, merge_sigs): it is okay to import
completely non-signed uids now (with --allow-non-selfsigned-uid).
* getkey.c (get_primary_uid, merge_selfsigs_main): do not choose
an attribute packet (i.e. photo) as primary uid. This prevents
oddities like "Good signature from [image of size 2671]". This is
still not perfect (one can still select an attribute packet as
primary in --edit), but is closer to the way the draft is going.
* g10.c (build_list): algorithms should include 110.
* g10.c (main): --pgp2 implies --no-ask-sig-expire and
--no-ask-cert-expire as those would cause a v4 sig/cert.
* armor.c (is_armor_header): be more lenient in what constitutes a
valid armor header (i.e. -----BEGIN blah blah-----) as some
Windows programs seem to add spaces at the end. --openpgp makes
it strict again.
2002-03-18 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_search_prompt): Properly handle a "no
keys found" case from the internal HKP code (external HKP is ok).
Also, make a COUNT -1 (i.e. streamed) keyserver response a little
more efficient.
* g10.c (main): Add --no-allow-non-selfsigned-uid
2002-03-17 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): --openpgp implies --allow-non-selfsigned-uid.
* getkey.c (merge_selfsigs_main): If none of the uids are primary
(because none are valid) then pick the first to be primary (but
still invalid). This is for cosmetics in case some display needs
to print a user ID from a non-selfsigned key. Also use
--allow-non-selfsigned-uid to make such a key valid and not
--always-trust. The key is *not* automatically trusted via
--allow-non-selfsigned-uid.
* mainproc.c (check_sig_and_print): Make sure non-selfsigned uids
print [uncertain] on verification even though one is primary now.
* getkey.c (merge_selfsigs): If the main key is not valid, then
neither are the subkeys.
* import.c (import_one): Allow --allow-non-selfsigned-uid to work
on completely unsigned keys. Print the uids in UTF8. Remove
mark_non_selfsigned_uids_valid().
* keyedit.c (show_key_with_all_names): Show revocation key as
UTF8.
* sign.c (clearsign_file): Allow --not-dash-escaped to work with
v3 keys.
2002-03-14 Werner Koch <wk@gnupg.org>
* main.h: Changed the default algorithms to CAST5 and SHA1.
2002-03-13 David Shaw <dshaw@jabberwocky.com>
* import.c (chk_self_sigs): Show which user ID a bad self-sig
(invald sig or unsupported public key algorithm) resides on.
* import.c (chk_self_sigs): any valid self-sig should mark a user
ID or subkey as valid - otherwise, an attacker could DoS the user
by inventing a bogus invalid self-signature.
2002-03-07 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): make a few more strings translatable.
* options.h, options.skel, g10.c (main), gpgv.c, mainproc.c
(check_sig_and_print), keyserver.c (parse_keyserver_options):
--auto-key-retrieve should really be a keyserver-option variable.
* import.c (revocation_present): new function to print a warning
if a key is imported that has been revoked by designated revoker,
but the designated revoker is not present to verify the
revocation. If keyserver-options auto-key-retrieve is set, try
and fetch the designated revoker from the keyserver.
* import.c (import_one): call revocation_present after importing a
new key. Note that this applies to --import, --recv-keys, and
--search-keys.
* keyserver-internal.h, keyserver.c (keyserver_import_fprint):
import via fingerprint (for revocation keys).
* keyserver.c (keyserver_import_keyid): much simpler
implementation now that we're using KEYDB_SEARCH_DESC internally.
2002-03-04 David Shaw <dshaw@jabberwocky.com>
* revoke.c (gen_revoke): do not prompt for revocation reason for
v3 revocations (unless force-v4-certs is on) since they wouldn't
be used anyway.
* keyedit.c (menu_revsig): show the status of the sigs
(exportable? revocable?) to the user before prompting for which
sig to revoke. Also, make sure that local signatures get local
revocations.
* keyedit.c (ask_revoke_sig): remind the user which sigs are
local.
* g10.c (main): Add "exec-path" variable to override PATH for
execing programs.
* export.c (do_export_stream): properly check return code from
classify_user_id to catch unclassifiable keys.
2002-03-03 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (parse_signature): variable type tweak for RISC
OS (from Stefan)
2002-02-28 David Shaw <dshaw@jabberwocky.com>
* getkey.c (check_revocation_keys): New function to check a
revocation against a list of potential revocation keys. Note the
loop-breaking code here. This is to prevent blowing up if A is
B's revocation key, while B is also A's. Note also that this is
written so that a revoked revoker can still issue revocations:
i.e. If A revokes B, but A is revoked, B is still revoked. I'm
not completely convinced this is the proper behavior, but it
matches how PGP does it. It does at least have the advantage of
much simpler code - my first version of this had lots of loop
maintaining code so you could chain revokers many levels deep and
if D was revoked, C was not, which meant that B was, and so on.
It was sort of scary, actually.
* getkey.c (merge_selfsigs_main): Add any revocation keys onto the
pk. This is particularly interesting since we normally only get
data from the most recent 1F signature, but you need multiple 1F
sigs to properly handle revocation keys (PGP does it this way, and
a revocation key could be marked "sensitive" and hence in a
different signature). Also, if a pk has a revocation key set,
check for revocation sigs that were not made by us - if made by a
valid revocation key, mark the pk revoked.
* packet.h, getkey.c (cache_public_key): do not cache key if
"dont_cache" is set. This allows the revocation key code to look
up a key and return information that may be inaccurate to prevent
loops without caching the fake data.
* packet.h, sig-check.c (do_signature_check): Record if a
signature was made by a revoked pk.
* packet.h, parse-packet.c (parse_one_sig_subpkt,
can_handle_critical, parse_signature): Get revocation key
information out of direct sigs.
* keylist.c (list_keyblock_print): don't assume that the presence
of a 0x20 signature means the key is revoked. With revocation
keys, this may not be true if the revocation key is not around to
verify it or if verification failed. Also, 0x1F should get listed
as "sig", and not "unexpected signature class".
* keyedit.c (show_key_with_all_names): Add a flag for printing
revoker information and change all callers.
* import.c (merge_blocks): merge in any new direct key (0x1F)
sigs.
* import.c (import_revoke_cert): don't keep processing after a
revocation is rejected.
* import.c (delete_inv_parts): Allow importing a revocation
signature even if it was not issued by the key. This allows a
revocation key to issue it. Of course, the sig still needs to be
checked before we trust it.
* free-packet.c (copy_public_key): Include a new copy of the
revocation keys when duping a pk.
* free-packet.c (free_seckey_enc, release_public_key_parts): Free
any revocation keys that are attached to a sig or pk.
* export.c (do_export_stream): Do not export signatures with
"sensitive" revocation keys in them.
2002-02-27 David Shaw <dshaw@jabberwocky.com>
* export.c (do_export_stream): Do not include v3 keys in a
--export-secret-subkeys export.
* getkey.c (merge_selfsigs_main): If a key isn't valid (say,
because of no self-signature), allow --always-trust to force it
valid so it can be trusted.
2002-02-25 David Shaw <dshaw@jabberwocky.com>
* hkp.c (hkp_ask_import), hkp.h, keyserver.c (all): treat key
lists internally as fingerprints when possible. All this is via
KEYDB_SEARCH_DESC - no point in reinventing the wheel. This allows
the helper program to search the keyserver by fingerprint if
desired (and the keyserver supports it). Note that automatic
fingerprint promotion during refresh only applies to v4 keys as a
v4 fingerprint can be easily changed into a long or short key id,
and a v3 cannot.
* pubkey-enc.c, getkey.c, misc.c, main.h: Take two copies of
hextobyte() from pubkey-enc.c and getkey.c and make them into one
copy in misc.c.
2002-02-22 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_search_prompt): Detect a "no keys found"
case even if the helper program does not explicitly say how many
keys were found.
* hkp.c (parse_hkp_index): Bug fix - don't report non-revoked keys
as revoked in HKP key searches.
2002-02-19 Werner Koch <wk@gnupg.org>
* parse-packet.c (parse_trust): Made parsing more robust.
2002-02-19 David Shaw <dshaw@jabberwocky.com>
* hkp.c (parse_hkp_index): Catch corruption in HKP index lines
(can be caused by broken or malicious keyservers).
* keyserver.c (keyserver_work): Add KEYSERVER_NOT_SUPPORTED for
unsupported actions (say, a keyserver that has no way to search,
or a readonly keyserver that has no way to add). Also add a
USE_EXTERNAL_HKP define to disable the internal HKP keyserver
code.
2002-02-14 Werner Koch <wk@gnupg.org>
* g10.c: New option --no-use-agent.
* pkclist.c (check_signatures_trust): Always print the warning for
unknown and undefined trust. Removed the did_add cruft. Reported
by Janusz A. Urbanowicz.
2002-02-11 David Shaw <dshaw@jabberwocky.com>
* hkp.c (parse_hkp_index): Bug fix - properly handle user IDs with
colons (":") in them while HKP searching.
2002-02-09 David Shaw <dshaw@jabberwocky.com>
* misc.c (pct_expando): More comments.
* keydb.h, sign.c (mk_notation_and_policy): Clarify what is a sig
and what is a cert. A sig has sigclass 0x00, 0x01, 0x02, or 0x40,
and everything else is a cert.
* g10.c (main), keyedit.c (keyedit_menu): Add a "nrlsign" for
nonrevocable and local key signatures.
* g10.c (main): Add a --no-force-mdc to undo --force-mdc.
* options.h, g10.c (main), cipher.c (write_header): Add a knob to
--disable-mdc/--no-disable-mdc. Off by default, of course, but is
used in --pgp2 and --pgp6 modes.
* pkclist.c (build_pk_list): Allow specifying multiple users in
the "Enter the user ID" loop. Enter a blank line to stop. Show
each key+id as it is added.
* keylist.c (show_policy_url), mainproc.c (print_notation_data):
It is not illegal (though possibly silly) to have multiple policy
URLs in a given signature, so print all that are present.
* hkp.c (hkp_search): More efficient implementation of URL-ifying
code.
2002-02-04 David Shaw <dshaw@jabberwocky.com>
* main.h, misc.c (pct_expando): New function to generalize
%-expando processing in any arbitrary string.
* photoid.c (show_photo): Call the new pct_expando function rather
than expand strings internally.
* sign.c (mk_notation_and_policy): Show policy URLs and notations
when making a signature if show-policy/show-notation is on.
%-expand policy URLs during generation. This lets the user have
policy URLs of the form "http://notary.jabberwocky.com/keysign/%K"
which will generate a per-signature policy URL.
* main.h, keylist.c (show_policy_url, show_notation): Add amount
to indent so the same function can be used in key listings as well
as during sig generation. Change all callers.
2002-02-04 David Shaw <dshaw@jabberwocky.com>
* keyserver.c, options.h (parse_keyserver_options, keyidlist):
Workaround for the pksd and OKS keyserver bug that calculates v4
RSA keyids as if they were v3. The workaround/hack is to fetch
both the v4 (e.g. 99242560) and v3 (e.g. 68FDDBC7) keyids. This
only happens for key refresh while using the HKP scheme and the
refresh-add-fake-v3-keyids keyserver option must be set. This
should stay off by default.
2002-02-03 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_spawn): Bug fix - do not append keys to
each other when --sending more than one.
2002-02-02 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c (main), keyedit.c (sign_uids), sign.c
(mk_notation_and_policy): Split "--set-policy-url" into
"--cert-policy-url" and "--sig-policy-url" so the user can set
different policies for key and data signing. For backwards
compatibility, "--set-policy-url" sets both, as before.
2002-01-30 Werner Koch <wk@gnupg.org>
* g10.c (main): --gen-random --armor does now output a base64
encoded string.
2002-01-28 David Shaw <dshaw@jabberwocky.com>
* g10.c (main), options.h, pkclist.c (algo_available): --pgp6
flag. This is not nearly as involved as --pgp2. In short, it
turns off force_mdc, turns on no_comment, escape_from, and
force_v3_sigs, and sets compression to 1. It also restricts the
user to IDEA (if present), 3DES, CAST5, MD5, SHA1, and RIPEMD160.
See the comments above algo_available() for lots of discussion on
why you would want to do this.
2002-01-27 David Shaw <dshaw@jabberwocky.com>
* keygen.c (keygen_set_std_prefs): Comment
* keyedit.c (sign_uids): Bug fix - when signing with multiple
secret keys at the same time, make sure each key gets the sigclass
prompt.
* exec.c (exec_finish): Close the iobuf and FILE before trying to
waitpid, so the remote process will get a SIGPIPE and exit. This
is only a factor when using a pipe to communicate.
* exec.c (exec_write): Disable cache-on-close of the fd iobuf (is
this right? Why is a fd iobuf cached at all?)
2002-01-26 Werner Koch <wk@gnupg.org>
* g10.c, options.h: New option --gpg-agent-info
* passphrase.c (agent_open): Let it override the environment info.
* seckey-cert.c (check_secret_key): Always try 3 times when the
agent is enabled.
* options.skel: Describe --use-agent.
2002-01-24 David Shaw <dshaw@jabberwocky.com>
* pubkey-enc.c (is_algo_in_prefs, get_it): Only check preferences
against keys with v4 self sigs - there is really little point in
warning for every single non-IDEA message encrypted to an old key.
* pkclist.c (select_algo_from_prefs): Only put in the fake IDEA
preference if --pgp2 is on.
* mainproc.c (check_sig_and_print): Print "Expired" for expired
but good signatures (this still prints "BAD" for expired but bad
signatures).
2002-01-23 David Shaw <dshaw@jabberwocky.com>
* keygen.c (ask_keysize): Cosmetic: don't present a RSA signing
key as a "keypair" which can be 768 bits long (as RSA minimum is
1024).
* pubkey-enc.c (is_algo_in_prefs): Allow IDEA as a fake preference
for v3 keys with v3 selfsigs.
2002-01-22 David Shaw <dshaw@jabberwocky.com>
* packet.h, getkey.c (merge_selfsigs_main), pkclist.c
(select_algo_from_prefs): Implement the fake IDEA preference as
per RFC2440:12.1. This doesn't mean that IDEA will be used (the
plugin may not be present), but it does mean that a v3 key with a
v3 selfsig has an implicit IDEA preference instead of 3DES. v3
keys with v4 selfsigs use preferences as normal.
* encode.c (encode_crypt): if select_algo_from_prefs fails, this
means that we could not find a cipher that both keys like. Since
all v4 keys have an implicit 3DES preference, this means there is
a v3 key with a v3 selfsig in the list. Use 3DES in this case as
it is the safest option (we know the v4 key can handle it, and
we'll just hope the v3 key is being used in an implementation that
can handle it). If --pgp2 is on, warn the user what we're doing
since it'll probably break PGP2 compatibility.
* g10.c (main): Do not force using IDEA for encrypted files in
--pgp2 mode - let the fake IDEA preference choose this for us for
better compatibility when encrypting to multiple keys, only some
of which are v3.
* keygen.c (keygen_set_std_prefs): Put 3DES on the end of the
default cipher pref list (RFC2440: "...it is good form to place it
there explicitly."). If the user has the IDEA plugin installed,
put a preference for IDEA *after* 3DES to effectively disable its
use for everything except encrypting along with v3 keys.
* encode.c, g10.c, sign.c: Change the PGP2 warning line from
"... will not be usable ..." to "... may not be usable ..." as the
user could be using one of the enhanced PGP2 variations.
* helptext.c: Revise the sign_uid.class help text as suggested by
Stefan.
2002-01-20 Werner Koch <wk@gnupg.org>
* passphrase.c (passphrase_to_dek): Add tryagain_text arg to be
used with the agent. Changed all callers.
(agent_get_passphrase): Likewise and send it to the agent
* seckey-cert.c (do_check): New arg tryagain_text.
(check_secret_key): Pass the string to do_check.
* keygen.c (ask_passphrase): Set the error text is required.
* keyedit.c (change_passphrase): Ditto.
* passphrase.c (agent_open): Disable opt.use_agent in case of a
problem with the agent.
(agent_get_passphrase): Ditto.
(passphrase_clear_cache): Ditto.
2002-01-19 Werner Koch <wk@gnupg.org>
* passphrase.c (agent_open): Add support for the new Assuan based
gpg-agent. New arg to return the used protocol version.
(agent_get_passphrase): Implemented new protocol here.
(passphrase_clear_cache): Ditto.
(readline): New.
2002-01-15 Timo Schulz <ts@winpt.org>
* encode.c (encode_crypt_files): Fail if --output is used.
* g10.c: New command --decrypt-files.
* decrypt.c (decrypt_messages): New.
2002-01-09 David Shaw <dshaw@jabberwocky.com>
* g10.c, misc.c, gpgv.c: move idea_cipher_warn to misc.c so gpgv.c
doesn't need a stub for it any longer.
* g10.c (get_temp_dir), main.h: no longer used (it's in exec.c now)
* g10.c (main), delkey.c (delete_keys), main.h : Allow
--delete-key (now --delete-keys, though --delete-key still works,
of course) to delete multiple keys in one go. This applies to
--delete-secret-key(s) and --delete-secret-and-public-key(s) as
well.
2002-01-09 Timo Schulz <ts@winpt.org>
* encode.c (encode_crypt_files): Now it behaves like verify_files.
* g10.c (main): We don't need to check argc for encode_crypt_files
any longer.
2002-01-09 Timo Schulz <ts@winpt.org>
* exec.c: Include windows.h for dosish systems.
2002-01-08 Timo Schulz <ts@winpt.org>
* g10.c (main): New description for --encrypt-files.
2002-01-08 Werner Koch <wk@gnupg.org>
* g10.c (main): Must register the secring for encryption because
it is needed to figure out the default recipient. Reported by
Roger Sondermann.
2002-01-05 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (menu_adduid): Require --expert before adding a photo
ID to a v3 key, and before adding a second photo ID to any key.
* keyedit.c (keyedit_menu): Don't allow adding photo IDs in
rfc1991 or pgp2 mode.
* getkey.c (merge_selfsigs_subkey): Permit v3 subkeys. Believe it
or not, this is allowed by rfc 2440, and both PGP 6 and PGP 7 work
fine with them.
* g10.c, options.h, keyedit.c, sign.c: Move the "ask for
expiration" switch off of --expert, which was getting quite
overloaded, and onto ask-sig-expire and ask-cert-expire. Both
default to off.
* g10.c (main): Change the default compression algo to 1, to be
more OpenPGP compliant (PGP also uses this, so it'll help with
interoperability problems as well).
* encode.c (encode_crypt): Handle compression algo 2, since the
default is now 1.
* build-packet.c (build_attribute_subpkt): Fix off-by-one error.
2002-01-05 Werner Koch <wk@gnupg.org>
* g10.c (main): Do not register the secret keyrings for certain
commands.
* keydb.c (keydb_add_resource): Use access to test for keyring
existence. This avoids cached opened files which are bad under
RISC OS.
2002-01-04 David Shaw <dshaw@jabberwocky.com>
* sign.c (sign_file, sign_symencrypt_file): always use one-pass
packets unless rfc1991 is enabled. This allows a signature made
with a v3 key to work in PGP 6 and 7. Signatures made with v4
keys are unchanged.
* g10.c (main): Disallow non-detached signatures in PGP2 mode.
Move the "you must use files and not pipes" PGP2 warning up so all
the PGP2 stuff is together.
* encode.c (encode_simple): Use the actual filesize instead of
partial length packets in the internal literal packet from a
symmetric message. This breaks PGP5(?), but fixes PGP2, 6, and 7.
It's a decent tradeoff. Note there was only an issue with
old-style RFC1991 symmetric messages. 2440-style messages in 6
and 7 work with or without partial length packets.
2002-01-03 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Removed --no-default-check-level option, as it is
not consistent with other "default" options. Plus, it is the same
as saying --default-check-level 0.
* exec.c (exec_read): Disallow caching tempfile from child
process, as this keeps the file handle open and can cause unlink
problems on some platforms.
* keyserver.c (keyserver_search_prompt): Minor tweak - don't
bother to transform keyids into textual form if they're just going
to be transformed back to numbers.
2002-01-03 Timo Schulz <ts@winpt.org>
* g10.c: New command --encrypt-files.
* verify.c (print_file_status): Removed the static because
encode_crypt_files also uses this function.
* main.h (print_files_status): New.
(encode_crypt_files): New.
* encode.c (encode_crypt_files): New.
2002-01-02 Stefan Bellon <sbellon@sbellon.de>
* keyserver.c: Moved util.h include down in order to avoid
redefinition problems on RISC OS.
* keyring.c (keyring_lock): Only lock keyrings that are writable.
* keyring.c (keyring_update_keyblock): Close unused iobuf.
* hkp.c (parse_hkp_index, hkp_search) [__riscos__]: Changed
unsigned char* to char* because of compiler issues.
* exec.c (exec_finish) [__riscos__]: Invalidate close cache so
that file can be unlinked.
2001-12-28 David Shaw <dshaw@jabberwocky.com>
* g10.c (main): Use a different strlist to check extensions since
they need to be handled seperately now.
* misc.c,main.h (check_permissions): Properly handle permission
and ownership checks on files in the lib directory
(e.g. /usr/local/lib/gnupg), which are owned by root and are
world-readable, and change all callers to specify extension or
per-user file.
* photoid.c (show_photo), keyserver.c (keyserver_spawn): Bug fix -
don't call exec_finish if exec_write fails.
* keyserver.c (keyserver_spawn): Look for OPTIONS from the
keyserver helper - specifically, a "OUTOFBAND" option for the
email keyserver.
* mainproc.c (list_node), keylist.c (list_keyblock_colon),
import.c (delete_inv_parts), export.c (do_export_stream): Use
signature flags for exportability check rather than re-parsing the
subpacket.
* keyid.c, keydb.h (get_lsign_letter): No longer needed.
2001-12-27 David Shaw <dshaw@jabberwocky.com>
* exec.c (exec_finish): Show errors when temp files cannot be
deleted for whatever reason.
* exec.c (exec_read): Don't rely on WEXITSTATUS being present.
* exec.c (make_tempdir): Add temp file creator for win32. Don't
create an incoming temp file if the exec is write-only.
* keyserver.c (keyserver_spawn): Clean up error handling, for when
the spawn fails.
* photoid.c (show_photo): Clean up error handling.
* misc.c (check_permissions): Neaten.
2001-12-25 David Shaw <dshaw@jabberwocky.com>
* mkdtemp.c (mkdtemp): Add copyleft info and tweak the 'X' counter
to be a bit simpler.
* keyserver.c, photoid.c: Remove unused headers left over from
when the exec functions lived there.
2001-12-23 Timo Schulz <ts@winpt.org>
* misc.c (check_permissions): Do not use it for W32 systems.
* tdbio.c (migrate_from_v2): Define ftruncate as chsize() for W32.
* mkdtemp.c: W32 support.
* photoid.c: Ditto.
* exec.c: Ditto.
2001-12-22 David Shaw <dshaw@jabberwocky.com>
* exec.c (make_tempdir): avoid compiler warning with const
* mkdtemp.c (mkdtemp): catch the empty ("") string case in case
someone repurposes mkdtemp at some point.
* photoid.c (generate_photo_id, show_photo): some type changes
from Stefan Bellon.
* exec.c (make_tempdir): handle Win32 systems, suggested by Timo
Schulz.
2001-12-22 Werner Koch <wk@gnupg.org>
* encode.c (encode_simple, encode_crypt): i18n 2 strings.
2001-12-22 Timo Schulz <ts@winpt.org>
* encode.c (encode_simple, encode_crypt): Use is_file_compressed
to avoid to compress compressed files.
2001-12-22 Werner Koch <wk@gnupg.org>
* keyserver.c (keyserver_spawn): Removed some variables
declaration due to shadowing warnings.
* build-packet.c (build_attribute_subpkt): s/index/idx/ to avoid
compiler warnig due to index(3).
* getkey.c (get_ctx_handle): Use KEYDB_HANDLE as return value.
* keylist.c (list_one): Made resname const.
* keyedit.c (keyedit_menu): Allow "addphoto" only when --openpgp is
not used.
* options.skel: Changed one example photo viewer to qiv.
2001-12-21 David Shaw <dshaw@jabberwocky.com>
* Makefile.am: add exec.c, exec.h, photoid.c, and photoid.h
* build-packet.c (build_attribute_subpkt): new function to build
the raw attribute subpacket. Note that attribute subpackets have
the same format as signature subpackets.
* exec.c: new file with generic exec-a-program functionality.
Used by both photo IDs and keyserver helpers. This is pretty much
the same code that used to be keyserver specific, with some
changes to be usable generically.
* free-packet.c (free_attributes (new)): function to free an
attribute packet.
* gpgv.c: added stub show_photo
* keyedit.c (keyedit_menu, menu_adduid, menu_showphoto): can add a
photo (calls generate_photo_id), or display a photo (calls
show_photo) from the --edit menu. New commands are "addphoto",
and "delphoto" (same as "deluid").
* keylist.c (list_keyblock_print): show photos during key list if
--show-photos enabled.
* keyserver.c (keyserver_spawn): use the generic exec_xxx
functions to call keyserver helper.
* g10.c, options.h: three new options - --{no-}show-photos, and
--photo-viewer to give the command line to display a picture.
* options.skel: instructions for the photo viewer
* parse-packet.c (parse_user_id, setup_user_id (new)): common code
for both user IDs and attribute IDs moved to setup_user_id.
* parse-packet.c (make_attribute_uidname (new)): constructs a fake
"name" for attribute packets (e.g. "[image of size ...]")
* parse-packet.c (parse_attribute (replaces parse_photo_id),
parse_attribute_subpkts): Builds an array of individual
attributes. Currently only handles attribute image / type jpeg
subpackets.
* sign.c (hash_uid): Fix bug in signing attribute (formerly
photo_id) packets.
* packet.h, and callers: globally change "photo_id" to "attribute"
and add structures for attributes. The packet format is generic
attributes, even though the only attribute type thus far defined
is jpeg.
2001-12-21 David Shaw <dshaw@jabberwocky.com>
* parse-packet.c (can_handle_critical): Can handle critical
revocation subpackets now.
* trustdb.c (mark_usable_uid_certs): Disregard revocations for
nonrevocable sigs. Note that this allows a newer revocable
signature to override an older nonrevocable signature.
* sign.c (make_keysig_packet): add a duration field and change all
callers. This makes make_keysig_packet closer to
write_signature_packets and removes some duplicated expiration
code.
* keyedit.c (keyedit_menu, menu_revsig, sign_uids,
sign_mk_attrib): Add nrsign command, don't allow revoking a
nonrevocable signature,
* g10.c (main): Add --nrsign option to nonrevocably sign a key
from the command line.
* build-packet.c (build_sig_subpkt_from_sig): Comment to explain
the use of CRITICAL.
2001-12-21 Werner Koch <wk@gnupg.org>
* g10.c. options.h : New option --show-keyring
* getkey.c (get_ctx_handle): New.
* keylist.c (list_one): Implement option here. By David Champion.
2001-12-20 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (keyserver_spawn): Use mkdtemp() to make temp
directory.
* mkdtemp.c: replacement function for those platforms that don't
have mkdtemp (make a temp directory securely).
2001-12-19 David Shaw <dshaw@jabberwocky.com>
* misc.c (check_permissions): New function to stat() and ensure
the permissions of GNUPGHOME and the files have safe permissions.
* keydb.c (keydb_add_resource): Check keyring permissions.
* tdbio.c (tdbio_set_dbname): Check permissions of trustdb.gpg
* keyserver.c (keyserver_spawn): Disable keyserver schemes that
involve running external programs if the options file has unsafe
permissions or ownership.
* g10.c, options.h: New option --no-permission-warning to disable
the permission warning message(s). This also permits use of the
keyserver if it had been disabled (see above). Also check the
permissions/ownership of random_seed.
* keyserver.c (keyserver_spawn): The new glibc prints a warning
when using mktemp() (the code was already secure, but the warning
was bound to cause confusion). Use a different implementation
based on get_random_bits() instead. Also try a few times to get
the temp dir before giving up.
2001-12-19 Werner Koch <wk@gnupg.org>
* g10.c, passphrase.c [CYGWIN32]: Allow this as an alias for MINGW32.
2001-12-18 David Shaw <dshaw@jabberwocky.com>
* g10.c (idea_cipher_warn): Add a flag to show the warning always
or once per session and change all callers (show always except for
the secret key protection and unknown cipher from an encrypted
message errors). Also make the strings translatable.
* pubkey-enc.c (get_it): Add the IDEA cipher warning if the user
tries to decrypt an IDEA encrypted message without the IDEA
plugin.
* keyserver.c (parse_keyserver_uri): More strict checking of the
keyserver URI. Specifically, fail if the ":port" section is
anything except a number between 1 and 65535.
2001-12-17 David Shaw <dshaw@jabberwocky.com>
* keyserver.c (print_keyinfo): No need to check for
control/illegal characters, as utf8_to_native does this for us.
* mainproc.c (proc_encrypted): Use generic IDEA warning.
* gpgv.c: add stub for idea_cipher_warn
* g10.c, hkp.c, keyserver.c: Fix capitalization and plural issues.
* encode.c (encode_crypt), sign.c (sign_file, clearsign_file):
disable pgp2 mode after the message is no longer pgp2 compatible.
* g10.c (main): Tweak the PGP2.x IDEA warning to use the generic
warning, and not merely fail if the IDEA plugin isn't there.
* g10.c (main, idea_cipher_warn), keygen.c (set_one_pref),
seckey-cert.c (do_check): Add a generic IDEA warning for when the
IDEA plugin is not present. This pops up when the user uses
"--cipher-algo idea", when setpref is used to set a "S1"
preference, and when a secret key protected with IDEA is used.
2001-12-15 Werner Koch <wk@gnupg.org>
* keyserver.c (keyserver_spawn): Assert that we have dropped privs.
2001-12-13 Werner Koch <wk@gnupg.org>
* pubkey-enc.c (get_session_key): Check that the public key
algorithm is indeed usable for en/decryption. This avoid a
strange error message from pubkey_decrypt if for some reasons a
bad algorithm indentifier is passed.
2001-12-12 David Shaw <dshaw@jabberwocky.com>
* Fixed some types for portability. Noted by Stefan Bellon.
2001-12-11 Werner Koch <wk@gnupg.org>
* hkp.c (hkp_export): Do not print possible control characters
from a keyserver response.
(parse_hkp_index): Made uid an unsigned char* because it is passed to
isspace().
(hkp_search): Ditto for the char* vars.
* g10.c (main): Print the IDEA warning also for -c and -se.
* g10.c (get_temp_dir): Assert that we have dropped privs
* encode.c (encode_crypt): Include the first key into the --pgp2
check.
2001-12-07 David Shaw <dshaw@jabberwocky.com>
* g10.c, options.h: New option --pgp2. This is identical to
"--rfc1991 --cipher-algo idea --compress-algo 1 --digest-algo md5
--force_v3_sigs" with the addition of an warning to advise the
user not to use a pipe (which would break pgp2 compatibility).
* encode.c (encode_crypt): warn if the user tries to encrypt to
any key that is not RSA and <= 2048 bits when the --pgp2 option is
used.
* sign.c (sign_file, clearsign_file): When using --pgp2, make a v3
sig, and warn if the signature is made with a non-v3 key.
2001-12-05 David Shaw <dshaw@jabberwocky.com>
* sign.c (sign_file, clearsign_file, sign_symencrypt_file): Prompt
for sig expiration if --expert is set and --force-v3-sigs is not
set (v3 sigs cannot expire).
* mainproc.c (check_sig_and_print): After checking a sig, print
expiration status. This causes a error return if the sig is
expired.
* build-packet.c (build_sig_subpkt_from_sig): Include a critical
sig expiration subpacket if the sig is to expire.
* keyedit.c (sign_uids): Do not sign an expired key unless
--expert is set, in which case prompt. Also, offer to expire a
signature when the key the user is signing expires.
* keygen.c (ask_expire_interval): Add a value to determine whether
to prompt for a key or sig expiration and change all callers.
* keyid.c: New functions: expirestr_from_sig and
colon_expirestr_from_sig.
* keylist.c (list_keyblock_colon): Show sig expiration date in the
--with-colons listing.
* sign.c (make_keysig_packet, write_signature_packets): Pass in an
optional timestamp for the signature packet, and change all
callers.
* keyedit.c (sign_mk_attrib): Include a critical expiration
subpacket in the signature if an expiration date is given.
2001-12-04 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (sign_uids): If the user tries to sign a
locally-signed key, allow the cert to be promoted to a full
exportable signature. This essentially deletes the old
non-exportable sig, and replaces it with a new exportable one.
2001-12-04 David Shaw <dshaw@jabberwocky.com>
* keyedit.c (keyedit_menu): Do not allow signing a revoked key
unless --expert is set, and ask even then.
* keyedit.c (sign_uids): Do not allow signing a revoked UID unless
--expert is set, and ask even then.
* g10.c, options.h : New option --expert
2001-11-16 David Shaw <dshaw@jabberwocky.com>
* Allow the user to select no compression via "--compress-algo 0"
on the command line.
* keyedit.c (show_prefs): Show compression preferences in the
long-form "showpref" style.
* keygen.c (set_one_pref): Permit setting a no-compression ("Z0")
preference.
* getkey.c (fixup_uidnode): Fix compression preference corruption
bug.
2001-12-02 David Shaw <dshaw@jabberwocky.com>
* g10.c: Add advisory --for-your-eyes-only option as per section
5.9 of 2440.
2001-12-05 David Shaw <dshaw@jabberwocky.com>
* Force a V4 sig if the user has a notation or policy URL set.
2001-12-04 David Shaw <dshaw@jabberwocky.com>
* g10.c: Add options --keyserver-options, --temp-directory, and
auto-key-retrieve (the opposite of no-auto-key-retrieve).
* hkp.c (hkp_search): New function to handle searching a HKP
keyserver for a key
* hkp.c (hkp_ask_import, hkp_export): Pretty large changes to make
them communicate via the generic functions in keyserver.c
* keyserver.c: new file with generic keyserver routines for
getting keys from a keyserver, sending keys to a keyserver, and
searching for keys on a keyserver. Calls the internal HKP stuff
in hkp.c for HKP keyserver functions. Other calls are handled by
an external program which is spawned and written to and read from
via pipes. Platforms that don't have pipes use temp files.
2001-11-20 David Shaw <dshaw@jabberwocky.com>
* options.h, g10.c: New options show-notation, no-show-notation,
default-check-level, no-default-check-level, show-policy-url,
no-show-policy-url.
* packet.h, sign.c (make_keysig_packet), parse-packet.c
(parse_signature), free-packet.c (free_seckey_enc): Fill in
structures for notation, policy, sig class, exportability, etc.
* keyedit.c, keylist.c (print_and_check_one_sig,
list_keyblock_print): Show flags in signature display for cert
details (class, local, notation, policy, revocable). If selected,
show the notation and policy url.
* keyedit.c (sign_uids): Prompt for and use different key sig
classes.
* helptext.c (helptexts): Add help text to explain different
key signature classes
2001-11-26 David Shaw <dshaw@jabberwocky.com>
* trustdb.c (mark_usable_uid_certs): Fix segfault from bad
initialization and fix reversed key signature expiration check.
2001-11-09 Werner Koch <wk@gnupg.org>
* export.c (do_export_stream): Put all given names into a search
description and change the loop so that all matching names are
returned.
2001-11-08 Werner Koch <wk@gnupg.org>
* pubkey-enc.c (get_it): To reduce the number of questions on the
MLs print the the name of cipher algorithm 1 with the error message.
* mainproc.c: Changed the way old rfc1991 encryption cipher is
selected. Based on a patch by W Lewis.
* pkclist.c (do_edit_ownertrust): Allow to skip over keys, the non
working "show info" is now assigned to "i"
* trustdb.c (ask_ownertrust, validate_keys): Implement a real quit
here. Both are by David Shaw.
* trustdb.c (validate_keys): Make sure next_exipire is initialized.
* sign.c (make_keysig_packet): Use SHA-1 with v4 RSA keys.
* g10.c, options.h : New option --[no-]froce-v4-certs.
* sign.c (make_keysig_packet): Create v4 sigs on v4 keys even with
a v3 key. Use that new option. By David Shaw
* revoke.c (ask_revocation_reason): Allow to select "no reason".
By David Shaw.
* keyid.c (fingerprint_from_sk): Calculation of an v3 fpr was
plain wrong - nearly the same code in fingerprint_from_pk is correct.
* build-packet.c (do_secret_key): Added a few comments to the code.
2001-11-07 Werner Koch <wk@gnupg.org>
* g10.c (main): Print a warning when -r is used w/o encryption.
Suggested by Pascal Scheffers.
2001-10-23 Werner Koch <wk@gnupg.org>
* keyedit.c (keyedit_menu): Changed helptext for showpref
command. Suggested by Reinhard Wobst.
* keyring.c (keyring_search): When marking the offtbl ready, take
into account that we may have more than one keyring.
2001-10-22 Werner Koch <wk@gnupg.org>
* Makefile.am: Do not use OMIT_DEPENDENCIES
* build-packet.c (build_sig_subpkt): Default is now to put all
types of subpackets into the hashed area and only list those which
should go into the unhashed area.
2001-10-18 Werner Koch <wk@gnupg.org>
* keydb.c (keydb_add_resource): Rearranged the way we keep track
of the resource. There will now be an entry for each keyring here
and not in keyring.c itself. Store a token to allow creation of a
keyring handle. Changed all functions to utilize this new design.
(keydb_locate_writable): Make a real implementation.
* keyring.c (next_kr): Removed and changed all callers to set the
resource directly from the one given with the handle.
(keyring_is_writable): New.
(keyring_rebuild_cache): Add an arg to pass the token from keydb.
2001-10-17 Werner Koch <wk@gnupg.org>
* keyring.c (keyring_search): Enabled word search mode but print a
warning that it is buggy.
2001-10-11 Werner Koch <wk@gnupg.org>
* hkp.c (hkp_ask_import): No more need to set the port number for
the x-hkp scheme.
(hkp_export): Ditto.
2001-10-06 Stefan Bellon <sbellon@sbellon.de>
* passphrase.c [__riscos__]: Disabled agent specific stuff.
* g10.c: New option --no-force-v3-sigs.
2001-10-04 Werner Koch <wk@gnupg.org>
* export.c (do_export_stream): Do not push the compress filter
here because the context would run out of scope due to the
iobuf_close done by the caller.
(do_export): Do it here instead.
2001-09-28 Werner Koch <wk@gnupg.org>
* keyedit.c (sign_uids): Always use the primary key to sign keys.
* getkey.c (finish_lookup): Hack to return only the primary key if
a certification key has been requested.
* trustdb.c (cmp_kid_for_make_key_array): Renamed to
(validate_one_keyblock): this and changed arg for direct calling.
(make_key_array): Renamed to
(validate_one_keyblock): this and changed args for direct calling.
(mark_usable_uid_certs, validate_one_keyblock)
(validate_key_list): Add next_expire arg to keep track of
expiration times.
(validate_keys): Ditto for UTKs and write the stamp.
* tdbio.c (migrate_from_v2): Check return code of tbdio_sync.
* tdbdump.c (import_ownertrust): Do a tdbio_sync().
* keyring.c: Made the offtbl an global object.
2001-09-27 Werner Koch <wk@gnupg.org>
* pkclist.c (do_edit_ownertrust): Allow settin of ultimate trust.
* trustdb.c (mark_keyblock_seen): New.
(make_key_array): Use it to mark the subkeys too.
(validate_keys): Store validity for ultimatly trusted keys.
2001-09-26 Werner Koch <wk@gnupg.org>
* pkclist.c (check_signatures_trust, do_we_trust): Removed the
invocation of add_ownertrust. Minor changes to the wording.
(add_ownertrust, add_ownertrust_cb): Removed.
* trustdb.c (get_validity): Allow to lookup the validity using a
subkey.
* trustdb.c (new_key_hash_table): Increased the table size to 1024
and changed the masks accordingly.
(validate): Changed stats printing.
(mark_usable_uid_certs): New.
(cmp_kid_for_make_key_array): Does now check the signatures and
figures out a usable one.
2001-09-25 Werner Koch <wk@gnupg.org>
* keyring.c (new_offset_item,release_offset_items)
(new_offset_hash_table, lookup_offset_hash_table)
(update_offset_hash_table, update_offset_hash_table_from_kb): New.
(keyring_search): Use a offset table to optimize search for
unknown keys.
(keyring_update_keyblock, keyring_insert_keyblock): Insert new
offsets.
* getkey.c (MAX_UNK_CACHE_ENTRIES): Removed the unknown keys
caching code.
* g10.c, options.h, import.c: Removed the entire
allow-secret-key-import stuff because the validity is now
controlled by other means.
* g10.c: New command --rebuild-keydb-caches.
* keydb.c (keydb_rebuild_caches): New.
* keyring.c (do_copy): Moved some code to
(create_tmp_file, rename_tmp_file, write_keyblock): new functions.
(keyring_rebuild_cache): New.
* packet.h (PKT_ring_trust): Add sigcache field.
* parse-packet.c (parse_trust): Parse sigcache.
* keyring.c (do_copy): Always insert a sigcache packet.
(keyring_get_keyblock): Copy the sigcache packet to the signature.
* sig-check.c (cache_sig_result): Renamed from
cache_selfsig_result. Changed implementation to use the flag bits
and changed all callers.
(mdc_kludge_check): Removed this unused code.
(do_check): Do not set the sig flags here.
* import.c (read_block): Make sure that ring_trust packets are
never imported.
* export.c (do_export_stream): and never export them.
* trustdb.c (make_key_array): Skip revoked and expired keys.
2001-09-24 Werner Koch <wk@gnupg.org>
* g10.c, options.h: New option --no-auto-check-trustdb.
* keygen.c (do_generate_keypair): Set newly created keys to
ultimately trusted.
* tdbio.h, tdbio.c: Removed all support for records DIR, KEY, UID,
PREF, SIG, SDIR and CACH. Changed migration function to work
direct on the file.
(tdbio_read_nextcheck): New.
(tdbio_write_nextcheck): New.
2001-09-21 Werner Koch <wk@gnupg.org>
Revamped the entire key validation system.
* trustdb.c: Complete rewrite. No more validation on demand,
removed some functions, adjusted to all callers to use the new
and much simpler interface. Does not use the LID anymore.
* tdbio.c, tdbio.h: Add new record types trust and valid. Wrote a
migration function to convert to the new trustdb layout.
* getkey.c (classify_user_id2): Do not allow the use of the "#"
prefix.
* keydb.h: Removed the TDBIDX mode add a skipfnc to the
descriptor.
* keyring.c (keyring_search): Implemented skipfnc.
* passphrase.c (agent_open): Add missing bracket. Include windows.h.
2001-09-19 Werner Koch <wk@gnupg.org>
* keylist.c (print_fingerprint): Renamed from fingerprint, made
global available. Added new arg to control the print style.
* mainproc.c (print_fingerprint): Removed.
* pkclist.c (print_fpr, fpr_info): Removed and changed callers to
use print_fingerprint.
* keyedit.c (show_fingerprint): Ditto.
* passphrase.c (writen, readn)
(agent_open, agent_close)
(agent_get_passphrase)
(passphrase_clear_cache): Support for W32. Contributed by Timo.
* import.c (import_one): Release keydb handles at 2 more places.
* keyring.c (keyring_release): Close the iobuf.
(keyring_get_keyblock): Init ret_kb to NULL and store error contidion.
* import.c (import_new_stats_handle): New.
(import_release_stats_handle): New.
(import_print_stats): Renamed from static fnc print_stats.
(import_keys, import_keys_stream): Add an optional status handle
arg and changed all callers.
* hkp.c (hkp_ask_import): Add an stats_handle arg and changed all
callers.
* mainproc.c (print_pkenc_list): Use print_utf8_string2().
2001-09-18 Werner Koch <wk@gnupg.org>
* g10.c: New command --refresh-keys.
* hkp.c (hkp_refresh_keys): New. Contributed by Timo Schulz.
* parse-packet.c (parse): Stop on impossible packet lengths.
2001-09-17 Werner Koch <wk@gnupg.org>
* mainproc.c (print_notation_data): Wrap notation data status lines
after 50 chars.
* mainproc.c (proc_pubkey_enc): Make option try-all-secrets work.
By disastry@saiknes.lv.
2001-09-14 Werner Koch <wk@gnupg.org>
* parse-packet.c (dump_sig_subpkt): List key server preferences
and show the revocable flag correctly. Contributed by David Shaw.
2001-09-09 Werner Koch <wk@gnupg.org>
* keyedit.c (keyedit_menu): No need to define another p.
* keylist.c (print_capabilities): s/used/use/ so that it
does not shadow a global.
* sign.c (sign_file): Renamed arg encrypt to encryptflag
* keygen.c: Replaced all "usage" by "use".
* misc.c (openpgp_pk_algo_usage): Ditto.
* pubkey-enc.c (get_it): Renamed arg k to enc so that the later
defined k does not shadow it.
* parse-packet.c (parse_gpg_control): No need to define another i.
* getkey.c (get_pubkey_byfprint): Must use the enum values and not
the fprint_len.
* keyring.c (keyring_search): Removed a non-sense break. Both
bugs pointed out by Stefan.
2001-09-07 Werner Koch <wk@gnupg.org>
* status.c, status.h: Added NO_RECP and ALREADY_SIGNED.
* pkclist.c (build_pk_list): Issue NO_RECP.
* keyedit.c (sign_uids): Added experimental ALREADY_SIGNED
* hkp.c (hkp_import): Use log_error. Bug reported by Neal H
Walfield.
* getkey.c (classify_user_id2): Change args to take the desc union
direct. It was a stupid idea to pass the individual fields of an
union to this function. Changed all callers.
(classify_user_id): Ditto and allow to pass NULL as the description.
2001-09-06 Werner Koch <wk@gnupg.org>
* getkey.c (fixup_uidnode): Features flag is now a bit vector.
* keygen.c (add_feature_mdc): Ditto.
Revamped the entire key I/O code to be prepared for other ways of
key storages and to get rid of the existing shit. GDBM support has
gone.
* keydb.c: New
* keyring.c, keyring.h: New.
* ringedit.c: Removed. Moved some stuff to keyring.c
* getkey.c: Changed everything related to the key retrieving
functions which are now using the keydb_ functions.
(prepare_search, word_match_chars, word_match)
(prepare_word_match, compare_name): Moved to keyring.c
(get_pubkey_byname): Removed ctx arg and add ret_kdbhd
arg. Changed all callers.
(key_byname): Use get_pubkey_end to release the context and take
new ret_kbdhd arg. Changed all callers.
(classify_user_id2): Fill the 16 byte fingerprint up with 4 null
bytes not with zero bytes of value 4, tsss.
* import.c (import_one): Updated to use the new keydb interface.
(import_secret_one): Ditto.
(import_revoke_cert): Ditto.
* delkey.c (do_delete_key): Ditto.
* keyedit.c (keyedit_menu): Ditto.
(get_keyblock_byname): Removed.
* revoke.c (gen_revoke): Ditto.
* export.c (do_export_stream): Ditto.
* trustdb.c (update_trustdb): Ditto.
* g10.c, gpgv.c (main): Renamed add_keyblock_resource to
keydb_add_resource.
* Makefile.am: Added and removed files.
* keydb.h: Moved KBNODE typedef and MAX_FINGERPRINT_LEN to
* global.h: this new header.
2001-09-03 Werner Koch <wk@gnupg.org>
* passphrase.c (agent_get_passphrase): Changed nread to size_t.
(passphrase_clear_cache): Ditto.
* keyid.c (mk_datestr): Avoid trigraphs.
(fingerprint_from_pk): Cache the keyid in the pk.
* options.h: Add opt.with_fingerprint so that we know whether the
corresponding options was used.
* g10.c (main): Set it here.
* pkclist.c (check_signatures_trust): Always print fingerprint
when this option is used. Mixed a minor memory leak.
* status.c, status.h: New status INV_RECP.
* pkclist.c (build_pk_list): Issue this status.
2001-08-31 Werner Koch <wk@gnupg.org>
* parse-packet.c (parse_key,parse_pubkeyenc)
(parse_signature): Return error on reading bad MPIs.
* mainproc.c (check_sig_and_print): Always print the user ID even
if it is not bound by a signature. Use the primary UID in the
status messages and encode them in UTF-8
* status.c (write_status_text_and_buffer): New.
2001-08-30 Werner Koch <wk@gnupg.org>
* packet.h (sigsubpkttype_t): Add SIGSUBPKT_FEATURES.
(PKT_public_key, PKT_user_id): Add a flag for it.
* parse-packet.c, build-packet.c: Add support for them.
* getkey.c (fixup_uidnode, merge_selfsigs): Set the MDC flags.
* keygen.c (add_feature_mdc): New.
(keygen_upd_std_prefs): Always set the MDC feature.
* keyedit.c (show_prefs): List the MDC flag
* pkclist.c (select_mdc_from_pklist): New.
* encode.c (encode_crypt, encrypt_filter): Test whether MDC
should be used.
* cipher.c (write_header): Set MDC use depending on the above test.
Print more status info.
* delkey.c (do_delete_key): Kludge to delete a secret key with no
public key available.
* ringedit.c (find_secret_keyblock_direct): New.
* getkey.c (seckey_available): Simplified.
* ringedit.c (cmp_seckey): Now compares the secret key against the
public key while ignoring all secret parts.
(keyring_search): Use a public key packet as arg. Allow to search
for subnkeys
(search): Likewise. Changed all callers.
(find_secret_keyblock_bypk): New.
(find_secret_keyblock_byname): First locate the pubkey and then
find the correponding secret key.
* parse-packet.c (parse): Renamed pkttype arg to onlykeypkts and
changed code accordingly. Changed all callers.
(search_packet): Removed pkttype arg.
* keyedit.c (keyedit_menu): First locate the public key and then
try to locate a secret key.
* ringedit.c (locate_keyblock_by_fpr): Removed.
(locate_keyblock_by_keyid): Removed.
(find_keyblock_bysk): Removed.
* sig-check.c (check_key_signature2): Print the keyid along with
the wrong sig class errors.
2001-08-24 Werner Koch <wk@gnupg.org>
* sign.c (sign_file): Stripped the disabled comment packet code.
(sign_file, sign_symencrypt_file): Moved common code to ..
(write_onepass_sig_packets): .. this new function.
(sign_file, clearsign_file, sign_symencrypt_file): Moved common
code to
(write_signature_packets): this new function.
(write_signature_packets, make_keysig_packet)
(update_keysig_packet): Moved common code to
(hash_uid, hash_sigclass_to_magic): these new functions
(sign_file, sign_symencrypt_file): Moved common code to
(write_plaintext_packet): this new function.
2001-08-21 Stefan Bellon <sbellon@sbellon.de>
* trustdb.c (query_trust_info): Changed trustlevel to signed int.
* g10.c [__riscos__]: Fixed handling of --use-agent --lock-multiple.
2001-08-20 Werner Koch <wk@gnupg.org>
* encr-data.c (decrypt_data): Keep track on whether we already
printed information about the used algorithm.
* mainproc.c (proc_encrypted): Removed the non-working IDEA hack
and print a message about the assumed algorithm.
* passphrase.c (passphrase_to_dek): Use the same algorithm as above.
(proc_symkey_enc): Print the algorithm, so that the user knows it
before entering the passphrase.
(proc_pubkey_enc, proc_pubkey_enc): Zero the DEK out.
* encode.c (encode_crypt, encrypt_filter): Ditto.
* g10.c: Allow for --sign --symmetric.
* sign.c (sign_and_symencrypt): New.
Applied patches from Stefan Bellon <sbellon@sbellon.de> to support
RISC OS. Nearly all of these patches are identified by the
__riscos__ macro.
* compress.c: Added a couple of casts.
* g10.c [__riscos__]: Some patches and new options foo-file similar
to all foo-fd options.
* gpgv.c, openfile.c, ringedit.c, tdbio.c: Minor fixes. Mainly
replaced hardcoded path separators with EXTSEP_S like macros.
* passprase.c [__riscos__]: Disabled agent stuff
* trustdb.c (check_trust): Changed r_trustlevel to signed int to
avoid mismatch problems in pkclist.c
* pkclist.c (add_ownertrust): Ditto.
* plaintext.c (handle_plaintext) [__riscos__]: Print a note when
file can't be created.
* options.h [__riscos__]: Use an extern unless included from the
main module.
* signal.c (got_fatal_signal) [__riscos__]: Close all files.
2001-08-14 Werner Koch <wk@gnupg.org>
* keygen.c (ask_algo): New arg r_usage. Allow for RSA keys.
(gen_rsa): Enabled the code.
(do_create): Enabled RSA branch.
(parse_parameter_usage): New.
(proc_parameter_file): Handle usage parameter.
(read_parameter_file): Ditto.
(generate_keypair): Ditto.
(generate_subkeypair): Ditto.
(do_generate_keypair): Ditto.
(do_add_key_flags): New.
(keygen_add_std_prefs): Use the new function.
(keygen_add_key_flags_and_expire): New.
(write_selfsig, write_keybinding): Handle new usage arg.
* build-packet.c (build_sig_subpkt): Make sure that key flags go
into the hashed area.
* keygen.c (write_uid): Initialize the reference cunter.
* keyedit.c (keyedit_menu): No more need to update the trustdb for
preferences. Added calls to merge keblock.
* kbnode.c (dump_kbnode): Print some more flags.
2001-08-10 Werner Koch <wk@gnupg.org>
Revamped the preference handling.
* packet.h (prefitem_t, preftype_t): New.
(PKT_public_key): Added a uid field.
(PKT_user_id): Added field to store preferences and a reference
counter.
* parse-packet.c (parse_user_id,parse_photo_id): Initialize them
* free-packet.c (free_user_id): Free them.
(copy_user_id): Removed.
(scopy_user_id): New.
(cmp_user_ids): Optimized for identical pointers.
(release_public_key_parts): Release the uid.
(copy_public_key_with_new_namehash): Removed.
(copy_prefs): New.
* keyedit.c (menu_adduid): Use the new shallow copy user id.
(show_prefs): Adjusted implementation.
(keyedit_menu): No more need to update the trustdb after changing
preferences.
* getkey.c (fixup_uidnode): Store preferences.
(find_by_name): Return a user id packet and remove namehash stuff.
(lookup): Removed the unused namehash stuff.
(finish_lookup): Added foundu arg.
(pk_from_block): Removed the namehash arg and changed all callers.
(merge_selfsigs): Copy prefs to all keys.
* trustdb.c (get_pref_data): Removed.
(is_algo_in_prefs): Removed.
(make_pref_record): Deleted and removed all class.
* pkclist.c (select_algo_from_prefs): Adjusted for the new
preference implementation.
* pubkey-enc.c (is_algo_in_prefs): New.
(get_it): Use that new function.
2001-08-09 Werner Koch <wk@gnupg.org>
* build-packet.c (build_sig_subpkt): Fixed calculation of
newarea->size.
* g10.c (main): New option "--preference-list"
* keyedit.c (keyedit_menu): New commands "setpref" and "updpref".
(menu_set_preferences): New.
* keygen.c (keygen_set_std_prefs): New.
(set_one_pref): New.
(check_zip_algo): New.
(keygen_get_std_prefs): New.
(keygen_upd_std_prefs): New
(keygen_add_std_prefs): Move the pref setting code into the above fnc.
* build-packet.c (build_sig_subpkt): Updated the list of allowed
to update subpackets.
2001-08-08 Werner Koch <wk@gnupg.org>
* packet.h (subpktarea_t): New.
(PKT_signature): Use that type for hashed_data and unhashed_data and
removed the _data prefix from those fields. Changed all users.
* parse-packet.c (parse_signature): Changed allocation for that.
(parse_sig_subpkt): Changed declaration
(enum_sig_subpkt): Ditto and changed implementation accordingly.
* free-packet.c (cp_subpktarea): Renamed from cp_data_block and
adjusted implementation. Changed caller.
* sig-check.c (mdc_kludge_check): Adjusted the hashing.
(do_check): Ditto.
* sign.c (sign_file, clearsign_file, make_keysig_packet,
update_keysig_packet): Ditto.
* build-packet.c (build_sig_subpkt): Partial rewrite.
(find_subpkt): Adjusted and made static.
(delete_sig_subpkt): Adjusted.
(do_signature): Ditto.
* keygen.c (ask_keysize): Do not print the notes about suggested
key sizes if just a DSA key is generated.
* trustdb.c (add_ultimate_key): s/log_error/log_info/ for
duplicated inserted trusted keys.
2001-08-07 Werner Koch <wk@gnupg.org>
* sign.c (sleep): Redefine for W32.
* g10.c, options.h: Set new flag opt.no_homedir_creation when
--no-options is given.
* openfile.c (try_make_homedir): Don't create the homedir in that case.
2001-08-03 Werner Koch <wk@gnupg.org>
* armor.c (armor_filter): Removed the default comment string
because it could get us in trouble due to translations using non
ascii characters.
2001-08-01 Werner Koch <wk@gnupg.org>
* keylist.c (list_keyblock_print): Do not list revoked UIDs unless
in verbose mode and we do no signature listing.
* getkey.c (finish_lookup): Skip subkeys which are not yet valid.
* g10.c, options.h: New option --ignore-valid-from.
* sign.c (make_keysig_packet): Added new sigversion argument to
allow the caller to force generation of required signature
version. Changed all callers. Suggested by Thomas Roessler.
* keyedit.c (sign_uids): Force v4 signature generation for local
sigs. Removed the check for local signature and pre-v4 keys.
2001-07-27 Werner Koch <wk@gnupg.org>
* keyedit.c (sign_uids): Check that we are not trying to to a
lsign with a pre-v4 key. Bug noticed by Thomas Roessler.
2001-07-26 Werner Koch <wk@gnupg.org>
* parse-packet.c (parse_photo_id): Reset all variables.
* getkey.c (merge_selfsigs_main): Removed checks on PHOTO_ID
because this is handled identically to a user ID.
2001-07-06 Werner Koch <wk@gnupg.org>
* cipher.c (write_header): Don't use MDC with --rfc1991. Suggested
by disastry@saiknes.lv.
2001-07-05 Werner Koch <wk@gnupg.org>
* g10.c, options.h: New option --preserve-permissions.
* ringedit.c (add_keyblock_resource): Use it here
(keyring_copy): and here.
* trustdb.c (verify_own_keys): Be more silent on --quiet.
Suggested by Thomas Roessler.
* sig-check.c (check_key_signature2): Ditto.
* mainproc.c (proc_encrypted, proc_tree): Ditto
* getkey.c (lookup): Ditto.
2001-07-04 Werner Koch <wk@gnupg.org>
* ringedit.c (add_keyblock_resource): Restore filename in case of error.
2001-06-25 Werner Koch <wk@gnupg.org>
* kbnode.c (dump_kbnode): Print the signature timestamp.
* keyedit.c (keyedit_menu): New menu point "primary".
(change_primary_uid_cb): New.
(menu_set_primary_uid): New.
* sign.c (update_keysig_packet): New.
* build-packet.c (build_sig_subpkt): Put the primary UID flag into
the hashed area. Allow update of some more packets.
2001-06-15 Werner Koch <wk@gnupg.org>
* getkey.c (merge_selfsigs): Exit gracefully when a secret key is
encountered. May happen if a secret key is in public keyring.
Reported by Francesco Potorti.
2001-06-12 Werner Koch <wk@gnupg.org>
* getkey.c (compare_name): Use ascii_memistr(), ascii_memcasecmp()
* keyedit.c (keyedit_menu): Use ascii_strcasecmp().
* armor.c (radix64_read): Use ascii_toupper().
* ringedit.c (do_bm_search): Ditto.
* keygen.c (read_parameter_file): Ditto.
* openfile.c (CMP_FILENAME): Ditto.
* g10.c (i18n_init): We can now use just LC_ALL.
2001-05-29 Werner Koch <wk@gnupg.org>
* keygen.c (generate_subkeypair): Print a warning if a subkey is
created on a v3 key. Suggested by Brian M. Carlson.
2001-05-27 Werner Koch <wk@gnupg.org>
* keyid.c (get_lsign_letter): New.
* keylist.c (list_keyblock_colon): Use it here.
* mainproc.c (list_node): and here.
* getkey.c, packet.h, free-packet.c: Removed that useless key
created field; I dunno why I introducded this at all - the
creation time is always bound to the key packet and subject to
fingerprint calculation etc.
* getkey.c (fixup_uidnode): Add keycreated arg and use this
instead of the signature timestamp to calculate the
help_key_expire. Bug reported by David R. Bergstein.
(merge_selfsigs_main): Correct key expiration time calculation.
(merge_selfsigs_subkey): Ditto.
2001-05-25 Werner Koch <wk@gnupg.org>
* revoke.c (gen_revoke): Add a cast to a tty_printf arg.
* delkey.c (do_delete_key): Ditto.
* keyedit.c (print_and_check_one_sig): Ditto.
(ask_revoke_sig): Ditto.
(menu_revsig): Ditto.
(check_all_keysigs): Removed unused arg.
2001-05-23 Werner Koch <wk@gnupg.org>
* g10.c (opts): Typo fix by Robert C. Ames.
2001-05-06 Werner Koch <wk@gnupg.org>
* revoke.c: Small typo fix
2001-05-04 Werner Koch <wk@gnupg.org>
* passphrase.c (passphrase_clear_cache): Shortcut if agent usage
is not enabled.
2001-05-01 Werner Koch <wk@gnupg.org>
* passphrase.c (writen): Replaced ssize_t by int. Thanks to
to Robert Joop for reporting that SunOS 4.1.4 does not have it.
2001-04-28 Werner Koch <wk@gnupg.org>
* getkey.c (merge_public_with_secret): pkttype was not set to subkey.
2001-04-27 Werner Koch <wk@gnupg.org>
* skclist.c (build_sk_list): Changed one log_debug to log_info.
2001-04-25 Werner Koch <wk@gnupg.org>
* keyedit.c (show_prefs): Add a verbose mode.
(show_key_with_all_names): Pass verbose flag for special value of
with_pref.
(keyedit_menu): New command "showpref"
(show_key_with_all_names): Mark revoked uids and the primary key.
2001-04-24 Werner Koch <wk@gnupg.org>
* getkey.c (get_primary_uid): Return a different string in case of
error and made it translatable.
* build-packet.c (do_secret_key): Ugly, we wrote a zero
instead of the computed ndays. Thanks to M Taylor for complaining
about a secret key import problem.
2001-04-23 Werner Koch <wk@gnupg.org>
* hkp.c (hkp_ask_import): Allow to specify a port number for the
keyserver. Add a kudge to set the no_shutdown flag.
(hkp_export): Ditto.
* options.skel: Document the changes
2001-04-20 Werner Koch <wk@gnupg.org>
* options.skel: Add some more comments.
2001-04-19 Werner Koch <wk@gnupg.org>
* keyid.c (mk_datestr): New. Handles negative times. We must do
this because Windoze segvs on negative times passed to gmtime().
Changed all datestr_from function to use this one.
* keyid.c, keyid.h (colon_strtime): New. To implement the
fixed-list-mode.
(colon_datestr_from_pk): New.
(colon_datestr_from_sk): New.
(colon_datestr_from_sig): New.
* keylist.c (list_keyblock_colon): Use these functions here.
* mainproc.c (list_node): Ditto.
2001-04-18 Werner Koch <wk@gnupg.org>
* openfile.c (open_sigfile): Fixed the handling of ".sign".
* mainproc.c (proc_tree): Use iobuf_get_real_fname.
Both are by Vincent Broman.
2001-04-14 Werner Koch <wk@gnupg.org>
* getkey.c (fixup_uidnode): Removed check for !sig which is
pointless here. Thanks to Jan Niehusmann.
2001-04-10 Werner Koch <wk@gnupg.org>
* sig-check.c (check_key_signature2): Use log_info instead of
log_error so that messed up keys do not let gpg return an error.
Suggested by Christian Kurz.
* getkey.c (merge_selfsigs_main): Do a fixup_uidnode only if we
have both, uid and sig. Thanks to M Taylor.
2001-04-05 Werner Koch <wk@gnupg.org>
* armor.c (unarmor_pump_new,unarmor_pump_release): New.
(unarmor_pump): New.
* pipemode.c (pipemode_filter): Use the unarmor_pump to handle
armored or non-armored detached signatures. We can't use the
regular armor_filter becuase this does only chack for armored
signatures the very first time. In pipemode we may have a mix of
armored and binary detached signatures.
* mainproc.c (proc_tree): Do not print the "old style" notice when
this is a pipemode processes detached signature.
(proc_plaintext): Special handling of pipemode detached sigs.
* packet.h (CTRLPKT_PLAINTEXT_MARK): New.
* parse-packet.c (create_gpg_control): New.
* kbnode.c (dump_kbnode): Support it here.
* mainproc.c (check_sig_and_print): Fixed the check for bad
sequences of multiple signatures.
(proc_plaintext): Add the marker packet.
(proc_tree): We can now check multiple detached signatures.
2001-04-02 Werner Koch <wk@gnupg.org>
The length of encrypted packets for blocksizes != 8 was not
correct encoded. I think this is a minor problem, because we
usually use partial length packets. Kudos to Kahil D. Jallad for
pointing this out.
* packet.h: Add extralen to PKT_encrypted.
* cipher.c (write_header): Set extralen.
* build-packet.c (do_encrypted): Use extralen instead of const 10.
(do_encrypted_mdc): Ditto.
* parse-packet.c (parse_encrypted): Set extralen to 0 because we
don't know it here.
2001-03-30 Werner Koch <wk@gnupg.org>
* getkey.c (premerge_public_with_secret): Changed wording an add
the keyID to the info message.
2001-03-29 Werner Koch <wk@gnupg.org>
* getkey.c (premerge_public_with_secret): Use log_info instead of
log_error when no secret key was found for a public one.
Fix the usage if the secret parts of a key are not available.
* openfile.c (ask_outfile_name): Trim spaces.
(open_outfile): Allow to enter an alternate filename. Thanks to
Stefan Bellon.
* plaintext.c (handle_plaintext): Ditto.
2001-03-28 Werner Koch <wk@gnupg.org>
* mainproc.c (do_check_sig): Allow direct key and subkey
revocation signature.
* sig-check.c (check_key_signature2): Check direct key signatures.
Print the signature class along with an error.
2001-03-27 Werner Koch <wk@gnupg.org>
* packet.h: Add a missing typedef to an enum. Thanks to Stefan Bellon.
* g10.c: New option --no-sig-create-check.
* sign.c (do_sign): Implement it here.
* g10.c: New option --no-sig-cache.
* sig-check.c (check_key_signature2): Implement it here.
(cache_selfsig_result): and here.
* keylist.c (list_keyblock): Removed debugging stuff.
* getkey.c (cache_public_key): Made global.
* keygen.c (write_selfsig, write_keybinding): Cache the new key.
* getkey.c (key_byname): Add new arg secmode and changed all
callers to request explicitly the mode. Deriving this information
from the other supplied parameters does not work if neither pk nor
sk are supplied.
2001-03-25 Werner Koch <wk@gnupg.org>
* packet.h (ctrlpkttype_t): New.
* mainproc.c (add_gpg_control,proc_plaintext,proc_tree): Use the
new enum values.
* pipemode.c (make_control): Ditto.
* armor.c (armor_filter): Ditto.
2001-03-24 Werner Koch <wk@gnupg.org>
* sign.c (do_sign): Verify the signature right after creation.
2001-03-23 Werner Koch <wk@gnupg.org>
* status.c, status.h (STATUS_UNEXPECTED): New.
* mainproc.c (do_proc_packets): And emit it here.
2001-03-21 Werner Koch <wk@gnupg.org>
* status.c: Add sys/types.h so that it runs on Ultrix. Reported
by Georg Schwarz.x
* build-packet.c (build_sig_subpkt): Fixed generaton of packet
length header in case where 2 bytes headers are needed. Thanks to
Piotr Krukowiecki.
2001-03-19 Werner Koch <wk@gnupg.org>
* g10.c (main): the default keyring is no always used unless
--no-default-keyring is given.
* ringedit.c (add_keyblock_resource): invalidate cache after file
creation.
2001-03-15 Werner Koch <wk@gnupg.org>
* keygen.c (ask_algo): Changed the warning of the ElGamal S+E Algo.
* keylist.c (print_capabilities): New.
(list_keyblock_colon): and use it here.
2001-03-13 Werner Koch <wk@gnupg.org>
* main.c, options.h: New option --fixed_list_mode.
* keylist.c (list_keyblock_colon): use it here.
* getkey.c (merge_keys_and_selfsig): Divert merging of public keys
to the function used in key selection..
* keylist.c (is_uid_valid): Removed.
(list_keyblock): Splitted into ..
(list_keyblock_print, list_keyblock_colon): .. these.
functions. Changed them to use the flags set in the key lookup code.
(reorder_keyblock): New, so that primary user IDs are listed first.
* ringedit.c (keyring_copy): flush the new iobuf chaces before
rename or remove operations. This is mainly needed for W32.
* hkp.c [HAVE_DOSISH_SYSTEM]: Removed the disabled code because we
have now W32 socket support in ../util/http.c
* skclist.c (key_present_in_sk_list): New.
(is_duplicated_entry): New.
(build_sk_list): Check for duplicates and do that before unlocking.
2001-03-12 Werner Koch <wk@gnupg.org>
* armor.c (parse_header_line): Removed double empty line check.
(parse_header_line): Replaced trim_trailing_ws with a counting
function so that we can adjust for the next read.
* options.skel: Fixed 3 typos. By Thomas Klausner. Replaced the
keyserver example by a better working server.
* parse-packet.c (parse_symkeyenc): Return Invalid_Packet on error.
(parse_pubkeyenc): Ditto.
(parse_onepass_sig): Ditto.
(parse_plaintext): Ditto.
(parse_encrypted): Ditto.
(parse_signature): Return error at other places too.
(parse_key): Ditto.
* g10.c (main): Set opt.list_packets to another value when invoked
with the --list-packets command.
* mainproc.c (do_proc_packets): Don's stop processing when running
under --list-packets command.
* signal.c (do_sigaction): Removed.
(init_one_signal): New to replace the above. Needed to support
systems without sigactions. Suggested by Dave Dykstra.
(got_fatal_signal,init_signals): Use the above here.
(do_block): Use sigset() if sigprocmask() is not available.
* armor.c (parse_hash_header): Test on TIGER192, which is the
correct value as per rfc2440. By Edwin Woudt.
2001-03-08 Werner Koch <wk@gnupg.org>
* misc.c: Include time.h. By James Troup.
* getkey.c: Re-enabled the unknown user Id and PK caches and
increased their sizes.
* getkey.c (merge_selfsigs_main): Set expire date and continue
processing even if we found a revoked key.
(merge_selfsigs_subkeys): Ditto.
* packet.h: Add an is_revoked flag to the user_id packet.
* getkey.c (fixup_uidnode): Set that flag here.
(merge_selfsigs_main): Fix so that the latest signature is used to
find the self-signature for an UID.
* parse-packet.c (parse_user_id): Zero out all fields.
* mainproc.c (check_sig_and_print): Print the primary user ID
according the the node flag and then all other non-revoked user IDs.
(is_uid_revoked): Removed; it is now handled by the key selection code.
Changed the year list of all copyright notices.
2001-03-07 Werner Koch <wk@gnupg.org>
* getkey.c (finish_lookup): Print an info message only in verbose mode.
2001-03-05 Werner Koch <wk@gnupg.org>
* packet.h: Replaced sigsubpkt_t value 101 by PRIV_VERIFY_CACHE.
We have never used the old value, so we can do this without any harm.
* parse-packet.c (dump_sig_subpkt): Ditto.
(parse_one_sig_subpkt): Parse that new sub packet.
* build-packet.c (build_sig_subpkt): Removed the old one from the
hashed area.
(delete_sig_subpkt): New.
(build_sig_subpkt): Allow an update of that new subpkt.
* sig-check.c (check_key_signature2): Add verification caching
(cache_selfsig_result): New.
* export.c (do_export_stream): Delete that sig subpkt before exporting.
* import.c (remove_bad_stuff): New.
(import): Apply that function to all imported data
2001-03-03 Werner Koch <wk@gnupg.org>
* getkey.c: Introduced a new lookup context flag "exact" and used
it in all place where we once used primary.
(classify_user_id2): Replaced the old function and add an extra
argument to return whether an exact keyID has been requested.
(key_byname): Removed the unused ctx.primary flag
(get_seckey_byname2): Ditto.
(finish_lookup): Changed debugging output.
2001-03-02 Werner Koch <wk@gnupg.org>
* keylist.c (list_one): Remove the merge key calls.
2001-03-01 Werner Koch <wk@gnupg.org>
* getkey.c (finish_lookup): Don't use it if we no specific usage
has been requested.
(merge_selfsigs_main): fix UID only if we have an signature.
(lookup): Return UNU_PUBKEY etc. instead of NO_PUBKEY if we found
a key but the requested usage does not allow this key.
* import.c (import_one): Take UNU_PUBKEY into account.
* mainproc.c (list_node): Ditto.
* keylist.c (list_keyblock): Ditto.
* keyedit.c (print_and_check_one_sig): Ditto.
2001-02-09 Werner Koch <wk@gnupg.org>
* delkey.c (delete_key): Removed that silly assert which rendered
the whole new stuff meaningless.
2001-02-08 Werner Koch <wk@gnupg.org>
* getkey.c (key_byname): It can happen that we have both, sk and pk
NULL, fix for that.
* parse-packet.c (parse_one_sig_subpkt): Add support for
primary_uid and key_flags.
(can_handle_critical): Ditto
* parse-packet.c (parse_encrypted): Fixed listing of pktlen for
MDC packets.
* getkey.c: Backported the version of this file from gpg 1.1. this
involved some changes in other files too.
* parse-packet.c (parse_key): Clear req_usage.
* skclist.c (build_sk_list): Use req_usage to pass the usage
information to the lookup function.
* pkclist.c (build_pk_list): Ditto.
* free-packet.c (copy_public_parts_to_secret_key): New.
* keydb.h: Add IS_* macros to check the sig_class.
* misc.c (openpgp_cipher_test_algo): New.
(openpgp_pk_test_algo): New.
(openpgp_pk_algo_usage): New.
(openpgp_md_test_algo): New.
* packet.h: Add a few fields to PKT_{public,secret}_key and
PKT_user_id.
* seckey-cert.c (do_check): Use the new main_keyid field.
2001-02-04 Werner Koch <wk@gnupg.org>
* encr-data.c (decrypt_data): Catch error when we had problems to
parse the encrypted packet. By Timo.
2001-01-29 Werner Koch <wk@gnupg.org>
* g10.c (main): --batch does now set nogreeting.
* delkey.c (do_delete_key): Fixed delete-both functionality.
2001-01-22 Werner Koch <wk@gnupg.org>
* g10.c: New command --delete-secret-and-public-key.
* delkey.c (delete_key): Add new arg allow_both.
(do_delete_key): Move most stuff from above to this new function.
2001-01-12 Werner Koch <wk@gnupg.org>
* passphrase.c (passphrase_to_dek): Use MD5 when IDEA is installed
and we have no S2K.
* mainproc.c (proc_encrypted): Likewise
2001-01-11 Werner Koch <wk@gnupg.org>
* sig-check.c (do_check): Print the signature key expire message
only in verbose mode and added the keyID.
2001-01-09 Werner Koch <wk@gnupg.org>
* status.c, status.h: New status USERID_HINT.
(write_status_text): Replace LF and CR int text by C-escape sequence.
* passphrase.c (passphrase_to_dek): Fixed the NEED_PASSPHRASE
output. It does now always print 2 keyIDs. Emit the new
USERID_HINT.
2001-01-08 Werner Koch <wk@gnupg.org>
* g10.c, options.h: New option --no-expensive-trust-checks.
* keylist.c (list_keyblock): Act on this option.
2001-01-04 Werner Koch <wk@gnupg.org>
* g10.c (main): Set homedir only in the pre-parsing phase and
replace backslashes in the W32 version.
2001-01-03 Werner Koch <wk@gnupg.org>
* status.c, status.h : New status KEY_CREATED
* keygen.c (do_generate_keypair,generate_subkeypair): Emit it.
2000-12-28 Werner Koch <wk@gnupg.org>
* signal.c (got_fatal_signal): Remove lockfiles here because the
atexit stuff does not work due to the use of raise. Suggested by
Peter Fales.
* gpgv.c (remove_lockfiles): New stub.
2000-12-19 Werner Koch <wk@gnupg.org>
* status.c, status.h (cpr_get_no_help): New.
* keyedit.c (keyedit_menu): Use it here because we have our own
help list here.
2000-12-18 Werner Koch <wk@gnupg.org>
* mainproc.c (print_failed_pkenc): Don't print the sometimes
confusing message about unavailabe secret key. Renamed ...
(print_pkenc_list): ... to this and introduced failed arg.
(proc_encrypted): Print the failed encryption keys and then
the one to be used.
(proc_pubkey_enc): Store also the key we are going to use.
* mainproc.c (check_sig_and_print): Don't list revoked user IDs.
(is_uid_revoked): New.
2000-12-08 Werner Koch <wk@gnupg.org>
* pipemode.c: Made the command work. Currently only for
non-armored detached signatures.
* mainproc.c (release_list): Reset the new pipemode vars.
(add_gpg_control): Handle the control packets for pipemode
* status.c, status.h: New stati {BEGIN,END}_STREAM.
2000-12-07 Werner Koch <wk@gnupg.org>
* g10.c: New option --allow-secret-key-import.
* import.c (import_keys,import_keys_stream): Honor this option.
(import): New arg allow_secret and pass that arg down to ...
(import_secret_one): to this and print a warning if secret key
importing is not allowed.
2000-12-05 Werner Koch <wk@gnupg.org>
* cipher.c (cipher_filter): Moved the end_encryption status ...
* encode.c (encode_simple,encode_crypt): to here
* sign.c (sign_file): and here.
* status.c (mywrite): Removed.
(get_status_string): Removed the LFs from the strings.
(set_status_fd,is_status_enabed,write_status_text,
write_status_buffer): Replaced all mywrite by stdio calls and use
fdopen to create a strem. This is needed to make things smoother
in the W32 version.
2000-12-04 Werner Koch <wk@gnupg.org>
* import.c (merge_blocks): Increment n_sigs for revocations.
2000-11-30 Werner Koch <wk@gnupg.org>
* g10.c (main): Use iobuf_translate_file_handle for all options
with filehandles as arguments. This is function does some magic
for the W32 API.
* verify.c (verify_signatures): Add a comment rant about the
detached signature problem.
* mainproc.c (proc_tree): Issue an error if a detached signature
is assumed but a standard one was found.
* plaintext.c (hash_datafiles): Don't fall back to read signature
from stdin.
* openfile.c (open_sigfile): Print verbose message only if the
file could be accessed.
2000-11-24 Werner Koch <wk@gnupg.org>
* passphrase.c [HAVE_DOSISH_SYSTEM]: Disabled all the agent stuff.
2000-11-16 Werner Koch <wk@gnupg.org>
* g10.c: New option --use-agent
* passphrase.c (agent_open,agent_close): New.
(agent_get_passphrase,agent_clear_passphrase): New.
(passphrase_clear_cache): New.
(passphrase_to_dek): Use the agent here.
* seckey-cert.c (do_check): Clear cached passphrases.
2000-11-15 Werner Koch <wk@gnupg.org>
* status.c (write_status_text): Moved the big switch to ...
(get_status_string): ... new function.
(write_status_buffer): New.
* status.c (mywrite): New and replaced all write() by this.
* status.c, status.h: Add 3 status lcodes for notaions and policy.
* mainproc.c (print_notation_data): Do status output of notations.
2000-11-13 Werner Koch <wk@gnupg.org>
* sign.c (clearsign_file): Use LF macro to print linefeed.
2000-11-11 Paul Eggert <eggert@twinsun.com>
Clean up the places in the code that incorrectly use "long" or
"unsigned long" for file offsets. The correct type to use is
"off_t". The difference is important on large-file hosts,
where "off_t" is longer than "long".
* keydb.h (struct keyblock_pos_struct.offset):
Use off_t, not ulong, for file offsets.
* packet.h (dbg_search_packet, dbg_copy_some_packets,
search_packet, copy_some_packets): Likewise.
* parse-packet.c (parse, dbg_search_packet, search_packet,
dbg_copy_some_packets, copy_some_packets): Likewise.
* ringedit.c (keyring_search): Likewise.
* parse-packet.c (parse): Do not use %lu to report file
offsets in error diagnostics; it's not portable.
* ringedit.c (keyring_search): Likewise.
2000-11-09 Werner Koch <wk@gnupg.org>
* g10.c (main): New option --enable-special-filenames.
2000-11-07 Werner Koch <wk@gnupg.org>
* g10.c (main): New command --pipemode.
* pipemode.c: New.
2000-10-23 Werner Koch <wk@gnupg.org>
* armor.c (armor_filter): Changed output of hdrlines, so that a CR
is emitted for DOS systems.
* keygen.c (read_parameter_file): Add a cast for isspace().
* status.c (myread): Use SIGINT instead of SIGHUP for DOS.
2000-10-19 Werner Koch <wk@gnupg.org>
* g10.c: New option --ignore-crc-error
* armor.c (invalid_crc): New.
(radix64_read): Act on new option.
* openfile.c (try_make_homedir): Klaus Singvogel fixed a stupid
error introduced on Sep 6th.
2000-10-18 Werner Koch <wk@gnupg.org>
* misc.c (print_cipher_algo_note): Don't print the note for AES.
Changed wording.
2000-10-16 Werner Koch <wk@gnupg.org>
* mainproc.c (do_proc_packets): Hack to fix the problem that
signatures are not detected when there is a MDC packet but no
compression packet.
* g10.c (print_hashline): New.
(print_mds): Use above func with --with-colons.
* mainproc.c (check_sig_and_print): Detect multiple signatures
and don't verify them.
2000-10-14 Werner Koch <wk@gnupg.org>
* mainproc.c (add_onepass_sig): There is an easier solution to the
error fixed yesterday; just check that we only have onepass
packets. However, the other solution provides an cleaner
interface and opens the path to get access to other information
from the armore headers.
(release_list): Reset some more variables.
2000-10-13 Werner Koch <wk@gnupg.org>
* mainproc.c (add_gpg_control): New.
(do_proc_packets): use it.
(proc_plaintext): Changed logic to detect clearsigns.
(proc_tree): Check the cleartext sig with some new code.
* packet.h: New packet PKT_GPG_CONTROL.
* parse-packet.c (parse_gpg_control): New.
* misc.c (get_session_marker): New.
* armor.c (armor_filter): Replaced the faked 1-pass packet by the
new control packet.
* keyedit.c (keyedit_menu): Allow batchmode with a command_fd.
* status.c (my_read): New.
(do_get_from_fd): use it.
2000-10-12 Werner Koch <wk@gnupg.org>
* keygen.c (keygen_add_std_prefs): Add Rijndael to the prefs.
2000-10-07 Werner Koch <wk@gnupg.org>
* gpgv.c: Add more stubs for ununsed code to make the binary smaller.
Wed Oct 4 15:50:18 CEST 2000 Werner Koch <wk@openit.de>
* sign.c (hash_for): New arg to take packet version in account, changed
call callers.
* gpgv.c: New.
* Makefile.am: Rearranged source files so that gpgv can be build with
at least files as possible.
Mon Sep 18 12:13:52 CEST 2000 Werner Koch <wk@openit.de>
* hkp.c (not_implemented): Print a notice for W32
Fri Sep 15 18:40:36 CEST 2000 Werner Koch <wk@openit.de>
* keygen.c (keygen_add_std_prefs): Changed order of preferences to
twofish, cast5, blowfish.
* pkclist.c (algo_available): Removed hack to disable Twofish.
Thu Sep 14 17:45:11 CEST 2000 Werner Koch <wk@openit.de>
* parse-packet.c (dump_sig_subpkt): Dump key flags. Print special
warning in case of faked ARRs.
* getkey.c (finsih_lookup): Hack so that for v4 RSA keys the subkey
is used for encryption.
Thu Sep 14 14:20:38 CEST 2000 Werner Koch <wk@openit.de>
* g10.c (main): Default S2K algorithms are now SHA1 and CAST5 - this
should solve a lot of compatibility problems with other OpenPGP
apps because those algorithms are SHOULD and not optional. The old
way to force it was by using the --openpgp option whith the drawback
that this would disable a couple of workarounds for PGP.
* g10.c (main): Don't set --quite along with --no-tty. By Frank Tobin.
* misc.c (disable_core_dump): Don't display a warning here but a return
a status value and ...
* g10.c (main): ...print warnining here. Suggested by Sam Roberts.
Wed Sep 13 18:12:34 CEST 2000 Werner Koch <wk@openit.de>
* keyedit.c (keyedit_menu): Allow to use "debug" on the secret key.
* ringedit.c (cmp_seckey): Fix for v4 RSA keys.
* seckey-cert.c (do_check): Workaround for PGP 7 bug.
Wed Sep 6 17:55:47 CEST 2000 Werner Koch <wk@openit.de>
* misc.c (print_pubkey_algo_note): Do not print the RSA notice.
* sig-check.c (do_signature_check): Do not emit the RSA status message.
* pubkey-enc.c (get_session_key): Ditto.
* encode.c (encode_simple, encode_crypt): Fix for large files.
* sign.c (sign_file): Ditto.
Wed Sep 6 14:59:09 CEST 2000 Werner Koch <wk@openit.de>
* passphrase.c (hash_passphrase): Removed funny assert. Reported by
David Mathog.
* openfile.c (try_make_homedir): Changes for non-Posix systems.
* g10.c (main): Take the default homedir from macro.
* g10.c: The --trusted-key option is back.
* trustdb.c (verify_own_key): Handle this option.
(add_ultimate_key): Moved stuff from verify_own_key to this new func.
(register_trusted_key): New.
Fri Aug 25 16:05:38 CEST 2000 Werner Koch <wk@openit.de>
* parse-packet.c (dump_sig_subpkt): Print info about the ARR.
* openfile.c (overwrite_filep): Always return okay if the file is
called /dev/null.
(make_outfile_name): Add ".sign" to the list of know extensions.
(open_sigfile): Ditto.
Wed Aug 23 19:52:51 CEST 2000 Werner Koch <wk@openit.de>
* g10.c: New option --allow-freeform-uid. By Jeroen C. van Gelderen.
* keygen.c (ask_user_id): Implemented here.
Fri Aug 4 14:23:05 CEST 2000 Werner Koch <wk@openit.de>
* status.c (do_get_from_fd): Ooops, we used fd instead of opt.command_fd.
Thanks to Michael Tokarev.
Tue Aug 1 20:06:23 CEST 2000 Werner Koch <wk@openit.de>
* g10.c: New opttion --try-all-secrets on suggestion from Matthias Urlichs.
* pubkey-enc.c (get_session_key): Quite easy to implement here.
Thu Jul 27 17:33:04 CEST 2000 Werner Koch <wk@openit.de>
* g10.c: New option --merge-only. Suggested by Brendan O'Dea.
* import.c (import_one): Implemented it here
(import_secret_one): Ditto.
(print_stats): and give some stats.
Thu Jul 27 12:01:00 CEST 2000 Werner Koch <wk@openit.de>
* g10.c: New options --show-session-key and --override-session-key
* pubkey-enc.c (hextobyte): New.
(get_override_session_key): New.
* mainproc.c (proc_pubkey_enc): Add session-key stuff.
* status.h, status.c (STATUS_SESSION_KEY): New.
Thu Jul 27 10:02:38 CEST 2000 Werner Koch <wk@openit.de>
* g10.c (main): Use setmode(O_BINARY) for MSDOS while generating random bytes
(print_mds): Likewise for stdin.
* plaintext.c (handle_plaintext): Likewise for stdout.
Mon Jul 24 10:30:17 CEST 2000 Werner Koch <wk@openit.de>
* keyedit.c (menu_expire): expire date for primary key can be set again.
Wed Jul 19 11:26:43 CEST 2000 Werner Koch <wk@openit.de>
* keylist.c (is_uid_valid): New.
(list_keyblock): Print validity information for all user IDs. Note, this
has to be done at other places too; for now we have only minimal support.
Wed Jul 12 13:32:06 CEST 2000 Werner Koch <wk@openit.de>
* helptext.c, pkclist.c: s/superseeded/superseded/
Mon Jul 10 16:08:57 CEST 2000 Werner Koch <wk@openit.de>
* parse-packet.c (enum_sig_subpkt): Fixed testing on crtitical bit in case
of a NULL buffer. Reported by Peter Marschall.
Wed Jul 5 13:28:45 CEST 2000 Werner Koch <wk@openit.de>
* keyedit.c, keyid.c: Add some _()
* argparse.c: Changed the flag to suppress --version handling to also
suppress --help.
Wed Jun 28 11:54:44 CEST 2000 Werner Koch <wk@openit.de>
* armor.c (armor_filter): Set sigclass to 0 in case of non-dash-escaped
clearsig. This makes this mode work again.
* mainproc.c (proc_tree): Fixed handling of one-pass-sig packets in textmode.
Disabled the ugly workaround for PGP 5 - let's see whether thi breaks less
cases. Found by Ted Cabeen.
* options.h (DBG_HASHING): New. All commented md_start_debug are now
controlled by this debug option.
* sign.c (print_status_sig_created): New and called from 2 places.
* keygen.c (gen_rsa): New, but commented.
(ask_algo): Commented support for RSA.
* seckey-cert.c (protect_secret_key): Started to fix the code for v4 RSA
keys - it is not solved yet. However, we have time until, Sep 20th ;)
Wed Jun 14 12:27:09 CEST 2000 Werner Koch <wk@openit.de>
* status.c (init_shm_coprocessing): Changed the sequence of the get,attach
to cope with the changes in newer Linux kernels. This bug has been found
by <dmitri@advantrix.com> who also proposed this solution. Hopefully
this does not break gpg on to many systems.
* cipher.c (write_header): Protect the IV with the MDC too.
* encr-data.c (decrypt_data): Likewise.
Fri Jun 9 10:09:52 CEST 2000 Werner Koch <wk@openit.de>
* g10.c: New options --no-auto-key-retrieve
* options.h (auto_key_retrieve): New.
* mainproc.c (check_sig_and_print): Implemented that.
Wed Jun 7 19:19:09 CEST 2000 Werner Koch <wk@openit.de>
* sig-check.c (do_check): Use EMULATE_MDENCODE also on v4 packets.
Wed Jun 7 17:25:38 CEST 2000 Werner Koch <wk@openit.de>
* cipher.c (write_header): Use plain CFB mode for MDC encrypted packets.
* encr-data.c (decrypt_data): Ditto.
Mon Jun 5 23:41:54 CEST 2000 Werner Koch <wk@openit.de>
* seskey.c (do_encode_md, encode_md_value): Add new arg v3compathack to work
around a bug in old versions.
* sig-check.c (do_check): use the aboved workaround when enabled.
* g10.c: New option --emulate-md-decode-bug
Mon Jun 5 12:37:43 CEST 2000 Werner Koch <wk@openit.de>
* build-packet.c (do_mdc): New.
(do_encrypted_mdc): Changed for the new proposal.
* parse-packet.c (parse_mdc): New.
(parse_encrypted): Fixed for the new proposal.
* packet.h (PKT_MDC): New.
* cipher.c (cipher_filter): Build the MDC packet here.
* g10.c (main): Enable --force-mdc.
* encr-data.c (mdc_decode_filter): Fixed for new MDC method
* options.h(rfc2440): New.
* g10.c (main): Changed the selected values for --openpgp to not include
optional algorithms.
Thu May 18 11:38:54 CEST 2000 Werner Koch <wk@openit.de>
* keyedit.c (keyedit_menu): Add a keyword arg to the prompt.
* status.c, status.h: Added 3 new status tokens.
* status.c (do_get_from_fd): New.
(cpr_enabled,cpr_get,cpr_get_hidden,cpr_kill_prompt,
cpr_get_answer_is_yes,cpr_get_answer_yes_no_quit): Modified to work
with the new function.
* g10.c: Add new option --command-fd.
* status.c (progress_cb): New.
(set_status_fd): Register progress functions
Fri May 12 14:01:20 CEST 2000 Werner Koch <wk@openit.de>
* delkey.c (delete_key): Add 2 new status messages
* status.c, status.h (STATUS_DELETE_PROBLEM): New.
Fixed years of copyright in all source files.
Mon May 1 17:08:14 CEST 2000 Werner Koch <wk@openit.de>
* trustdb.c (propagate_validity): Fixed the bug that only one uid
gets fully trusted even when all are signed by an ultimate key.
Mon May 1 15:38:04 CEST 2000 Werner Koch <wk@openit.de>
* getkey.c (key_byname): Always returned a defined context. Fixed
a segv for invalid user id specifications. Reported by Walter Koch.
* getkey.c (get_user_id): I18ned "no user id" string. By Walter.
* pkclist.c (do_show_revocation_reason): Typo fixes.
* helptext.c: Ditto.
* armor.c (armor_filter): Fixed some CRLF issues. By Mike McEwan.
Fri Apr 14 19:37:08 CEST 2000 Werner Koch <wk@openit.de>
* pkclist.c (do_show_revocation_reason): New.
(show_revocation_reason): New and called at various places.
* g10.c (main): Fixed small typo.
* pkclist.c (do_we_trust): Act on always_trust but not for revoked
keys. Suggested by Chip Salzenberg.
* g10.c: New option --lock-never.
* ringedit.c (get_writable_keyblock_file): New.
* keygen.c (do_generate_keypair): Use this instead of the hardwired one.
* keygen.c (ask_user_id): Check that the email address is in the
correct field. Suggested by Christian Kurz.
Mon Apr 10 13:34:19 CEST 2000 Werner Koch <wk@openit.de>
* keyedit.c (show_key_with_all_names): s/sbb/ssb/
Tue Mar 28 14:26:58 CEST 2000 Werner Koch <wk@openit.de>
* trustdb.c (verify_own_keys): Do not print warning about unprotected
key when in quiet mode.
Wed Mar 22 13:50:24 CET 2000 Werner Koch <wk@openit.de>
* mainproc.c (print_userid): Do UTF8 conversion before printing.
* import.c (import_one): Ditto.
(import_secret_one): Ditto.
(delete_inv_parts): Ditto.
Thu Mar 16 16:20:23 CET 2000 Werner Koch <wk@openit.de>
* keylist.c (print_key_data): Handle a NULL pk gracefully.
* getkey.c (merge_one_pk_and_selfsig): Fixed silly code for
getting the primary keys keyID but kept using the one from the
subkey.
* pubkey-enc.c (get_it): Print a note for expired subkeys.
* getkey.c (has_expired): New.
(subkeys_expiretime): New.
(finish_lookup): Check for expired subkeys needed for encryption.
(merge_keys_and_selfsig): Fixed expiration date merging for subkeys.
* keylist.c (list_keyblock): Print expiration time for "sub".
(list_one): Add missing merging for public keys.
* mainproc.c (list_node): Ditto.
2000-03-14 13:49:38 Werner Koch (wk@habibti.openit.de)
* keygen.c (keyedit_menu): Do not allow to use certain commands
while the secret key is selected.
2000-03-09 12:53:09 Werner Koch (wk@habibti.openit.de)
* keygen.c (ask_expire_interval): Movede parsig to ...
(parse_expire_string): ... this new function. And some new control
commands.
(proc_parameter_file): Add expire date parsing.
(do_generate_keypair): Allow the use of specified output files.
2000-03-08 10:38:38 Werner Koch (wk@habibti.openit.de)
* keygen.c (ask_algo): Removed is_v4 return value and the commented
code to create Elg keys in a v3 packet. Removed the rounding
of key sizes here.
(do_create): Likewise removed arg v4_packet.
(gen_elg): Likewise removed arg version. Now rounding keysizes here.
(gen_dsa): Rounding keysize now here.
(release_parameter_list): New
(get_parameter*): New.
(proc_parameter_file): New.
(read_parameter_file): New.
(generate_keypair): Splitted. Now uses read_parameter_file when in
batch mode. Additional argument to specify a parameter file.
(do_generate_keypair): Main bulk of above fucntion and uses the
parameter list.
(do_create): Don't print long notice in batch mode.
* g10.c (main): Allow batched key generation.
Thu Mar 2 15:37:46 CET 2000 Werner Koch <wk@gnupg.de>
* pubkey-enc.c (get_it): Print a note about unknown cipher algos.
* g10.c (opts): Add a note to the help listing about the man page
and removed some options from the help listing.
* keyedit.c (print_and_check_one_sig): Use a new function to truncate
the output of the user ID. Suggested by Jan-Benedict Glaw.
Wed Feb 23 10:07:57 CET 2000 Werner Koch <wk@gnupg.de>
* helptext.c: typo fix.
Thu Feb 17 13:39:32 CET 2000 Werner Koch <wk@gnupg.de>
* revoke.c: Removed a bunch of commented code.
* packet.h (SIGSUBPKT_REVOC_REASON): New.
* build-packet.c (build_sig_subpkt): Support new sub packet.
* parse-packet.c (parse_one_sig_subpkt): Ditto.
(dump_sig_subpkt): Ditto.
* revoke.c (ask_revocation_reason): New.
(release_revocation_reason_info): New.
(revocation_reason_build_cb): New.
(gen_revoke): Ask for reason.
* main.h (struct revocation_reason_info): Add declaration.
* keyedit.c (menu_revsig): Add support for revocation reason.
(menu_revkey): Ditto.
(sign_uid_mk_attrib): Renamed to ...
(sign_mk_attrib): ... this, made static and add support for reasons.
Tue Feb 15 08:48:13 CET 2000 Werner Koch <wk@gnupg.de>
* build-packet.c (build_packet): Fixed fixing of old comment packets.
* import.c (import_keys): Fixed importing from stdin when called with
nnames set to zero as it normally happens.
Mon Feb 14 14:30:20 CET 2000 Werner Koch <wk@gnupg.de>
* sig-check.c (check_key_signature2): Add new arg r_expired.
(do_signature_check): New arg to pass it down to ...
(do_check): New arg r-expire which is set when the signature
has expired.
* trustdb.c (check_sig_record): Set SIGF_EXPIRED flag and set
the expiretime to zero so that thi signature will not be checked
anymore.
Fri Feb 11 17:44:40 CET 2000 Werner Koch <wk@gnupg.de>
* g10.c (g10_exit): Update the random seed_file.
(main): Set the random seed file. New option --no-random-seed-file.
Thu Feb 10 17:39:44 CET 2000 Werner Koch <wk@gnupg.de>
* keyedit.c (menu_expire): Fixed segv due to unitialized sub_pk.
By Rémi.
Thu Feb 10 11:39:41 CET 2000 Werner Koch <wk@gnupg.de>
* keylist.c (list_keyblock): Don't print warnings in the middle of
regulat output lines. By Rémi.
* sig-check.c: Include options.h
Wed Feb 9 15:33:44 CET 2000 Werner Koch <wk@gnupg.de>
* gpg.c: New option --ignore-time-conflict
* sig-check.c (do_check): Implemented this option.
* trustdb.c (check_trust): Ditto.
* sign.c (do_sign): Ditto.
* keygen.c (generate_subkeypair): Ditto.
* encode.c (encode_simple): use iobuf_cancel after open failure.
Reported by Huy Le.
Fri Jan 14 18:32:01 CET 2000 Werner Koch <wk@gnupg.de>
* packet.h (STRING2KEY): Changed mode from byte to int.
* parse-packet.c (parse_key): Add the special GNU protection stuff
* build-packet.c (so_secret_key): Ditto.
* seckey-cert.c (do_check): Ditto.
* keyedit.c (change_passphrase): Ditto.
* export.c (export_secsubkeys): New.
(do_export_stream): Hack to export the primary key using mode 1001.
* g10.c: New command --export-secret-subkeys
Thu Jan 13 19:31:58 CET 2000 Werner Koch <wk@gnupg.de>
* armor.c (is_armored): Check for 1-pass-sig packets. Reported by
David Hallinan <hallinan@rtd.com>.
(armor_filter): Replaced one LF by the LF macro. Reported by
Wolfgang Redtenbacher.
Wed Jan 5 11:51:17 CET 2000 Werner Koch <wk@gnupg.de>
* g10.c (main): Reset new global flag opt.pgp2_workarounds
when --openpgp is used.
* mainproc.c (proc_plaintext): Do the PGP2,5 workarounds only
when the global flag is set.
(proc_tree): Ditto.
* textfilter.c (copy_clearsig_text): Ditto.
* armor.c (armor_filter): Ditto.
* g10.c: New option --list-only
* mainproc.c (proc_tree): Don't do it if opt.list_only is active.
(proc_pubkey_enc): Implement option.
* status.h, status.c ({BEGIN,END}_{EN,DE}CRYPTION): New.
* cipher.c (cipher_filter): New status outputs.
* mainproc.c (proc_encrypted): New status outputs.
Fri Dec 31 14:08:15 CET 1999 Werner Koch <wk@gnupg.de>
* armor.c (armor_filter): Made the "Comment:" header translatable.
* hkp.c (hkp_import): Make sure that the program does not return
success when there is a connection problem. Reported by Phillip Jones.
Sun Dec 19 15:22:26 CET 1999 Werner Koch <wk@gnupg.de>
* armor.c (LF): Use this new macro at all places where a line LF
is needed. This way DOSish textfiles should be created when the
input data is also in dos mode.
* sign.c (LF): Ditto.
* textfilter.c (LF): Ditto.
(copy_clearsig_text): Disabled the forcing of CR,LF sequences
for DOS systems.
* plaintext.c (handle_plaintext): Fixes for line endings on DOS.
and react on a LF in cleartext.
* armor.c (fake_packet): Restore the original line ending after
removing trailing spaces.
* signal.c (got_fatal_signal): DOS fix.
Thu Dec 16 10:07:58 CET 1999 Werner Koch <wk@gnupg.de>
* mainproc.c (print_failed_pkenc): Fix for unknown algorithm.
Found by fygrave@epr0.org.
Thu Dec 9 10:31:05 CET 1999 Werner Koch <wk@gnupg.de>
* hkp.c: i18n the strings.
Sat Dec 4 15:32:20 CET 1999 Werner Koch <wk@gnupg.de>
* trustdb.c (verify_key): Shortcut for ultimately trusted keys.
Sat Dec 4 12:30:28 CET 1999 Werner Koch <wk@gnupg.de>
* pkclist.c (build_pk_list): Validate the trust using the namehash
if this one has been set by the key lookup.
* g10.c: Add --delete-secret-key to the help page.
* openfile.c (copy_options_file): Made static.
(try_make_homedir): New.
* ringedit.c (add_keyblock_resource): Use the try_make_hoemdir logic.
* tdbio.c (tdbio_set_dbname): Likewise.
* keygen.c (generate_user_id): Use m_alloc_clear() here. We should
better use an allocation function specific to the user_id packet.
* keygen.c (keygen_add_std_prefs): Changed symmetric preferences
to include Blowfish again. This is due to it's better speed compared
to CAST5.
* g10.c (strusage): Print the home directory.
* armor.c (armor_filter): Take action on the cancel control msg.
* filter.h (armor_filter_context_t): Add cancel flag.
Mon Nov 29 21:52:11 CET 1999 Werner Koch <wk@gnupg.de>
* g10.c: New option --fast-list-mode ..
* keylist.c (list_keyblock): .. and implemented.
* mainproc.c (list_node): Ditto.
* import.c (mark_non_selfsigned_uids_valid): Fixed the case that there
is a uid without any packet following.
Mon Nov 22 11:14:53 CET 1999 Werner Koch <wk@gnupg.de>
* mainproc.c (proc_plaintext): Never enable the hash processing
when skip_verify is active.
* armor.c (parse_header_line): Stop parsing on a WS line too.
Suggested by Aric Cyr.
* tdbdump.c (HEXTOBIN): Changed the name of the argument, so that
traditional cpp don't mess up the macros. Suggested by Jos Backus.
* mainproc.c (list_node): Print the PK algo in the --with-colon mode.
* keylist.c (list_keyblock): Ditto.
* signal.c (got_fatal_signal): Found the reason why exit(8) did not
work - it is better to set the disposition back to default before
raising the signal. Print the notice on stderr always.
Fri Nov 12 20:33:19 CET 1999 Werner Koch <wk@gnupg.de>
* g10.c (make_username): Swapped the logic.
* keylist.c (public_key_list): Now takes a STRLIST as arg and moved
the creation ot this list to the caller, so that he can copy with
UTF-conversion of user IDs. Changed all callers.
(secret_key_list): Likewise.
* getkey.c (get_user_id_string_native): New and ...
* encode.c (write_pubkey_enc_from_list): ... use it here.
* pubring.asc: Updated.
* packet.h (PKT_PHOTO_ID): New.
* parse-packet.c (parse_photo_id): New.
* build-packet.c (do_user_id: Handle photo IDs.
(build_packet): Change CTB for photo IDs
* free-packet.c (free_user_id): Release memory used for photo IDs
* sig-check.c (hash_uid_node): Handle photo IDs too.
* trustdb.c (print_uid_from_keyblock): Hash photo ID.
(make_uid_records): Ditto.
* getkey.c (find_by_name): Ditto.
* keyedit.c (show_prefs): Ditto.
* keylist.c (list_keyblock): Ditto.
Thu Oct 28 16:08:20 CEST 1999 Werner Koch <wk@gnupg.de>
* keygen.c (ask_expire_interval): Print a warning for systems
with a signed 32 time_t if the exiration time is beyoind 2038.
Fri Oct 8 20:40:50 CEST 1999 Werner Koch <wk@gnupg.de>
* ringedit.c (enum_keyblocks): The last fix way really stupid;
reverted and set rt to Unknown.
Fri Oct 8 20:32:01 CEST 1999 Werner Koch <wk@gnupg.de>
* ringedit.c (enum_keyblocks): Zero the entire kbpos out on open.
* g10.c (oEntropyDLL): Removed option.
(main): Made the warning on development versions more verbose.
* g10.c (oHonorHttpProxy): New option.
* hkp.c (hkp_ask_import,hkp_export): Implement this option.
* options.skel: Enable this option for new installations
Mon Oct 4 21:23:04 CEST 1999 Werner Koch <wk@gnupg.de>
* import.c (import_keys): Changed calling interface, adjusted caller.
(import): Moved printing of stats out ...
(print_stats): New. ... to here.
(import_keys_stream): Call stats print here.
(import_keys): Print stats as totals for all files.
* tdbio.h (DIRF_NEWKEYS): New
* tdbio.c (tdbio_dump_record): Print the new flag.
* trustdb.c (check_trust_record): New arg sigs_only. Adapted all
callers.
(do_update_trust_record): Removed recheck arg and add a new sigs_only
do we can later improve on the performance. Changed all callers too.
(check_trustdb): Evalutate the new flag and add a status output.
Do a check when the dir record has not been checked.
(build_cert_tree): Evaluate the new flag.
(check_trust): Ditto. Do a trust_record check, when the dir record
is not marked as checked.
(mark_fresh_keys): New.
(clear_lid_table): New.
(sync_trustdb): New.
* import.c (import_keys): Call sync_trustdb() after processing.
(import_keys_stream): Ditto.
* tdbdump.c (import_ownertrust): Ditto.
* import.c (import_revoke_cert): Notify the trust DB.
(do_update_trust_record): Use |= to set the REVOKED bit and not &=;
shame on me for this bad copy+paste introduced bug.
(do_we_trust): Add trustmask to allow revoked key override to work.
Chnaged are to allow return of a mofified trustlevel. Adapted the
one caller.
* g10.c: New options --emulate-3des-s2k-bug
* passphrase.c (hash_passphrase): Implemented above.
* mainproc.c (proc_tree): Check for standalone signatures.
(do_check_sig): Print a notice for a standalone revocation
(check_sig_and_print): Do not print an error for unchecked standalone
revocations.
Tue Sep 28 20:54:37 CEST 1999 Werner Koch <wk@gnupg.de>
* encode.c (encode_simple): Use new CTB when we don't have the
length of the file. This is somewhat strange as the comment above
indicates that this part is actually fixed for PGP 5 - maybe I simply
lost the source line, tsss.
* armor.c (armor_filter): Set a flag if no OpenPGP data has been found.
* verify.c (verify_signatures): Add an error helptext.
Thu Sep 23 19:24:30 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* openfile.c (open_outfile): Fixed the 8dot3 handling.
* passphrase.c (passphrase_to_dek): Print uid using utf8 func.
* delkey.c (delete_key): Ditto.
* pkclist.c (show_paths,do_edit_ownertrust,do_we_trust): Ditto
(do_we_trust_pre): Ditto.
* trustdb.c (print_user_id,check_uidsigs): Ditto.
* revoke.c (gen_revoke,ask_revoke_sig): Ditto.
Thu Sep 23 09:52:58 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* verify.c (print_file_status): New.
(verify_one_file): Moved status print to th new fnc. Add error status.
* status.c, status.h (STATUS_FILE_ERROR): New
Wed Sep 22 10:14:17 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* openfile.c (make_outfile_name): Use case-insenstive compare for
DOS systems. Add ".pgp" to the list of know extensions.
(open_outfile): For DOS systems try to replace the suffiy instead of
appending it.
* status.c, status.h: Add STATUS_FILE_{START,DONE}.
* verify.c (verify_one_file): Emit these new stati.
* sign.c (clearsign_file): Avoid duplicated Entries in the "Hash:"
line. Those headers are now only _not_ printed when there are
only old-style keys _and_ all hashs are MD5.
Mon Sep 20 12:24:41 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* verify.c (verify_files, ferify_one_file): New.
* g10.c: New command --verify-files
Fri Sep 17 12:56:42 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* g10.c: Add UK spelling as alias for armor options ;-)
* import.c (append_uid): Fixed a SEGV when there is no selfsig and
no subkey.
(merge_sigs): Ditto. Removed the assertion.
Wed Sep 15 16:22:17 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* g10.c: New option --entropy-dll-name
Mon Sep 13 10:51:29 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* signal.c (got_fatal_signal): Print message using write(2) and
only for development versions.
Mon Sep 6 19:59:08 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* tdbio.c (tdbio_set_dbname): Use mkdir macro
* ringedit.c (add_keyblock_resource): Ditto.
Fri Sep 3 10:04:45 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* pkclist.c (build_pk_list): Skip keys set with --encrypt-to also
when asking for a key.
* plaintext.c (handle_plaintext): Make sure that we don't read a
second EOF in the read loop for partial length packets.
* mainproc.c (check_sig_and_print): print user ID as utf-8.
Thu Sep 2 16:40:55 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* import.c (merge_blocks): First add new subkeys, then merge subkey
certificates.
(merge_sigs): Don't merge subkey signatures here.
Wed Sep 1 15:30:44 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* keygen.c (ask_expire_interval): Fixed bug related to cpr_xx (tnx
Francis J. Lacoste).
Tue Aug 31 17:20:44 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* plaintext.c (do_hash): Hash CR,LF for a single CR.
(ask_for_detached_datafile): Changed arguments to be closer to
those of hash_datafiles and cleanup the code a bit.
* mainproc.c (proc_tree): Workaround for pgp5 textmode detached
signatures. Changed behavior of asking for data file to be the same
as with provided data files.
* keylist.c (list_keyblock): Use UTF8 print functions.
Mon Aug 30 20:38:33 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* import.c (chk_self_sigs): some s/log_error/log_info/ so that gpg
does not return an error if a key has some invalid packets.
* helptext.c: Fixed some typos and changed the way the
translation works. The english text is now the keyword for gettext
and not anymore the keyword supplied to the function. Done after
some discussion with Walter who thinks this is much easier for the
translators.
* misc.c (disable_core_dumps): Don't do it for DOSish systems.
* signal.c (signal_name): Bounds check on signum.
Wed Aug 4 10:34:18 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* pubring.asc: Updated.
* pkclist.c (do_we_trust_pre,check_signatures_trust): Do not print
the warning about --always_trust when --quiet is used.
* pkclist.c (fpr_info): New and called at several places.
* parse-packet.c (dump_sig_subpkt): List revocation key contents.
Mon Jul 26 09:34:46 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* pkclist.c (build_pk_list): Fixed typo in format string.
* trustdb.c (create_shadow_dir): Don't translate the error string.
* g10.c (main): Fixed spelling of user-id.
* getkey.c (find_by_name_pk,find_by_name_sk,
find_by_keyid,find_by_keyid_sk): Ditto and translate it.
* import.c (mark_non_selfsigned_uids_valid,delete_inv_parts): Ditto.
Mon Jul 26 01:01:39 CEST 1999 Michael Roth <mroth@nessie.de>
* g10.c, options.h: New options --no-literal and --set-filesize
* encode.c (encode_simple, encode_crypt): Support for the options
--no-literal and --set-filesize.
* sign.c (sign_file): ditto.
Fri Jul 23 13:53:03 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* ringedit.c (enum_keyblocks): Removed annoying error message in cases
when we have no keyring at all to enum.
* getkey.c (classify_user_id): Rewrote to relax the recognition of
keyIDs and fingerprints (Michael).
* mainproc.c (check_sig_and_print): Print status NO_PUBKEY.
(print_failed_pkenc): Print status NO_SECKEY.
* import.c (mark_non_selfsigned_uids_valid): New.
* g10.c: New option --allow-non-selfsigned-uid.
* pkclist.c (print_fpr): New.
(do_we_trust_pre): Print the fpr before asking whether to use the key
anyway.
(do_edit_ownertrust): Likewise.
Thu Jul 22 20:03:03 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* ringedit.c (enum_keyblocks): Removed annoying error message in cases
when we have no keyring at all to enum.
* getkey.c (classify_user_id): Rewrote to relax the recognition of
keyIDs and fingerprints (Michael).
* mainproc.c (check_sig_and_print): Print status NO_PUBKEY.
(print_failed_pkenc): Print status NO_SECKEY.
* import.c (mark_non_selfsigned_uids_valid): New.
* g10.c: New option --allow-non-selfsigned-uid.
Thu Jul 15 10:15:35 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* g10.c: New options --disable-{cipher,pubkey}-algo.
Wed Jul 14 19:42:08 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* status.h (STATUS_IMPORTED): New.
* import.c (import): Print some status information (Holger Schurig).
* g10.c (main): Make --no-greeting work again. Add a warning when
--force-mds is used.
Tue Jul 13 17:39:25 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* pkclist.c (do_edit_ownertrust): Changed the way help works.
(build_pk_list): Implemented default recipient stuff.
* g10.c: New options --default-recipient[-self]
(main): Suppress greeting in most cases, entering a passphrase or
a missing value is not considered to be interactive use.
Merged --print-md and --print-mds; the latter is now obsolete.
Changed the way --gen-random works and documented it.
Changed the way --gen-prime works and add a man entry.
* g10.c (MAINTAINER_OPTIONS): Removed.
Mon Jul 12 18:45:57 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* keyedit.c (keyedit_menu): Add arg sign_mode and changed callers
* g10.c (main): New command --lsign-key.
Mon Jul 12 14:55:34 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* mainproc.c (kidlist_item): New.
(release_list): Release failed pk-enc-list.
(print_failed_pkenc): New
(proc_encrypted): Print info about failed PK enc.
* openfile.c (make_outfile_name): s/error/info/
* passphrase.c (passphrase_to_dek): Return an empty passphrase when
in batch mode and don't make the warning message fatal
* seckey-cert.c (check_secret_key): Try only once when in batch mode.
* g10.c (make_username): New.
Thu Jul 8 16:21:27 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* packet.h (PKT_ring_trust): New
* parse-packet.c (parse_trust): Store trust value
* build-packet (build_packet): Ignore ring trust packets.
* mainproc.c (add_ring_trust): New.
(list_node): Print "rtv" records.
* g10.c: New option --with-fingerprint.
* trustdb.c (verify_own_keys): Don't insert if we are dry running
(check_trust): Ditto.
Wed Jul 7 13:08:40 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* Makefile.am: Support for libtool.
* keygen.c (ask_expire_interval): Hack to allow for an expire date.
* trustdb.c (do_update_trust_record,update_trust_record): Splitted.
(check_trust_record): New.
(check_trust,build_cert_tree): Check the dir record as needed.
(upd_pref_record): Removed.
(make_pref_record): New.
(propagate_validity): Stop as soon as we have enough validity.
* tbdio.c (MAX_CACHE_ENTRIES_HARD): Increased the limit.
Fri Jul 2 11:45:54 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* g10.c (g10_exit): Dump random stats.
* sig-check.c (check_key_signature,check_key_signature2): Enhanced
version and wrapper for old function.
(do_signature_check,signature_check): Ditto.
Thu Jul 1 12:47:31 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* keyedit.c (show_key_with_all_names): Print a notice for disabled keys.
(enable_disable_keys): Add functionality
* pkclist.c (edit_ownertrust): preserve disabled state.
(build_pk_list): Skip disabled keys.
* trustdb.c (upd_one_ownertrust): Ditto.
(build_cert_tree): Mask the ownertrust.
(trust_letter): Mask the value.
(do_check): Take disabled flag into account.
* passphrase.c (passphrase_to_dek): Add a pubkey_algo arg and changed
all callers.
* g10.c (utf8_strings): 2 new options.
* trustdb.c (insert_trust_record_by_pk): New, replaces the next one.
(insert_trust_record): Now takes a keyblock as arg. Changed all
callers to use the appropritae function.
* openfile.c (ask_outfile_name): New.
* plaintext.c (handle_plaintext): Ask for filename if there is
no valid syntax. Don't use fname varbatim but filter it.
Tue Jun 29 21:44:25 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* trustdb.h (TRUST_FLAG_DISABLED): New.
* status.c (USE_CAPABILITIES): Capabilities support (Remi).
* tdbio.c : Added new fields to the DIR record.
(tdbio_write_record): Fixed the update of the hash tables.
(tdbio_delete_record): Drop the record from the hash tables.
(drop_from_hashtbl): New.
* status.c (cpr_get): Special online help mode.
* helptext.c ("keyedit.cmd"): Removed.
* keyedit.c (keyedit_menu): Use only help system.
(enable_disable_key): New bit doies not yet work.
Sat Jun 26 12:15:59 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* dearmor.c (enarmor_file): Fixed comment string.
* tdbdump.c (export_ownertrust): Text fix.
* tbio.c (tdbio_invalid): Ditto.
* parse-packet.c (parse_key): Made temp buffer larger.
* Makefile.am (install-data-local): Add missing backslashes
Tue Jun 15 12:21:08 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* g10.c (main): Made iterated+salted the default S2K method.
* Makefile.am (install-data-local): Use DESTDIR.
* passphrase.c (passphrase_to_dek): Emit missing-passphrase while in
batchmode.
* parse-packet.c (parse_pubkeyenc): Fixed a SEGV.
Mon Jun 14 21:18:54 CEST 1999 Michael Roth <mroth@nessie.de>
* g10.c: New options --openpgp, --no-tty, --emit-version,
--default-comment and --lock-multiple
Thu Jun 10 14:18:23 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* free-packet.c (free_encrypted): Fixed EOF case (Remi).
(free_plaintext): Ditto.
* helptext.c (keyedit.delsig.unknown): New (Remi).
* keyedit.c (print_and_check_one_sig): Add arg print_without_key and
changed all callers to make use of it (Remi):
Tue Jun 8 13:36:25 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* keylist.c (print_key_data): New and called elsewhere.
* g10.c: New option --with-key-data
Wed Jun 2 14:17:19 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* mainproc.c (proc_tree): Yet another bad hack to cope with
broken pgp2 created detached messages in textmode.
Tue Jun 1 16:01:46 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* openfile.c (make_outfile_name): New.
* plaintext.c (handle_plaintext): Outputfile is now the inputfile
without the suffix.
* g10.c: New option --use-embedded-filename
Mon May 31 19:41:10 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* g10.c (main): Fix for SHM init (Michael).
* compress.c, encr-data.c, mdfilter.c,
plaintext.c, free-packet.c: Speed patches (Rémi).
Thu May 27 09:40:55 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* status.c (cpr_get_answer_yes_no_quit): New.
* keyedit.c (menu_delsig): New.
(check_all_keysigs): Splitted.
(print_and_check_one_sig): New.
Wed May 26 14:36:29 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* build-packet.c (build_sig_subpkt): Support large packets.
* parse-packet.c (enum_sig_subpkt): Replaces parse_sig_subpkt.
* mainproc.c (print_notation_data): Print all notation packets.
* g10.c (add_notation_data): Add a way to specify the critical flag.
(main): Add option --set-policy-url.
(check_policy_url): Basic checks.
* sign.c (mk_notation_and_policy): Replaces mk_notation.
* parse-packet.c (can_handle_critical): Moved decision whether we can
handle critical subpacket to an extra function.
Tue May 25 19:50:32 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* sign.c (sign_file): Always use compression algo 1 for signed
onyl file becuase we can´ be sure the the verifier supports other
algorithms.
* build-packet.c (build_sig_subpkt): Support for notation data.
* sign.c (sign_file,clearsign_file,make_keysig_packet): Ditto.
(mk_notation): New.
* g10.c (add_notation_data): New and add option -N
* mainproc.c (print_notation_data): New.
(check_sig_and_print): Print any notation data of the signed text.
Sun May 23 14:20:22 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* pkclist.c (check_signatures_trust): Print a warning and return
immediateley if opt.always_trust is true.
* g10.c (main): Corrected handling of no-default-keyring
* pkclist.c (algo_available): Disable Twofish until we have settled
how to do the MDC.
* hkp.c: Disable everything for mingw32
Sat May 22 22:47:26 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* mainproc.c (check_sig_and_print): Add sig creation time to the
VALIDSIG status output. Add more info to the ERRSIG output.
* sig-check.c (signature_check): Add sig time after epoch to SIG_ID.
* import.c (import_one): Merge duplicate user IDs.
(collapse_uids): New.
* kbnode.c (move_kbnode): New.
(remove_kbnode): New.
* keyedit.c (keyedit_menu): Call collapse_uids.
* g10.c: new option --logger-fd.
* import.c: s/log_*_f/log_*/
Thu May 20 14:04:08 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* misc.c (pull_in_libs): do the volatile only for gcc
* sig-check (signature_check): Emit SIG_iD only for classes 0 and 1.
* armor.c (armor_filter): Add detection of PGP2 created clearsigs.
(fake_packet): A tab is not a WS for pgp2 - handle this.
* textfilter.c (len_without_trailing_chars): New.
(copy_clearsig_text): Add pgp2mode arg.
* sign.c (clearsign_file): pass old_style to the above fnc.
Wed May 19 16:04:30 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* g10.c: New option --interactive.
* mainproc.c (proc_plaintext): Add workaround for pgp2 bug
(do_check_sig): Ditto.
(proc_tree): Ditto.
* plaintext.c (do_hash): Ditto.
(hash_datafiles): Ditto, add an arg, changed all callers.
* mdfilter.c (md_filter): Add support for the alternate hash context.
Mon May 17 21:54:43 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* parse-packet.c (parse_encrypted): Support for PKT_ENCRYPTED_MDC.
* build-packet.c (do_encrypted_mdc): Ditto.
* cipher.c (write_header): Add mdc hashing.
(cipher_filter): write out the hash.
* mainproc.c (do_proc_packets): Add PKT_ENCRYPTED_MDC.
* encr-data.c (decrypt_data): Add mdc hashing.
(mdc_decode_filter): New.
* parse-packet.c (parse_sig_subpkt): Fixed stupid bug for subpkt
length calculation
(parse_signature): Fixed even more stupid bug.
Sat May 8 19:28:08 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* build-packet.c (do_signature): Removed MDC hack.
* encode.c (encode_crypt_mdc): Removed.
* mainproc.c (do_check_sig): Removed MDC hack.
(check_sig_and_print): Ditto.
* parse-packet.c (parse_signature): Ditto.
* sig-check.c (mdc_kludge_check): Ditto.
* free-packte.c (copy_signature, free_seckey_enc): Ditto.
* parse-packet.c (parse_signature,parse_key): Store data of
unknown algorithms with mpi_set_opaque inseatd of the old
faked data stuff.
(read_rest): Removed.
(read_rest2): Renamed to read_rest
* build-packet.c (write_fake_data): Use mpi_get_opaque.
* free-packet.c (cp_fake_data): Removed and cahnged all callers
to use mpi_copy.
(free_pubkey_enc,free_seckey_enc,release_public_key_parts,
release_secret_key_parts): Use mpi_free for opaque data.
Thu May 6 14:18:17 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* trustdb.c (check_trust): Check for revoked subkeys.
* pkclist.c (do_we_trust): Handled revoked subkeys.
(do_we_trust_pre): Ditto.
(check_signatures_trust): Ditto.
* build-packet.c (hash_public_key): Fix for ancient g10 keys.
* mainproc.c (do_proc_packets): Return EOF if no data has been read.
* g10.c (main): Catch errors for default operation.
Thu Apr 29 12:29:22 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* sign.c (sign_file): Fixed hashing in case of no subpackets.
(clearsign_file): Ditto.
(make_keysig_packet): Ditto.
Wed Apr 28 13:03:03 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* keyedit.c (keyedit_menu): Add new command revkey.
* (menu_revkey): New.
Mon Apr 26 17:48:15 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* parse-packet.c (parse_signature): Add the MDC hack.
* build-packet.c (do_signature): Ditto.
* free-packet.c (free_seckey_enc,copy_signature,cmp_signatures): Ditto.
* mainproc.c (do_check_sig): Ditto.
* sig-check.c (mdc_kludge_check): New.
* encode.c (encrypt_mdc_file): New.
* keyedit.c (check_all_keysigs): List revocations.
* (menu_revsig): New.
* sign (make_keysig_packet): Support for class 0x30.
Sun Apr 18 20:48:15 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* pkclist.c (select_algo_from_prefs): Fixed the case that one key
has no preferences (Remi Guyomarch).
keylist.c (list_keyblock): ulti_hack to propagate trust to all uids.
Sun Apr 18 10:11:28 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* seckey-cert.c (do_check): Use real IV instead of a 0 one, so that
it works even if the length of the IV doesn't match the blocksize.
Removed the save_iv stuff.
(protect_secret_key): Likewise. Create the IV here.
* packet.h (PKT_secret_key): Increased size of IV field and add a
ivlen field.
* parse-packet.c (parse_key): Use the len protect.ivlen.
* build-packet.c (do_secret_key). Ditto.
* getkey.c (key_byname): Close keyblocks.
* Makefile.am (gpgm): Removed this
* g10.c: Merged gpg and gpgm
* import.c (import): Utilize option quiet.
* tdbio.c (tdbio_set_dbname): Ditto.
* ringedit.c (add_keyblock_resource,keyring_copy): Ditto.
* keyedit.c (sign_uids): Add some batch support.
* g10.c (main): add call to tty_batchmode.
Fri Apr 9 12:26:25 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* status.c (write_status_text): Some more status codes.
* passphrase_to_dek (passphrase_to_dek): add a status code.
* seckey_cert.c (check_secret_key): Likewise.
* encr-data.c (decrypt_data): Reverse the last changes
* cipher.c (write_header): Ditto.
* parse-packet.c (parse_key): Dropped kludge for ancient blowfish mode.
Thu Apr 8 09:35:53 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* mainproc.c (proc_encrypted): Add a new status output
* passphrase.c (passphrase_to_dek): Ditto.
* status.h status.c: Add new status tokens.
Wed Apr 7 20:51:39 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* encr-data.c (decrypt_data): Fixes for 128 bit blocksize
* cipher.c (write_header): Ditto.
* seckey-cert.c (do_check): Ditto.
(protect_secret_key). Ditto.
* misc.c (print_cipher_algo_note): Twofish is now a standard algo.
* keygen.c (do_create): Fixed spelling (Gaël Quéri)
(ask_keysize): Only allow keysizes up to 4096
* ringedit.c (add_keyblock_resource): chmod newly created secrings.
* import.c (delete_inv_parts): Fixed accidently deleted subkeys.
Tue Apr 6 19:58:12 CEST 1999 Werner Koch <wk@isil.d.shuttle.de>
* armor.c: Removed duped include (John Bley)
* mainproc.c: Ditto.
* build-packet.c (hash_public_key): Fixed hashing of the header.
* import.c (delete_inv_parts): Allow import of own non-exportable sigs.
Sat Mar 20 13:59:47 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* armor.c (fake_packet): Fix for not not-dash-escaped
Sat Mar 20 11:44:21 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* g10.c (main): Added command --recv-keys
* hkp.c (hkp_import): New.
Wed Mar 17 13:09:03 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* trustdb.c (check_trust): add new arg add_fnc and changed all callers.
(do_check): Ditto.
(verify_key): Ditto.
(propagate_validity): Use the new add_fnc arg.
(print_user_id): Add the FILE arg.
(propagate_ownertrust): New.
* pkclist.c (add_ownertrust_cb): New and changed the add_ownertrust
logic.
* getkey.c (get_keyblock_bylid): New.
* trustdb.c (print_uid_from_keyblock): New.
(dump_tn_tree_with_colons): New.
(list_trust_path): Add colon print mode.
* trustdb.c (insert_trust_record): Always use the primary key.
* encode.c (encode_simple): Added text_mode filter (Rémi Guyomarch)
(encode_crypt): Ditto.
* mainproc.c (proc_pubkey_enc): Added status ENC_TO.
* armor.c (armor_filter): Added status NODATA.
* passphrase.c (passphrase_to_dek): Always print NEED_PASSPHRASE
* seckey_cert.c (check_secret_key): Added BAD_PASS status.
* g10.c (main): Set g10_opt_homedir.
Sun Mar 14 19:34:36 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* keygen.c (do_create): Changed wording of the note (Hugh Daniel)
Thu Mar 11 16:39:46 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* tdbdump.c: New
* trustdb.c (walk_sigrecs,do_list_sigs,list_sigs,
list_records,list_trustdb,export_ownertrust,import_ownertrust): Moved
to tdbdump.c
(init_trustdb): renamed to setup_trustdb. Changed all callers.
(do_init_trustdb): renamed to init_trustdb().
* trustdb.c (die_invalid_db): replaced by tdbio_invalid.
* tdbio.c (tdbio_invalid): New.
* import.c (delete_inv_parts): Skip non exportable signatures.
* keyedit.c (sign_uid_mk_attrib): New.
(sign_uids): Add the local argument.
(keyedit_menu): New "lsign" command.
* trustdb.c (register_trusted_key): Removed this and all related stuff.
* g10.c (oTrustedKey): Removed option.
* tdbio.h (dir.valcheck): New trustdb field.
* tdbio.c: Add support for this field
(tdbio_read_modify_stamp): New.
(tdbio_write_modify_stamp): New.
* trustdb.c (do_check): Check against this field. Removed cache update.
(verify_key): Add cache update.
(upd_uid_record): Some functional changes.
(upd_cert_record): Ditto
Wed Mar 10 11:26:18 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* keylist.c (list_keyblock): Fixed segv in uid. Print 'u' as
validity of sks.
Mon Mar 8 20:47:17 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* getkey.c (classify_user_id): Add new mode 12 (#<lid>).
* seckey-cert.c (check_secret_key): replaced error by info.
* trustdb.c (query_trust_info): Add another arg, changed all callers.
(check_trust): Ditto.
(do_check): Ditto.
(verify_key): Handle namehash.
* keylist.c (list_keyblock): print trust info for user ids.
* sig-check.c (signature_check): Add sig-created to status output.
Tue Mar 2 16:44:57 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* textfilter.c (copy_clearsig_text): New.
(clearsign): Removed.
* sign.c (clearsign_file): does not use textfiler anymore.
* keygen.c (ask_user_id): print a note about the used charset.
Tue Mar 2 10:38:42 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* sig-check.c (signature_check): sig-id now works for all algos.
* armor.c (armor_filter): Fixed armor bypassing.
Sun Feb 28 19:11:00 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* keygen.c (ask_user_id): Don't change the case of email addresses.
(has_invalid_email_chars): Adjusted.
* keylist.c (list_one): Really list serect keys (Remi Guyomarch)
* keyedit.c (menu_select_uid): Add some braces to make egcs happy.
(menu_select_key): Ditto.
* mainproc.c (do_proc_packets): List sym-enc packets (Remi Guyomarch)
Fri Feb 26 17:55:41 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* pkclist.c (build_pk_list): Return error if there are no recipients.
* sig-check.c (signature_check): New signature id feature.
* armor.c (make_radic64_string): New.
* mainproc.c (proc_pubkey_enc): early check for seckey availability.
* pkclist.c (do_we_trust_pre): print user id before asking.
* ringedit.c (add_keyblock_resource,get_keyblock_handle): Cleaner
handling of default resource.
Thu Feb 25 18:47:39 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* pkclist.c (algo_available): New.
(select_algo_from_prefs): Check whether algo is available.
* ringedit.c (keyring_copy): Take care of opt.dry_run.
(do_gdbm_store): Ditto.
* openfile.c (open_outfile). Ditto.
(copy_options_file): Ditto.
* trustdb.c (update_trustdb): Ditto.
(clear_trust_checked_flag): Ditto.
(update_trust_record): Ditto.
(insert_trust_record): Ditto.
Wed Feb 24 11:07:27 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* keylist.c (secret_key_list): Now really list the secret key.
* trustdb.c (do_init_trustdb): New. Init is now deferred.
Mon Feb 22 20:04:00 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* getkey.c (lookup_sk): Return G10ERR_NO_SECKEY and not x_PUBKEY.
Fri Feb 19 15:49:15 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* pkclist.c (select_algo_from_prefs): retrieve LID if not there.
* armor.c (fake_packet): Replaced ugly lineending handling.
* g10.c (oNoEncryptTo): New.
* pkclist.c (build_pk_list): Implemented this option.
* g10.c (main): Greeting is now printed to stderr and not to tty.
Use add_to_strlist() instead of direct coding.
* import.c (import): Use iobuf_push_filter2.
* mainproc.c (check_sig_and_print): Print all user ids
for good signatures.
* getkey.c (get_pubkeyblock): New.
* import.c (chk_self_sigs): Fixed SEGV for unbounded class 0x18 keys.
(delete_inv_parts): Delete special marked packets.
Tue Feb 16 14:10:02 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* g10.c (main): New option --encrypt-to
* pkclist.c (build_pk_list): Implemented encrypt-to.
* parse-packet.c (parse_user_id): Removed the hack to work with
utf-8 strings.
* g10.c (main): Install lockfile cleanup handler.
* tdbio.c (cleanup): Removed: this is now handled by dotlock.
Sat Feb 13 14:13:04 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* tdbio.c (tdbio_set_dbname): Init lockhandle for a new trustdb
Wed Feb 10 17:15:39 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* g10.c (main): check for development version now in configure
* tdbio.c (tdbio_write_record): Add uid.validity
(tdbio_read_record) : Ditto.
(tdbio_dump_record) : Ditto.
* keygen.c (keygen_add_std_prefs): Replaced Blowfish by Twofish,
removed MD5 and Tiger.
* pubkey-enc.c (get_it): Suppress warning about missing Blowfish
in preferences in certain cases.
* ringedit.c (lock_rentry,unlock_rentry): New.
* getkey.c (key_byname): Pass ret_kb down to lookup_xx.
* armor.c (armor_filter): No output of of empty comment lines.
Add option --no-version to suppress the output of the version string.
* getkey.c: Release the getkey context for auto context variables.
Sun Jan 24 18:16:26 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* getkey.c: Changed the internal design to allow simultaneous
lookup of multible user ids
(get_pubkey_bynames): New.
(get_seckey_bynames): New.
(get_seckey_next): New.
(get_seckey_end): New.
* keylist.c (list_one): Use the new functions.
* keylist.c (list_keyblock): add a newline for normal listings.
* g10.c (--recipient): New option name to replace --remote-user
Wed Jan 20 18:59:49 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* textfilter.c: Mostly rewritten
* plaintext.c (handle_plaintext): Use now text_filter semantics.
Tue Jan 19 19:34:58 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* export.c (export_pubkeys_stream): New.
(do_export_stream): New.
* g10.c (aSendKeys): New command.
* hkp.c (hkp_export): New.
* compress.c (do_uncompress): Hack for algo 1 and 1.1.3
Sun Jan 17 11:04:33 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* textfilter.c (text_filter): Now uses iobuf_read_line().
(read_line): Removed.
* armor.c (trim_trailing_spaces): Removed and replaced
by trim_trailing_ws from libutil
Sat Jan 16 12:03:27 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* hkp.c (hkp_ask_import): Use only the short keyid
Sat Jan 16 09:27:30 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* import.c (import_key_stream): New
(import): New, moved most of import_keys here.
* g10.c: New option --keyserver
* mainproc.c (check_sig_and_print): Hook to import a pubkey.
* pref.c pref.h : Removed
* hkp.c hkp.h: New
Wed Jan 13 14:10:15 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* armor.c (radix64_read): Print an error if a bad armor was detected.
Wed Jan 13 12:49:36 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* armor.c (radix64_read): Now handles malformed armors produced
by some buggy MUAs.
Tue Jan 12 11:17:18 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* ringedit.c (find_keyblock_bysk): New.
* skc_list.c (is_insecure): New.
(build_sk_list): usage check for insecure keys.
* import.c (chk_self_sigs): Add handling for subkeys.
(delete_inv_parts): Skip unsigned subkeys
* sig-check.c (do_check): Print info if the signature is older
than the key.
* keygen.c (generate_subkeypair): Fail on time warp.
* sign.c (do_sign): Ditto.
Sun Jan 10 15:10:02 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* armor.c (fake_packet): Fixed not-dash-escaped bug.
Sat Jan 9 16:02:23 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* sig-check.c (do_check): Output time diff on error
* status.c (STATUS_VALIDSIG): New.
(is_status_enabled): New.
* mainproc.c (check_sig_and_print): Issue that status message.
* plaintext.c (special_md_putc): Removed
* armor.c (armor_filter): print error for truncated lines.
* free-packet.c (free_encrypted): Revomed call to set_block_mode.
(free_plaintext): Ditto.
Thu Jan 7 18:00:58 CET 1999 Werner Koch <wk@isil.d.shuttle.de>
* pkclist.c (add_ownertrust): Fixed return value.
* encr-data.c (decrypt_data): Disabled iobuf_set_limit and
iobuf_pop_filter stuff.
* compress.c (handle_compressed): Disabled iobuf_pop_filter.
* packet.h (PKT_secret_key): Add is_primary flag.
* parse-packet.c (parse_key): Set this flag.
* passphrase.c (passphrase_to_dek): Kludge to print the primary
keyid - changed the API: keyid must now hold 2 keyids.
* getkey.c (get_primary_seckey): New.
* seckey-cert.c (do_check): pass primary keyid to passphrase query
* tbdio.c (open_db): removed the atexit
(tdbio_set_dbname): and moved it to here.
* armor.c: Rewrote large parts.
Tue Dec 29 19:55:38 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* revoke.c (gen_revoke): Removed compression.
* pkclist.c (do_we_trust_pre): special check for revoked keys
* trustdb.c (update_trust_record): Fixed revoke flag.
Tue Dec 29 14:41:47 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* misc.c (disable_core_dumps): Check for EINVAL (Atari)
* getkey (merge_one_pk_and_selfsig): Fixed search of expiredate.
(merge_keys_and_selfsig): Ditto.
* free-packet.c (cmp_public_keys): cmp expire only for v3 packets
(cmp_secret_keys): Ditto.
(cmp_public_secret_key): Ditto.
Wed Dec 23 17:12:24 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* armor.c (find_header): Reset not_dashed at every header
Wed Dec 23 13:18:14 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* pkclist.c (add_ownertrust): Refresh validity values.
* trustdb.c (enum_cert_paths_print): New arg refresh.
* ringedit.c: Fixed problems fix keyrings
* parse-packet.c (dbg_parse_packet): New debug functions.
* getkey.c (getkey_disable_caches): New.
* import.c (import_keys): Disable caches.
Thu Dec 17 18:31:15 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* misc.c (trap_unaligned): Only for glibc 1
* sign.c (write_dash_escaped): Now escapes "From " lines
* g10.c: New option --escape-from-lines
* trustdb.c (sort_tsl_list): New
(list_trust_path): Now prints sorted list.
(enum_cert_paths): Likewise.
(enum_cert_paths_print): New.
(print_paths): New printing format.
* pkclist.c (add_ownertrust): New arg quit.
(edit_ownertrust): New quit selection and does not query
the recipients ownertrust anymore.
(add_ownertrust): Print the ceritficate path.
Mon Dec 14 21:18:49 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* parse-packet.c (parse_signature): Now checks for critical bit
(parse_sig_subpkt): Splitted.
(parse_one_sig_subpkt): New.
* sig-check.c (do_check): handle critical bit.
Sun Dec 13 14:10:56 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* pcklist.c (select_algo_from_prefs): Preferences should
now work (lost the != ? )
Thu Dec 10 20:15:36 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* ringedit.c (gdbm_store): Fix for inserts
* g10.c (main): New option --export-all
* export.c (export_pubkeys): New arg.
(do_export): Now may skip old keys.
* status.c: Minor patches for Sun's cc
* keygen.c (ask_algo): Disabled v3 ElGamal choice, rearranged
the numbers. Add a warning question when a sign+encrypt key
is selected.
* g10.c (do_not_use_RSA): Removed.
* misc.c (print_pubkey_algo_note): New as replacement for the
do_not_use_RSA() and chnaged all callers.
(print_cipher_algo_note): New.
(print_hash_algo_note): New.
* cipher.c (write_header): Add a call to print_cipher_algo_note.
* seckey-cert.c (protect_secret_key): Ditto
* sign.c (do_sign): Add a call to print_digest_algo_note.
* getkey.c (get_long_user_id_string): New.
* mainproc.c (check_sig_and_print): Changed the format of the
status output.
* encrypt.c (write_pubkey_enc_from_list): print used symmetric cipher.
* pkclist.c (do_we_trust): Changed a message.
Wed Dec 9 13:41:06 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* misc.c (trap_unaligned) [ALPHA]: Only if UAC_SIGBUS is defined.
* sign.c (write_dash_escaped): Add the forgotten patch by Brian Moore.
* compress.c (do_uncompress): Fixed the inflating bug.
Tue Dec 8 13:15:16 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* trustdb.c (upd_uid_record): Now uses the newest self-signature
(insert_trust_record): Now calls update with recheck set to true.
(register_trusted_key): New.
(verify_own_keys): Enhanced by list of trusted keys.
* g10.c (main): Print a warning when a devel version is used.
(main): New option --trusted-key
* import.c (merge_blocks): Fixed merging of new user ids and
added merging of subkeys.
(append_uid): Ditto.
(merge_keysig): New.
(append_key): New.
* getkey.c (merge_one_pk_and_selfsig): Get the expiration time
from the newest self-signature.
(merge_keys_and_selfsig): Ditto.
* free-packet.c (cmp_secret_key): New.
Fri Nov 27 21:37:41 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* g10.c: New option --lock-once
* tdbio.c (open_db): Add an atexit
(cleanup): New.
(tdbio_sync): Add locking.
(tdbio_end_transaction): Ditto.
(put_record_into_cache): Ditto.
* ringedit.c (keyring_copy): Ditto.
(cleanup): New.
(add_keyblock_resource): Add an atexit.
Fri Nov 27 15:30:24 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* armor.c (find_header): Another fix for clearsigs.
Fri Nov 27 12:39:29 CET 1998 Werner Koch <wk@isil.d.shuttle.de>
* status.c (display_help): Removed.
* helptext.c: New and removed the N_() from all cpr_gets.
Fri Nov 20 16:54:52 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (main): New option --not-dash-escaped
* sign.c (write_dashed_escaped): Ditto.
* armor.c (find_header): Support for NotDashEscaped header.
* getkey.c: print "disabled cache.." only if verbose is used.
Thu Nov 19 07:17:31 1998 Werner Koch <werner.koch@guug.de>
* parse-packet.c (dump_sig_subpkt): Fixed expire listing
* getkey.c (merge_keys_and_selfsig): Fixed expire calculation.
(merge_one_pk_and_selfsig): Ditto.
* keyedit.c (menu_expire). Ditto.
* keygen.c (keygen_add_key_expire): Ditto.
(ask_expire_interval): New and changed all local function to use
this instead.
(keygen_add_key_expire): Opaque should now be a public key;
changed all callers.
* parse.packet.c (parse): use skip_rest to skip packets.
* keyedit.c (keyedit_menu): New arg for cmdline cmds.
Wed Nov 18 20:33:50 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c (check_trustdb): Now rechecks all gived userids.
(collect_paths): Some fixes.
(upd_pref_records): Skips empty items, evaluate all items.
* parse-packet.c (dump_sig_subpkt): Better listing of prefs.
(skip_packet): Now knows about marker packet
* g10.c: removed cmd "--edit-sig".
* pubring.asc: Updated.
Sat Nov 14 14:01:29 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (main): Changed syntax of --list-trust-path
* trustdb.c (list_trust_path): Replaced max_depth by
opt.max_cert_depth
Fri Nov 13 07:39:58 1998 Werner Koch <werner.koch@guug.de>
* trustdb.c (collect_paths): Removed a warning message.
(enum_trust_web): Removed.
(enum_cert_paths): New.
* pkclist.c (add_ownertrust): Changed to use enum_cert_paths.
(edit_ownertrust): Now list ceritficates on request.
(show_paths): New.
Wed Nov 11 18:05:44 1998 Werner Koch <werner.koch@guug.de>
* g10.c (main): New option --max-cert-depth
* tdbio.h: add new fields to ver and dir record.
* tdbio.c: read/write/dump of these fields.
(tdbio_db_matches_options): New.
* trustdb.c: replaced MAC_CERT_DEPTH by opt.max_cert_depth.
(do_check): cache validity and changed other functions
to reset the cached value.
* keylist.c (list_one): Now lists the ownertrust.
* mainproc.c (list_node): Ditto.
Tue Nov 10 10:08:59 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (g10_exit): Now looks at the new g10_errors_seen.
* mainproc.c (check_sig_and_print): Sets g10_errors_seen.
* *.c : i18n many more strings.
* ringedit.c (locate_keyblock_by_keyid): Add HAVE_LIBGDBM
(locate_keyblock_by_fpr): Ditto.
* g10.c (main): removed unsused "int errors".
(main): Add new option --charset.
* g10.c (main): special message for the unix newbie.
Mon Nov 9 07:17:42 1998 Werner Koch <werner.koch@guug.de>
* getkey.c (finish_lookup): Kludge to prefere algo 16.
* trustdb.c (new_lid_table): Clear cached item.
* status.c (cpr_get_utf8): New.
* pkclist.c (build_pk_list): Uses this.
Sun Nov 8 17:20:39 1998 Werner Koch (wk@isil.d.shuttle.de)
* mainproc.c (check_sig_and_print): Why did I use strlen()-1
in the printf? - This truncated the TZ.
Sat Nov 7 15:57:28 1998 me,,, (wk@tobold)
* getkey.c (lookup): Changes to support a read_next.
(get_pubkey): Fixed a memory leak.
* keylist.c (list_one): Now lists all matching user IDs.
Tue Nov 3 16:19:21 1998 Werner Koch (wk@isil.d.shuttle.de)
* keygen.c (ask_user_id): Now converted to UTF-8
* g10.c (main): Kludge for pgp clearsigs and textmode.
Fri Oct 30 16:40:39 1998 me,,, (wk@tobold)
* signal.c (block_all_signals): New.
(unblock_all_signals): New
* tdbio.c (tdbio_end_transaction): Now blocks all signals.
* trustdb.c (new_lid_table): Changed the representation of the
former local_lid_info stuff.
* trustdb.c (update_trust_record): Reorganized the whole thing.
* sig-check.c (check_key_signature): Now handles class 0x28
Wed Oct 28 18:56:33 1998 me,,, (wk@tobold)
* export.c (do_export): Takes care of the exportable sig flag.
Tue Oct 27 14:53:04 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c (update_trust_record): New "fast" parameter.
Sun Oct 25 19:32:05 1998 Werner Koch (wk@isil.d.shuttle.de)
* openfile.c (copy_options_File): New.
* ringedit.c (add_keyblock_resource): Creates options file
* tdbio.c (tdbio_set_dbname): Ditto.
Sat Oct 24 14:10:53 1998 brian moore <bem@cmc.net>
* mainproc.c (proc_pubkey_enc): Don't release the DEK
(do_proc_packets): Ditto.
Fri Oct 23 06:49:38 1998 me,,, (wk@tobold)
* keyedit.c (keyedit_menu): Comments are now allowed
* trustdb.c: Rewrote large parts.
Thu Oct 22 15:56:45 1998 Michael Roth (mroth@nessie.de)
* encode.c: (encode_simple): Only the plain filename without
a given directory is stored in generated packets.
(encode_crypt): Ditto.
* sign.c: (sign_file) Ditto.
Thu Oct 22 10:53:41 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c (update_trust_record): Add new optional arg.
* import.c (import_keys): Add statistics output
* trustdb.c (update_trustdb): Ditto.
(insert_trustdb): Ditto.
* tdbio.c (tdbio_begin_transaction): New.
(tdbio_end_transaction): New.
(tdbio_cancel_transaction): New.
* g10.c (main): New option --quit.
* trustdb.c (check_hint_sig): No tests for user-id w/o sig.
This caused an assert while checking the sigs.
* trustdb.c (upd_sig_record): Splitted into several functions.
* import.c (import_keys): New arg "fast".
* g10.c (main): New command --fast-import.
Wed Oct 21 18:19:36 1998 Michael Roth <mroth@nessie.de>
* ringedit.c (add_keyblock_resource): Directory is now created.
* tdbio.c (tdbio_set_dbname): New info message.
Wed Oct 21 11:52:04 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c (update_trustdb): released keyblock in loop.
* keylist.c (list_block): New.
(list_all): Changed to use list_block.
* trustdb.c: Completed support for GDBM
* sign.c (only_old_style): Changed the way force_v3 is handled
(sign_file): Ditto.
(clearsign_file): Ditto.
* keygen.c (has_invalid_email_chars): Splitted into mailbox and
host part.
* keylist.c (list_one): Add a merge_keys_and_selfsig.
* mainproc.c (proc_tree): Ditto.
Sun Oct 18 11:49:03 1998 Werner Koch (wk@isil.d.shuttle.de)
* sign.c (only_old_style): Add option force_v3_sigs
(sign_file): Fixed a bug in sig->version
(clearsign_file): Ditto.
* parse-packet.c (dump_sig_subpkt): New
* keyedit.c (menu_expire): New.
* free-packet.c (cmp_signatures): New
Sat Oct 17 10:22:39 1998 Werner Koch (wk@isil.d.shuttle.de)
* armor.c: changed output line length from 72 to 64.
* keyedit.c (fix_keyblock): New.
Fri Oct 16 10:24:47 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c: Rewrote most.
* tdbio.c: Add cache and generalized hash tables.
* options.h (ENABLE_COMMENT_PACKETS): New but undef'ed.
* encode.c, sign.c, keygen.c: Disabled comment packets.
* export.c (do_export): Comment packets are never exported,
except for those in the secret keyring.
* g10.c (main): Removed option do-no-export-rsa; should be
be replaced by a secpial tool.
* export.c (do_export): Removed the code for the above option.
* armor.c (find_header): Support for new only_keyblocks.
* import.c (import_keys): Only looks for keyblock armors.
* packet.h: replaced valid_days by expiredate and changed all users.
* build-packet.c (do_public_key): calculates valid-days
(do_secret_key): Ditto.
* parse-packet.c (parse_key): expiredate is calucated from the
valid_period in v3 packets.
* keyid.c (do_fingerprint_md): calculates valid_dates.
* keygen.c (add_key_expire): fixed key expiration time for v4 packets.
* armor.c (find_header): A LF in the first 28 bytes
was skipped for non-armored data.
Thu Oct 8 11:35:51 1998 Werner Koch (wk@isil.d.shuttle.de)
* armor.c (is_armored): Add test on old comment packets.
* tdbio.c (tdbio_search_dir_bypk): fixed memory leak.
* getkey.c: Changed the caching algorithms.
Wed Oct 7 19:33:28 1998 Werner Koch (wk@isil.d.shuttle.de)
* kbnodes.c (unused_nodes): New.
Wed Oct 7 11:15:36 1998 Werner Koch (wk@isil.d.shuttle.de)
* keyedit.c (sign_uids): Fixed a problem with SK which could caused
a save of an unprotected key.
(menu_adduid): Ditto.
* keyedit.c (keyedit_menu): Prefs are now correctly listed for
new user ids.
* trustdb.c (update_trust_record): New.
(insert_trust_record): Now makes use of update_trust_record.
Tue Oct 6 16:18:03 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c (read_record): replaces most of the tdbio_read_records.
(write_record): Ditto.
Sat Oct 3 11:01:21 1998 Werner Koch (wk@isil.d.shuttle.de)
* keygen.c (ask_alogo): enable ElGamal enc-only only for addmode.
Wed Sep 30 10:15:33 1998 Werner Koch (wk@isil.d.shuttle.de)
* import.c (import_one): Fixed update of wrong keyblock.
Tue Sep 29 08:32:08 1998 me,,, (wk@tobold)
* mainproc.c (proc_plaintext): Display note for special filename.
* plaintext.c (handle_plaintext): Suppress output of special file.
Mon Sep 28 12:57:12 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (verify_own_keys): Add warning if a key is not protected.
* passphrase (hash_passphrase): Fixed iterated+salted mode and
setup for keysizes > hashsize.
* g10.c (main): New options: --s2k-{cipher,digest,mode}.
Fri Sep 25 09:34:23 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c: Chnaged some help texts.
Tue Sep 22 19:34:39 1998 Werner Koch (wk@isil.d.shuttle.de)
* passphrase.c (read_passphrase_from_fd): fixed bug for long
passphrases.
Mon Sep 21 11:28:05 1998 Werner Koch (wk@(none))
* getkey.c (lookup): Add code to use the sub key if the primary one
does not match the usage.
* armor.c (armor_filter): New error message: no valid data found.
(radix64_read): Changes to support multiple messages.
(i18n.h): New.
* mainproc.c (add_onepass_sig): bug fix.
Mon Sep 21 08:03:16 1998 Werner Koch (wk@isil.d.shuttle.de)
* pkclist.c (do_we_trust): Add keyid to most messages.
* passphrase.c (read_passphrase_from_fd): New.
(have_static_passphrase): New
(get_passphrase_fd): Removed.
(set_passphrase_fd): Removed.
* g10.c (main): passphrase is now read here.
* keyedit.c (keyedit_menu): "help" texts should now translate fine.
Mon Sep 21 06:40:02 1998 Werner Koch (wk@isil.d.shuttle.de)
* encode.c (encode_simple): Now disables compression
when --rfc1991 is used.
(encode_crypt): Ditto.
Fri Sep 18 16:50:32 1998 Werner Koch (wk@isil.d.shuttle.de)
* getkey.c (merge_key_and_selfsig): New.
Fri Sep 18 10:20:11 1998 Werner Koch (wk@isil.d.shuttle.de)
* pkclist.c (select_algo_from_prefs): Removed 3DES kludge.
* seskey.c (make_session_key): Fixed SERIOUS bug introduced
by adding the weak key detection code.
* sign.c (sign_file): Changed aremor header in certain cases.
Tue Sep 15 17:52:55 1998 Werner Koch (wk@isil.d.shuttle.de)
* mainproc.c (check_sig_and_print): Replaced ascime by asctimestamp.
Mon Sep 14 11:40:52 1998 Werner Koch (wk@isil.d.shuttle.de)
* seskey.c (make_session_key): Now detects weak keys.
* trustdb (clear_trust_checked_flag): New.
* plaintext.c (handle_plaintext): Does no anymore suppress CR from
cleartext signed messages.
Sun Sep 13 12:54:29 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c (insert_trust_record): Fixed a stupid bug in the free
liunked list loops.
Sat Sep 12 15:49:16 1998 Werner Koch (wk@isil.d.shuttle.de)
* status.c (remove_shmid): New.
(init_shm_comprocess): Now sets permission to the real uid.
Wed Sep 9 11:15:03 1998 Werner Koch (wk@isil.d.shuttle.de)
* packet.h (PKT_pubkey_enc): New flah throw_keyid, and add logic to
implement it.
* g10.c (main): New Option --throw-keyid
* getkey.c (enum_secret_keys): Add new ar and changed all callers.
Tue Sep 8 20:04:09 1998 Werner Koch (wk@isil.d.shuttle.de)
* delkey.c (delete_key): Moved from keyedit.c.
Mon Sep 7 16:37:52 1998 Werner Koch (wk@isil.d.shuttle.de)
* build-packet.c (calc_length_header): New arg new_ctb to correctly
calculate the length of new style packets.
* armor.c (is_armored): Checks for symkey_enc packets.
* pkclist.c (select_algo_from_prefs): 3DEs substitute is now CAST5.
Tue Aug 11 17:54:50 1998 Werner Koch (wk@isil.d.shuttle.de)
* build-packet.c (do_secret_key): Fixed handling of old keys.
* getkey.c (compare_name): Fixed exact and email matching
* openfile.c (open_outfile): Changed arguments and all callers.
Tue Aug 11 09:14:35 1998 Werner Koch (wk@isil.d.shuttle.de)
* encode.c (encode_simple): Applied option set-filename and comment.
(encode_crypt): Ditto.
* sign.c (sign_file): Ditto.
* armor.c (armor_filter): Applied option comment.
* encode.c (encode_crypt): Moved init_packet to the begin.
(encode_simple): add an init_packet().
* comment (write_comment): Now enforces a hash sign as the 1st byte.
* import.c (import_one): Add explanation for "no user ids".
* compress.c (do_uncompress): Applied Brian Warner's patch to support
zlib 1.1.3 etc.
* trustdb.c (check_trust): Fixed a problem after inserting new keys.
* getkey (lookup): do not return the primary key if usage is given
(lookup_sk): Ditto and take usage into account.
* status.c (cpr_get_answer_is_yes): add display_help.
Mon Aug 10 10:11:28 1998 Werner Koch (wk@isil.d.shuttle.de)
* getkey.c (lookup_sk): Now always returns the primary if arg
primary is true.
(lookup): Likewise.
(get_pubkey_byname): Now returns the primary key
(get_seckey_byname): Ditto.
Mon Aug 10 08:34:03 1998 Werner Koch (wk@isil.d.shuttle.de)
* keyid.c (pubkey_letter): ELG_E is now a small g.
Sat Aug 8 17:26:12 1998 Werner Koch (wk@isil.d.shuttle.de)
* openfile (overwrite_filep): Changed semantics and all callers.
Sat Aug 8 12:17:07 1998 Werner Koch (wk@isil.d.shuttle.de)
* status.c (display_help): New.
Thu Aug 6 16:30:41 1998 Werner Koch,mobil,,, (wk@tobold)
* seskey.c (encode_session_key): Now uses get_random_bits().
Thu Aug 6 07:34:56 1998 Werner Koch,mobil,,, (wk@tobold)
* ringedit.c (keyring_copy): No more backupfiles for
secret keyrings and add additional warning in case of
a failed secret keyring operation.
Wed Aug 5 11:54:37 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (check_opts): Moved to main. Changed def_cipher_algo
semantics and chnaged all users.
* pubkey-enc.c (get_sssion_key): New informational output
about preferences.
* parse-packet.c (parse_symkeyenc): Fixed salted+iterated S2K
(parse_key): Ditto.
* build-packet.c (do_secret_key): Ditto.
(do_symkey_enc): Ditto.
Tue Aug 4 08:59:10 1998 Werner Koch (wk@isil.d.shuttle.de)
* getkey.c (enum_secret_keys): Now returns only primary keys.
* getkey (lookup): Now sets the new namehash field.
* parse-packet.c (parse_sig_subpkt2): New.
* sign.c (sign_file): one-pass sigs are now emiited reverse.
Preference data is considered when selecting the compress algo.
Wed Jul 29 12:53:03 1998 Werner Koch (wk@isil.d.shuttle.de)
* free-packet.c (copy_signature): New.
* keygen.c (generate_subkeypair): rewritten
* g10.c (aKeyadd): Removed option --add-key
Mon Jul 27 10:37:28 1998 Werner Koch (wk@isil.d.shuttle.de)
* seckey-cert.c (do_check): Additional check on cipher blocksize.
(protect_secret_key): Ditto.
* encr-data.c: Support for other blocksizes.
* cipher.c (write_header): Ditto.
Fri Jul 24 16:47:59 1998 Werner Koch (wk@isil.d.shuttle.de)
* kbnode.c (insert_kbnode): Changed semantics and all callers.
* keyedit.c : More or less a complete rewrite
Wed Jul 22 17:10:04 1998 Werner Koch (wk@isil.d.shuttle.de)
* build-packet.c (write_sign_packet_header): New.
Tue Jul 21 14:37:09 1998 Werner Koch (wk@isil.d.shuttle.de)
* import.c (import_one): Now creates a trustdb record.
* g10.c (main): New command --check-trustdb
Mon Jul 20 11:15:07 1998 Werner Koch (wk@isil.d.shuttle.de)
* genkey.c (generate_keypair): Default key is now DSA with
encryption only ElGamal subkey.
Thu Jul 16 10:58:33 1998 Werner Koch (wk@isil.d.shuttle.de)
* keyid.c (keyid_from_fingerprint): New.
* getkey.c (get_pubkey_byfprint): New.
Tue Jul 14 18:09:51 1998 Werner Koch (wk@isil.d.shuttle.de)
* keyid.c (fingerprint_from_pk): Add argument and changed all callers.
(fingerprint_from_sk): Ditto.
Tue Jul 14 10:10:03 1998 Werner Koch (wk@isil.d.shuttle.de)
* plaintext.c (handle_plaintext): Now returns create error if
the file could not be created or the user responded not to overwrite
the file.
* mainproc.c (proc_plaintext): Tries again if the file could not
be created to check the signature without output.
* misc.c (disable_core_dumps): New.
* g10.c (main): disable coredumps for gpg
* g10.c (MAINTAINER_OPTIONS): New to disable some options
Mon Jul 13 16:47:54 1998 Werner Koch (wk@isil.d.shuttle.de)
* plaintext.c (hash_datafiles): New arg for better support of
detached sigs. Changed all callers.
* mainproc.c (proc_signature_packets): Ditto.
* g10.c (main): New option "compress-sigs"
* sig.c (sign_file): detached signatures are not anymore compressed
unless the option --compress-sigs is used.
Thu Jul 9 19:54:54 1998 Werner Koch (wk@isil.d.shuttle.de)
* armor.c: Fixes to allow zero length cleartext signatures
Thu Jul 9 14:52:47 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (build_list): Now drops setuid.
(main): Changed the way keyrings and algorithms are registered .
Wed Jul 8 14:17:30 1998 Werner Koch (wk@isil.d.shuttle.de)
* packet.h (PKT_public_key): Add field keyid.
* parse-packet.c (parse_key): Reset the above field.
* keyid.c (keyid_from_pk): Use above field as cache.
* tdbio.c, tdbio.h: New
* trustdb.c: Moved some functions to tdbio.c.
(print_keyid): New.
* pkclist.c (check_signatures_trust): New.
Wed Jul 8 10:45:28 1998 Werner Koch (wk@isil.d.shuttle.de)
* plaintext.c (special_md_putc): New.
(handle_plaintext): add clearsig argument
* mainproc.c (proc_plaintext): detection of clearsig
* sign.c (write_dased_escaped): Changed clearsig format
Tue Jul 7 18:56:19 1998 Werner Koch (wk@isil.d.shuttle.de)
* armor.c (find_header): Now makes sure that there is only one
empty line for clearsigs, as this is what OP now says.
Mon Jul 6 13:09:07 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (main): New option default-secret-key
* getkey.c (get_seckey_byname): support for this option.
Mon Jul 6 09:03:49 1998 Werner Koch (wk@isil.d.shuttle.de)
* getkey.c (add_keyring): Keyrings are now added to end of the
list of keyrings. The first added keyringwill be created.
(add_secret_keyring): Likewise.
* ringedit.c (add_keyblock_resource): Files are created here.
* g10.c (aNOP): Removed
* getkey.c (lookup): Add checking of usage for name lookups
* packet.h (pubkey_usage): Add a field which may be used to store
usage capabilities.
* pkclist.c (build_pk_list): getkey now called with usage arg.
* skclist.c (build_sk_list): Ditto.
* sign.c (clearsign_file): Fixed "Hash:" headers
Sat Jul 4 13:33:31 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c (list_ownertrust): New.
* g10.c (aListOwnerTrust): New.
* g10.c (def_pubkey_algo): Removed.
* trustdb.c (verify_private_data): Removed and also the call to it.
(sign_private_data): Removed.
Fri Jul 3 13:26:10 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (aEditKey): was aEditSig. Changed usage msg.
* keyedit.c: Done some i18n stuff.
* g10.c (do_not_use_RSA): New.
* sign.c (do_sign): Add call to above function.
* encode.c (write_pubkey_enc_from_list): Ditto.
Thu Jul 2 21:01:25 1998 Werner Koch (wk@isil.d.shuttle.de)
* parse-packet.c: Now is able sto store data of unknown
algorithms.
* free-packet.c: Support for this.
* build-packet.c: Can write data of packet with unknown algos.
Thu Jul 2 11:46:36 1998 Werner Koch (wk@isil.d.shuttle.de)
* parse-packet.c (parse): fixed 4 byte length header
Wed Jul 1 12:36:55 1998 Werner Koch (wk@isil.d.shuttle.de)
* packet.h (new_ctb): New field for some packets
* build-packet.c (build_packet): Support for new_ctb
* parse-packet.c (parse): Ditto.
Mon Jun 29 12:54:45 1998 Werner Koch (wk@isil.d.shuttle.de)
* packet.h: changed all "_cert" to "_key", "subcert" to "subkey".
* free-packet.c (free_packet): Removed memory leak for subkeys.
Sun Jun 28 18:32:27 1998 Werner Koch (wk@isil.d.shuttle.de)
* import.c (import_keys): Renamed from import_pubkeys.
(import_secret_one): New.
* g10.c (aExportSecret): New.
* export.c (export_seckeys): New.
* parse-packet.c (parse_certificate): Cleaned up.
(parse_packet): Trust packets are now considered as unknown.
(parse_pubkey_warning): New.
Fri Jun 26 10:37:35 1998 Werner Koch (wk@isil.d.shuttle.de)
* keygen.c (has_invalid_email_chars): New.
Wed Jun 24 16:40:22 1998 Werner Koch (wk@isil.d.shuttle.de)
* armor.c (armor_filter): Now creates valid onepass_sig packets
with all detected hash algorithms.
* mainproc.c (proc_plaintext): Now uses the hash algos as specified
in the onepass_sig packets (if there are any)
Mon Jun 22 11:54:08 1998 Werner Koch (wk@isil.d.shuttle.de)
* plaintext.c (handle_plaintext): add arg to disable outout
* mainproc.c (proc_plaintext): disable output when in sigs_only mode.
Thu Jun 18 13:17:27 1998 Werner Koch (wk@isil.d.shuttle.de)
* keygen.c: Removed all rsa packet stuff, chnaged defaults
for key generation.
Sun Jun 14 21:28:31 1998 Werner Koch (wk@isil.d.shuttle.de)
* misc.c (checksum_u16): Fixed a stupid bug which caused a
wrong checksum calculation for the secret key protection and
add a backward compatibility option.
* g10.c (main): Add option --emulate-checksum-bug.
Thu Jun 11 13:26:44 1998 Werner Koch (wk@isil.d.shuttle.de)
* packet.h: Major changes to the structure of public key material
which is now stored in an array and not anaymore in a union of
algorithm specific structures. These is needed to make the system
more extendable and makes a lot of stuff much simpler. Changed
all over the system.
* dsa.c, rsa.c, elg.c: Removed.
Wed Jun 10 07:22:02 1998 Werner Koch,mobil,,, (wk@tobold)
* g10.c ("load-extension"): New option.
Mon Jun 8 22:23:37 1998 Werner Koch (wk@isil.d.shuttle.de)
* seckey-cert.c (do_check): Removed cipher constants
(protect_secret_key): Ditto.
Fri May 29 10:00:28 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c (query_trust_info): New.
* keylist.c (list_one): Add output of trust info
* mainproc (list_node): ditto.
* g10.c (main): full trustdb init if -with-colons and any of the
key list modes.
Thu May 28 10:34:42 1998 Werner Koch (wk@isil.d.shuttle.de)
* status.c (STATUS_RSA_OR_IDEA): New.
* sig-check.c (check_signature): Output special status message.
* pubkey-enc.c (get_session_key): Ditto.
* mainproc.c (check_sig_and_print): Changed format of output.
* passpharse.c (passphrase_to_dek): Likewise.
Wed May 27 13:46:48 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (aListSecretKeys): New option --list-secret-keys
* keylist.c (std_key_list): Renamed to public_key_list.
(secret_key_list): New
(list_one, list_all): Add support for secret keys.
* getkey.c (get_secret_keyring): New.
* mainproc.c (list_node): Add option --with-colons for secret keys
* sig-check.c (check_key_signature): detection of selfsigs
* mainproc.c (list_node): fixed listing.
* g10.c (aListSecretKeys): New option --always-trust
* pkclist.c (do_we_trust): Override per option added
* status.c (write_status_text): Add a prefix to every output line.
Wed May 27 07:49:21 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10 (--compress-keys): New.
* options.h (compress_keys): New.
* export.c (export_pubkeys): Only compresses with the new option.
Tue May 26 11:24:33 1998 Werner Koch (wk@isil.d.shuttle.de)
* passphrase.c (get_last_passphrase): New
(set_next_passphrase): New.
(passphrase_to_dek): add support for the above functions.
* keyedit.c (make_keysig_packet): Add sigclass 0x18,
changed all callers due to a new argument.
* keygen.c (write_keybinding): New
(generate_subkeypair): Add functionality
(ask_algo, ask_keysize, ask_valid_days): Broke out of generate_keypair
(ask_user_id, ask_passphrase): Ditto.
Thu May 21 11:26:13 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c,gpgd.c (main): Does now return an int, so that egcs does
not complain.
* armor.c (fake_packet): Removed erro message and add a noticed
that this part should be fixed.
* sign.c (sign_file): Compression now comes in front of encryption.
* encode.c (encode_simple): Ditto.
(encode_crypt): Ditto.
Tue May 19 16:18:19 1998 Werner Koch (wk@isil.d.shuttle.de)
* armor.c (fake_packet): Changed assertion to log_error
Sat May 16 16:02:06 1998 Werner Koch (wk@isil.d.shuttle.de)
* build-packet.c (build_packet): Add SUBKEY packets.
Fri May 15 17:57:23 1998 Werner Koch (wk@isil.d.shuttle.de)
* sign.c (hash_for): New and used in all places here.
* main.h (DEFAULT_): new macros.
* g10.c (opt.def_digest_algo): Now set to 0
* compress.c (init_compress): Add support for algo 1
* options.h (def_compress_algo): New
* g10.c (main): New option --compress-algo
Fri May 15 13:23:59 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (print_mds): New feature to print only one hash,
chnaged formatting.
Thu May 14 15:36:24 1998 Werner Koch (wk@isil.d.shuttle.de)
* misc.c (trap_unaligned) [__alpha__]: New
* g10.c (trap_unaligned): Add call to this to track down SIGBUS
on Alphas (to avoid the slow emulation code).
Wed May 13 11:48:27 1998 Werner Koch (wk@isil.d.shuttle.de)
* build-packet.c (do_signature): Support for v4 pakets.
* keyedit.c (make_keysig_packet): Ditto.
* build-packet.c (build_sig_subpkt_from_sig): New.
(build_sig_subpkt): New.
* elg.c (g10_elg_sign): removed keyid_from_skc.
* dsa.c (g10_dsa_sign): Ditto.
* rsa.c (g10_rsa_sign): Ditto.
* keyedit.c (make_keysig_packet): Add call to keyid_from_skc
* sign.c (clearsign_file): Support for v4 signatures.
(sign_file): Ditto.
Wed May 6 09:31:24 1998 Werner Koch (wk@isil.d.shuttle.de)
* parse-packet.c (do_parse): add support for 5 byte length leader.
(parse_subpkt): Ditto.
* build-packet.c (write_new_header): Ditto.
* packet.h (SIGSUBPKT_): New constants.
* parse-packet.c (parse_sig_subpkt): Changed name, made global,
and arg to return packet length, chnaged all callers
Tue May 5 22:11:59 1998 Werner Koch (wk@isil.d.shuttle.de)
* keygen.c (gen_dsa): New.
* build_packet.c (do_secret_cert): Support for DSA
Mon May 4 19:01:25 1998 Werner Koch (wk@isil.d.shuttle.de)
* compress.c: doubled buffer sizes
* parse-packet.c (do_plaintext): now uses iobuf_read/write.
Mon May 4 09:35:53 1998 Werner Koch (wk@isil.d.shuttle.de)
* seskey.c (encode_md_value): Add optional argument hash_algo,
changed all callers.
* passphrase.c (make_dek_from_passphrase): Removed
* (get_passhrase_hash): Changed name to passphrase_to_dek, add arg,
changed all callers.
* all: Introduced the new ELG identifier and added support for the
encryption only one (which is okay to use by GNUPG for signatures).
Sun May 3 17:50:26 1998 Werner Koch (wk@isil.d.shuttle.de)
* packet.h (PKT_OLD_COMMENT): New name for type 16.
* parse-packet.c (parse_comment): Now uses type 61
Fri May 1 12:44:39 1998 Werner Koch,mobil,,, (wk@tobold)
* packet.h (count): Chnaged s2k count from byte to u32.
* seckey-cert.c (do_check): Changed s2k algo 3 to 4, changed
reading of count.
* build-packet.c (do_secret_cert): ditto.
* parse-packet.c (parse_certificate): ditto.
* parse-packet.c (parse_symkeyenc): New.
* build-packet.c (do_symkey_enc): New.
Thu Apr 30 16:33:34 1998 Werner Koch (wk@isil.d.shuttle.de)
* sign.c (clearsign_file): Fixed "Hash: " armor line.
Tue Apr 28 14:27:42 1998 Werner Koch (wk@isil.d.shuttle.de)
* parse-packet.c (parse_subpkt): Some new types.
Mon Apr 27 12:53:59 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (main): Add option --skip-verify.
* mainproc.c (check_sig_and_print): Ditto.
* g10.c (print_mds): Add output for Tiger.
* sign.c (sign_file): Now uses partial length headers if used
in canonical textmode (kludge to fix a bug).
* parse-packet.c (parse_certificate): Changed BLOWFISH id.
* pubkey-enc.c (get_session_key): Ditto.
* seskey.c (make_session_key): Ditto.
* seckey-cert.c (protect_secret_key,do_check): Add BLOWFISH160.
Fri Apr 24 17:38:48 1998 Werner Koch,mobil,,, (wk@tobold)
* sig-check.c (check_key_signature): Add sig-class 0x14..0x17
* keyedit.c (sign-key): Some changes to start with support of
the above new sig-classes.
Wed Apr 22 09:01:57 1998 Werner Koch,mobil,,, (wk@tobold)
* getkey.c (compare_name): add email matching
Tue Apr 21 16:17:12 1998 Werner Koch,mobil,,, (wk@tobold)
* armor.c (armor_filter): fixed missing last LF before CSUM.
Thu Apr 9 11:35:22 1998 Werner Koch (wk@isil.d.shuttle.de)
* seckey-cert.c (do_check): New; combines all the check functions
into one.
* sign.c: removed all key management functions
* keyedit.c: New.
Thu Apr 9 09:49:36 1998 Werner Koch (wk@isil.d.shuttle.de)
* import.c (chk_self_sigs): Changed an error message.
Wed Apr 8 16:19:39 1998 Werner Koch (wk@isil.d.shuttle.de)
* packet.h: packet structs now uses structs from the pubkey,
removed all copy operations from packet to pubkey structs.
Wed Apr 8 13:40:33 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c (verify_own_certs): Fixed "public key not found".
* getkey.c (key_byname): New, combines public and secret key search.
* pkclist.c (build_pkc_list): Add new arg usage, changed all callers.
* skclist.c (build_skc_list): Likewise.
* ringedit.c (find_keyblock, keyring_search2): Removed.
Wed Apr 8 09:47:21 1998 Werner Koch (wk@isil.d.shuttle.de)
* sig-check.c (do_check): Applied small fix from Ulf Möller.
Tue Apr 7 19:28:07 1998 Werner Koch (wk@isil.d.shuttle.de)
* cipher.c, encr-data.c, seckey-cert.c: Now uses cipher_xxxx
functions instead of blowfish_xxx or cast_xxx
Tue Apr 7 11:04:02 1998 Werner Koch (wk@isil.d.shuttle.de)
* Makefile.am (g10maint.o): Changed the way it is created.
Mon Apr 6 11:17:08 1998 Werner Koch (wk@isil.d.shuttle.de)
* misc.c: New.
* keygen.c (checksum,checksum_u16,checksum_mpi): Moved to misc.c
* seckey-cert.c: Kludge for wrong ELG checksum implementation.
Sat Apr 4 20:07:01 1998 Werner Koch (wk@isil.d.shuttle.de)
* cipher.c (cipher_filter): Support for CAST5
* encr-data.c (decode_filter): Ditto.
(decrypt_data): Ditto.
* seskey.c (make_session_key): Ditto.
* seckey-cert.c (check_elg, check_dsa): Ditto,
(protect_secret_key): Ditto.
* pubkey-enc.c (get_session_key): Ditto.
* passphrase.c (hash_passphrase): Ditto.
Thu Apr 2 20:22:35 1998 Werner Koch (wk@isil.d.shuttle.de)
* gpgd.c: New
Thu Apr 2 10:38:16 1998 Werner Koch (wk@isil.d.shuttle.de)
* keygen.c (generate_keypair): Add valid_days stuff.
* trustdb.c (check_trust): Add check for valid_days.
Wed Apr 1 16:15:58 1998 Werner Koch (wk@isil.d.shuttle.de)
* keygen.c (generate_keypair): Addional question whether the
selected large keysize is really needed.
Wed Apr 1 15:56:33 1998 Werner Koch (wk@isil.d.shuttle.de)
* seckey-cert.c (protect_secret_key): merged protect_xxx to here.
Wed Apr 1 10:34:46 1998 Werner Koch (wk@isil.d.shuttle.de)
* Makefile.am (g10maint.c): Changed creation rule, so that it works
on FreeBSD (missing CFLAGS).
* parse-packet.c (parse_subkey): Removed.
Thu Mar 19 15:22:36 1998 Werner Koch (wk@isil.d.shuttle.de)
* ringedit.c (keyring_enum): Fixed problem with reading too
many packets. Add support to read secret keyrings.
* getkey.c (scan_keyring): Removed
(lookup): New to replace scan_keyring.
(scan_secret_keyring): Removed.
(lookup_skc): New.
Wed Mar 18 11:47:34 1998 Werner Koch (wk@isil.d.shuttle.de)
* ringedit.c (enum_keyblocks): New read mode 11.
* keyid.c (elg_fingerprint_md): New and changed all other functions
to call this if the packet version is 4 or above.
Tue Mar 17 20:46:16 1998 Werner Koch (wk@isil.d.shuttle.de)
* parse-packet.c (parse_certificate): Add listing support for subkeys.
Tue Mar 17 20:32:22 1998 Werner Koch (wk@isil.d.shuttle.de)
* armor.c (is_armored): Allow marker packet.
Thu Mar 12 13:36:49 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c (check_trust): Checks timestamp of pubkey.
* sig-check. (do_check): Compares timestamps.
Tue Mar 10 17:01:56 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (main): Add call to init_signals.
* signal.c: New.
Mon Mar 9 12:43:42 1998 Werner Koch (wk@isil.d.shuttle.de)
* dsa.c: New
* packet.h, free-packet.c, parse-packet.c : Add support for DSA
* sig-check.c, getkey.c, keyid.c, ringedit.c: Ditto.
* seckey-cert.c: Ditto.
* packet.h : Moved .digest_algo of signature packets to outer
structure. Changed all references
Sun Mar 8 13:06:42 1998 Werner Koch (wk@isil.d.shuttle.de)
* openfile.c : Support for stdout filename "-".
* mainproc.c (check_sig_and_print): Enhanced status output:
* status.c (write_status_text): New.
Fri Mar 6 16:10:54 1998 Werner Koch (wk@isil.d.shuttle.de)
* kbnode.c (clone_kbnode): Fixed private_flag.
* mainproc.c (list_node): Output of string "Revoked" as user-id.
Fri Mar 6 14:26:39 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (main): Add userids to "-kv" and cleaned up this stuff.
Fri Mar 6 12:45:58 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (main): Changed semantics of the list-... commands
and added a new one. Removed option "-d"
* decrypt.c: New.
* trustdb.c (init_trustdb): Autocreate directory only if it ends
in "/.gnupg".
Thu Mar 5 12:12:11 1998 Werner Koch (wk@isil.d.shuttle.de)
* mainproc.c (do_proc_packets): New. Common part of proc_packet.
(proc_signature_packets): special version to handle signature data.
* verify.c: New.
* g10.c (aVerify): New.
* plaintext.c (hash_datafiles): New.
* compress.c (handle_compressed): Add callback arg, changed caller.
Thu Mar 5 10:20:06 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c: Is nom the common source for gpg and gpgm
* g10maint.c: Removed
* Makefile.am: Add rule to build g10maint.c
Thu Mar 5 08:43:59 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (main): Changed the way clear text sigs are faked.
Wed Mar 4 19:47:37 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10maint.c (aMuttKeyList): New
* keylist.c: New.
Wed Mar 4 17:20:33 1998 Werner Koch (wk@isil.d.shuttle.de)
* getkey.c (get_pubkey_byname): Kludge to allow 0x prefix.
Tue Mar 3 13:46:55 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10maint.c (main): New option --gen-random.
Tue Mar 3 09:50:08 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (aDeleteSecretKey): New.
(aEditSig): Add option "--edit-key" as synonym for "--edit-sig".
(aDeleteSecretKey): New.
* getkey.c (seckey_available): New.
* sign.c (delete_key): Enhanced to delete secret keys, changed all
callers.
Mon Mar 2 21:23:48 1998 Werner Koch (wk@isil.d.shuttle.de)
* pkc_list.c (build_pkc_list): Add interactive input of user ID.
Mon Mar 2 20:54:05 1998 Werner Koch (wk@isil.d.shuttle.de)
* pkclist.c (do_we_trust_pre): New.
(add_ownertrust): Add message.
* trustdb.c (enum_trust_web): Quick fix.
Mon Mar 2 13:50:53 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (main): New action aDeleteKey
* sign.c (delete_key): New.
Sun Mar 1 16:38:58 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c (do_check): No returns TRUST_UNDEFINED instead of
eof error.
Fri Feb 27 18:14:03 1998 Werner Koch (wk@isil.d.shuttle.de)
* armor.c (find_header): Removed trailing CR on headers.
Fri Feb 27 18:02:48 1998 Werner Koch (wk@isil.d.shuttle.de)
* ringedit.c (keyring_search) [MINGW32]: Open and close file here
because rename does not work on open files. Chnaged callers.
Fri Feb 27 16:43:11 1998 Werner Koch (wk@isil.d.shuttle.de)
* sig-check.c (do_check): Add an md_enable.
* mainproc.c (do_check_sig): Use md_open in case of detached sig
(proc_tree): Take detached sigs into account.
Fri Feb 27 15:22:46 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (main): Make use of GNUPGHOME envvar.
* g10main.c (main): Ditto.
Wed Feb 25 11:40:04 1998 Werner Koch (wk@isil.d.shuttle.de)
* plaintext.c (ask_for_detached_datafile): add opt.verbose to
info output.
* openfile.c (open_sigfile): Try also name ending in ".asc"
Wed Feb 25 08:41:00 1998 Werner Koch (wk@isil.d.shuttle.de)
* keygen.c (generate_keypair): Fixed memory overflow.
Tue Feb 24 15:51:55 1998 Werner Koch (wk@isil.d.shuttle.de)
* parse-packet.c (parse_certificate): Support for S2K.
* build-packet.c (do_secret_cert): Ditto.
* keygen.c (gen_elg): Ditto.
* seckey-cert.c (check_elg): Ditto
(protect_elg): Ditto.
* sign.c (chnage_passphrase): Ditto.
* passphrase.c (get_passphrase_hash): Support for a salt and
changed all callers.
(make_dek_from_passphrase): Ditto.
Tue Feb 24 12:30:56 1998 Werner Koch (wk@isil.d.shuttle.de)
* build-packet.c (hash_public_cert): Disabled debug output.
Fri Feb 20 17:22:28 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c (init_trustdb) [MINGW32]: Removed 2nd mkdir arg.
(keyring_copy) [MINGW32]: Add a remove prior to the renames.
Wed Feb 18 18:39:02 1998 Werner Koch (wk@isil.d.shuttle.de)
* Makefile.am (OMIT_DEPENDENCIES): New.
* rsa.c: Replaced log_bug by BUG.
Wed Feb 18 13:35:58 1998 Werner Koch (wk@isil.d.shuttle.de)
* mainproc.c (do_check_sig): Now uses hash_public_cert.
* parse-packet.c (parse_certificate): Removed hashing.
* packet.h (public_cert): Removed hash variable.
* free-packet.c (copy_public_cert, free_public_cert): Likewise.
* sig-check.c (check_key_signatures): Changed semantics.
Wed Feb 18 12:11:28 1998 Werner Koch (wk@isil.d.shuttle.de)
* trustdb.c (do_check): Add handling for revocation certificates.
(build_sigrecs): Ditto.
(check_sigs): Ditto.
Wed Feb 18 09:31:04 1998 Werner Koch (wk@isil.d.shuttle.de)
* armor.c (armor_filter): Add afx->hdrlines.
* revoke.c (gen_revoke): Add comment line.
* dearmor.c (enarmor_file): Ditto.
* sig-check.c (check_key_signature): Add handling for class 0x20.
* mainproc.c : Ditto.
Tue Feb 17 21:24:17 1998 Werner Koch (wk@isil.d.shuttle.de)
* armor.c : Add header lines "...ARMORED FILE .."
* dearmor.c (enarmor_file): New.
* g10maint.c (main): New option "--enarmor"
Tue Feb 17 19:03:33 1998 Werner Koch (wk@isil.d.shuttle.de)
* mainproc.c : Changed a lot, because the packets are now stored
a simple linlked list and not anymore in a complicatd tree structure.
Tue Feb 17 10:14:48 1998 Werner Koch (wk@isil.d.shuttle.de)
* free_packet.c (cmp_public_certs): New.
(cmp_user_ids): New.
* kbnode.c (clone_kbnode): New.
(release_kbnode): Add clone support.
* ringedit.c (find_keyblock_bypkc): New.
* sign.c (remove_keysigs): Self signatures are now skipped,
changed arguments and all callers.
* import.c : Add functionality.
Tue Feb 17 09:31:40 1998 Werner Koch (wk@isil.d.shuttle.de)
* options.h (homedir): New option.
* g10.c, g10maint.c, getkey.c, keygen.c, trustdb.c (opt.homedir): New.
* trustdb.c (init_trustdb): mkdir for hoem directory
(sign_private_data): Renamed "sig" to "g10.sig"
Mon Feb 16 20:02:03 1998 Werner Koch (wk@isil.d.shuttle.de)
* kbnode.c (commit_kbnode): New.
(delete_kbnode): removed unused first arg. Changed all Callers.
* ringedit.c (keyblock_resource_name): New.
(get_keyblock_handle): NULL for filename returns default resource.
Mon Feb 16 19:38:48 1998 Werner Koch (wk@isil.d.shuttle.de)
* sig-check.s (check_key_signature): Now uses the supplied
public key to check the signature and not any more the one
from the getkey.c
(do_check): New.
(check_signature): Most work moved to do_check.
Mon Feb 16 14:48:57 1998 Werner Koch (wk@isil.d.shuttle.de)
* armor.c (find_header): Fixed another bug.
Mon Feb 16 12:18:34 1998 Werner Koch (wk@isil.d.shuttle.de)
* getkey.c (scan_keyring): Add handling of compressed keyrings.
Mon Feb 16 10:44:51 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c, g10maint.c (strusage): Rewrote.
(build_list): New
Mon Feb 16 08:58:41 1998 Werner Koch (wk@isil.d.shuttle.de)
* armor.c (use_armor): New.
Sat Feb 14 14:30:57 1998 Werner Koch (wk@isil.d.shuttle.de)
* mainproc.c (proc_tree): Sigclass fix.
Sat Feb 14 14:16:33 1998 Werner Koch (wk@isil.d.shuttle.de)
* armor.c (armor_filter): Changed version and comment string.
* encode.c, sign.c, keygen.c: Changed all comment packet strings.
Sat Feb 14 12:39:24 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (aGenRevoke): New command.
* revoke.c: New.
* sign.c (make_keysig_packet): Add support for sigclass 0x20.
Fri Feb 13 20:18:14 1998 Werner Koch (wk@isil.d.shuttle.de)
* ringedit.c (enum_keyblocks, keyring_enum): New.
Fri Feb 13 19:33:40 1998 Werner Koch (wk@isil.d.shuttle.de)
* export.c: Add functionality.
* keygen.c (generate_keypair): Moved the leading comment behind the
key packet.
* kbnode.c (walk_kbnode): Fixed.
* g10.c (main): listing armored keys now work.
Fri Feb 13 16:17:43 1998 Werner Koch (wk@isil.d.shuttle.de)
* parse-packet.c (parse_publickey, parse_signature): Fixed calls
to mpi_read used for ELG b.
Fri Feb 13 15:13:23 1998 Werner Koch (wk@isil.d.shuttle.de)
* g10.c (main): changed formatting of help output.
Thu Feb 12 22:24:42 1998 Werner Koch (wk@frodo)
* pubkey-enc.c (get_session_key): rewritten
Copyright 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006,
2007, 2008, 2009 Free Software Foundation, Inc.
This file is free software; as a special exception the author gives
unlimited permission to copy and/or distribute it, with or without
modifications, as long as this notice is preserved.
This file is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY, to the extent permitted by law; without even the
implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
diff --git a/g10/tdbio.c b/g10/tdbio.c
index 606194f53..282411e5d 100644
--- a/g10/tdbio.c
+++ b/g10/tdbio.c
@@ -1,1632 +1,1633 @@
/* tdbio.c
* Copyright (C) 1998, 1999, 2000, 2001, 2002 Free Software Foundation, Inc.
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 3 of the License, or
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, see <http://www.gnu.org/licenses/>.
*/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <errno.h>
#include <assert.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <unistd.h>
#include "errors.h"
#include "iobuf.h"
#include "memory.h"
#include "util.h"
#include "options.h"
#include "main.h"
#include "i18n.h"
#include "trustdb.h"
#include "tdbio.h"
#if defined(HAVE_DOSISH_SYSTEM)
#define ftruncate chsize
#endif
#if defined(HAVE_DOSISH_SYSTEM) || defined(__CYGWIN__)
#define MY_O_BINARY O_BINARY
#else
#define MY_O_BINARY 0
#endif
/****************
* Yes, this is a very simple implementation. We should really
* use a page aligned buffer and read complete pages.
* To implement a simple trannsaction system, this is sufficient.
*/
typedef struct cache_ctrl_struct *CACHE_CTRL;
struct cache_ctrl_struct {
CACHE_CTRL next;
struct {
unsigned used:1;
unsigned dirty:1;
} flags;
ulong recno;
char data[TRUST_RECORD_LEN];
};
#define MAX_CACHE_ENTRIES_SOFT 200 /* may be increased while in a */
#define MAX_CACHE_ENTRIES_HARD 10000 /* transaction to this one */
static CACHE_CTRL cache_list;
static int cache_entries;
static int cache_is_dirty;
/* a type used to pass infomation to cmp_krec_fpr */
struct cmp_krec_fpr_struct {
int pubkey_algo;
const char *fpr;
int fprlen;
};
/* a type used to pass infomation to cmp_[s]dir */
struct cmp_xdir_struct {
int pubkey_algo;
u32 keyid[2];
};
static char *db_name;
static DOTLOCK lockhandle;
static int is_locked;
static int db_fd = -1;
static int in_transaction;
static void open_db(void);
static void migrate_from_v2 (void);
/*************************************
************* record cache **********
*************************************/
/****************
* Get the data from therecord cache and return a
* pointer into that cache. Caller should copy
* the return data. NULL is returned on a cache miss.
*/
static const char *
get_record_from_cache( ulong recno )
{
CACHE_CTRL r;
for( r = cache_list; r; r = r->next ) {
if( r->flags.used && r->recno == recno )
return r->data;
}
return NULL;
}
static int
write_cache_item( CACHE_CTRL r )
{
int n;
if( lseek( db_fd, r->recno * TRUST_RECORD_LEN, SEEK_SET ) == -1 ) {
log_error(_("trustdb rec %lu: lseek failed: %s\n"),
r->recno, strerror(errno) );
return G10ERR_WRITE_FILE;
}
n = write( db_fd, r->data, TRUST_RECORD_LEN);
if( n != TRUST_RECORD_LEN ) {
log_error(_("trustdb rec %lu: write failed (n=%d): %s\n"),
r->recno, n, strerror(errno) );
return G10ERR_WRITE_FILE;
}
r->flags.dirty = 0;
return 0;
}
/****************
* Put data into the cache. This function may flush the
* some cache entries if there is not enough space available.
*/
int
put_record_into_cache( ulong recno, const char *data )
{
CACHE_CTRL r, unused;
int dirty_count = 0;
int clean_count = 0;
/* see whether we already cached this one */
for( unused = NULL, r = cache_list; r; r = r->next ) {
if( !r->flags.used ) {
if( !unused )
unused = r;
}
else if( r->recno == recno ) {
if( !r->flags.dirty ) {
/* Hmmm: should we use a a copy and compare? */
if( memcmp(r->data, data, TRUST_RECORD_LEN ) ) {
r->flags.dirty = 1;
cache_is_dirty = 1;
}
}
memcpy( r->data, data, TRUST_RECORD_LEN );
return 0;
}
if( r->flags.used ) {
if( r->flags.dirty )
dirty_count++;
else
clean_count++;
}
}
/* not in the cache: add a new entry */
if( unused ) { /* reuse this entry */
r = unused;
r->flags.used = 1;
r->recno = recno;
memcpy( r->data, data, TRUST_RECORD_LEN );
r->flags.dirty = 1;
cache_is_dirty = 1;
cache_entries++;
return 0;
}
/* see whether we reached the limit */
if( cache_entries < MAX_CACHE_ENTRIES_SOFT ) { /* no */
r = xmalloc( sizeof *r );
r->flags.used = 1;
r->recno = recno;
memcpy( r->data, data, TRUST_RECORD_LEN );
r->flags.dirty = 1;
r->next = cache_list;
cache_list = r;
cache_is_dirty = 1;
cache_entries++;
return 0;
}
/* cache is full: discard some clean entries */
if( clean_count ) {
int n = clean_count / 3; /* discard a third of the clean entries */
if( !n )
n = 1;
for( unused = NULL, r = cache_list; r; r = r->next ) {
if( r->flags.used && !r->flags.dirty ) {
if( !unused )
unused = r;
r->flags.used = 0;
cache_entries--;
if( !--n )
break;
}
}
assert( unused );
r = unused;
r->flags.used = 1;
r->recno = recno;
memcpy( r->data, data, TRUST_RECORD_LEN );
r->flags.dirty = 1;
cache_is_dirty = 1;
cache_entries++;
return 0;
}
/* no clean entries: have to flush some dirty entries */
if( in_transaction ) {
/* but we can't do this while in a transaction
* we increase the cache size instead */
if( cache_entries < MAX_CACHE_ENTRIES_HARD ) { /* no */
if( opt.debug && !(cache_entries % 100) )
log_debug("increasing tdbio cache size\n");
r = xmalloc( sizeof *r );
r->flags.used = 1;
r->recno = recno;
memcpy( r->data, data, TRUST_RECORD_LEN );
r->flags.dirty = 1;
r->next = cache_list;
cache_list = r;
cache_is_dirty = 1;
cache_entries++;
return 0;
}
log_info(_("trustdb transaction too large\n"));
return G10ERR_RESOURCE_LIMIT;
}
if( dirty_count ) {
int n = dirty_count / 5; /* discard some dirty entries */
if( !n )
n = 1;
if( !is_locked ) {
if( make_dotlock( lockhandle, -1 ) )
log_fatal("can't acquire lock - giving up\n");
else
is_locked = 1;
}
for( unused = NULL, r = cache_list; r; r = r->next ) {
if( r->flags.used && r->flags.dirty ) {
int rc = write_cache_item( r );
if( rc )
return rc;
if( !unused )
unused = r;
r->flags.used = 0;
cache_entries--;
if( !--n )
break;
}
}
if( !opt.lock_once ) {
if( !release_dotlock( lockhandle ) )
is_locked = 0;
}
assert( unused );
r = unused;
r->flags.used = 1;
r->recno = recno;
memcpy( r->data, data, TRUST_RECORD_LEN );
r->flags.dirty = 1;
cache_is_dirty = 1;
cache_entries++;
return 0;
}
BUG();
}
int
tdbio_is_dirty()
{
return cache_is_dirty;
}
/****************
* Flush the cache. This cannot be used while in a transaction.
*/
int
tdbio_sync()
{
CACHE_CTRL r;
int did_lock = 0;
if( db_fd == -1 )
open_db();
if( in_transaction )
log_bug("tdbio: syncing while in transaction\n");
if( !cache_is_dirty )
return 0;
if( !is_locked ) {
if( make_dotlock( lockhandle, -1 ) )
log_fatal("can't acquire lock - giving up\n");
else
is_locked = 1;
did_lock = 1;
}
for( r = cache_list; r; r = r->next ) {
if( r->flags.used && r->flags.dirty ) {
int rc = write_cache_item( r );
if( rc )
return rc;
}
}
cache_is_dirty = 0;
if( did_lock && !opt.lock_once ) {
if( !release_dotlock( lockhandle ) )
is_locked = 0;
}
return 0;
}
#if 0
/* The transaction code is disabled in the 1.2.x branch, as it is not
yet used. It will be enabled in 1.3.x. */
/****************
* Simple transactions system:
* Everything between begin_transaction and end/cancel_transaction
* is not immediatly written but at the time of end_transaction.
*
*/
int
tdbio_begin_transaction()
{
int rc;
if( in_transaction )
log_bug("tdbio: nested transactions\n");
/* flush everything out */
rc = tdbio_sync();
if( rc )
return rc;
in_transaction = 1;
return 0;
}
int
tdbio_end_transaction()
{
int rc;
if( !in_transaction )
log_bug("tdbio: no active transaction\n");
if( !is_locked ) {
if( make_dotlock( lockhandle, -1 ) )
log_fatal("can't acquire lock - giving up\n");
else
is_locked = 1;
}
block_all_signals();
in_transaction = 0;
rc = tdbio_sync();
unblock_all_signals();
if( !opt.lock_once ) {
if( !release_dotlock( lockhandle ) )
is_locked = 0;
}
return rc;
}
int
tdbio_cancel_transaction()
{
CACHE_CTRL r;
if( !in_transaction )
log_bug("tdbio: no active transaction\n");
/* remove all dirty marked entries, so that the original ones
* are read back the next time */
if( cache_is_dirty ) {
for( r = cache_list; r; r = r->next ) {
if( r->flags.used && r->flags.dirty ) {
r->flags.used = 0;
cache_entries--;
}
}
cache_is_dirty = 0;
}
in_transaction = 0;
return 0;
}
#endif
/********************************************************
**************** cached I/O functions ******************
********************************************************/
static void
cleanup(void)
{
if( is_locked ) {
if( !release_dotlock(lockhandle) )
is_locked = 0;
}
}
/* Caller must sync */
int
tdbio_update_version_record (void)
{
TRUSTREC rec;
int rc;
memset( &rec, 0, sizeof rec );
rc=tdbio_read_record( 0, &rec, RECTYPE_VER);
if(rc==0)
{
rec.r.ver.created = make_timestamp();
rec.r.ver.marginals = opt.marginals_needed;
rec.r.ver.completes = opt.completes_needed;
rec.r.ver.cert_depth = opt.max_cert_depth;
rec.r.ver.trust_model = opt.trust_model;
rc=tdbio_write_record(&rec);
}
return rc;
}
static int
create_version_record (void)
{
TRUSTREC rec;
int rc;
memset( &rec, 0, sizeof rec );
rec.r.ver.version = 3;
rec.r.ver.created = make_timestamp();
rec.r.ver.marginals = opt.marginals_needed;
rec.r.ver.completes = opt.completes_needed;
rec.r.ver.cert_depth = opt.max_cert_depth;
if(opt.trust_model==TM_PGP || opt.trust_model==TM_CLASSIC)
rec.r.ver.trust_model = opt.trust_model;
else
rec.r.ver.trust_model = TM_PGP;
rec.rectype = RECTYPE_VER;
rec.recnum = 0;
rc = tdbio_write_record( &rec );
if( !rc )
tdbio_sync();
return rc;
}
int
tdbio_set_dbname( const char *new_dbname, int create )
{
char *fname;
static int initialized = 0;
if( !initialized ) {
atexit( cleanup );
initialized = 1;
}
if(new_dbname==NULL)
fname=make_filename(opt.homedir,"trustdb" EXTSEP_S "gpg", NULL);
else if (*new_dbname != DIRSEP_C )
{
if (strchr(new_dbname, DIRSEP_C) )
fname = make_filename (new_dbname, NULL);
else
fname = make_filename (opt.homedir, new_dbname, NULL);
}
else
fname = xstrdup (new_dbname);
if( access( fname, R_OK ) ) {
if( errno != ENOENT ) {
log_error( _("can't access `%s': %s\n"), fname, strerror(errno) );
xfree(fname);
return G10ERR_TRUSTDB;
}
if( create ) {
FILE *fp;
TRUSTREC rec;
int rc;
char *p = strrchr( fname, DIRSEP_C );
mode_t oldmask;
assert(p);
*p = 0;
if( access( fname, F_OK ) ) {
try_make_homedir( fname );
- log_fatal( _("%s: directory does not exist!\n"), fname );
+ if (access (fname, F_OK ))
+ log_fatal (_("%s: directory does not exist!\n"), fname);
}
*p = DIRSEP_C;
xfree(db_name);
db_name = fname;
#ifdef __riscos__
if( !lockhandle )
lockhandle = create_dotlock( db_name );
if( !lockhandle )
log_fatal( _("can't create lock for `%s'\n"), db_name );
if( make_dotlock( lockhandle, -1 ) )
log_fatal( _("can't lock `%s'\n"), db_name );
#endif /* __riscos__ */
oldmask=umask(077);
if (is_secured_filename (fname)) {
fp = NULL;
errno = EPERM;
}
else
fp =fopen( fname, "wb" );
umask(oldmask);
if( !fp )
log_fatal( _("can't create `%s': %s\n"), fname, strerror(errno) );
fclose(fp);
db_fd = open( db_name, O_RDWR | MY_O_BINARY );
if( db_fd == -1 )
log_fatal( _("can't open `%s': %s\n"), db_name, strerror(errno) );
#ifndef __riscos__
if( !lockhandle )
lockhandle = create_dotlock( db_name );
if( !lockhandle )
log_fatal( _("can't create lock for `%s'\n"), db_name );
#endif /* !__riscos__ */
rc = create_version_record ();
if( rc )
log_fatal( _("%s: failed to create version record: %s"),
fname, g10_errstr(rc));
/* and read again to check that we are okay */
if( tdbio_read_record( 0, &rec, RECTYPE_VER ) )
log_fatal( _("%s: invalid trustdb created\n"), db_name );
if( !opt.quiet )
log_info(_("%s: trustdb created\n"), db_name);
return 0;
}
}
xfree(db_name);
db_name = fname;
return 0;
}
const char *
tdbio_get_dbname()
{
return db_name;
}
static void
open_db()
{
byte buf[10];
int n;
TRUSTREC rec;
assert( db_fd == -1 );
if (!lockhandle )
lockhandle = create_dotlock( db_name );
if (!lockhandle )
log_fatal( _("can't create lock for `%s'\n"), db_name );
#ifdef __riscos__
if (make_dotlock( lockhandle, -1 ) )
log_fatal( _("can't lock `%s'\n"), db_name );
#endif /* __riscos__ */
db_fd = open (db_name, O_RDWR | MY_O_BINARY );
if (db_fd == -1 && (errno == EACCES
#ifdef EROFS
|| errno == EROFS
#endif
)
) {
db_fd = open (db_name, O_RDONLY | MY_O_BINARY );
if (db_fd != -1)
log_info (_("NOTE: trustdb not writable\n"));
}
if ( db_fd == -1 )
log_fatal( _("can't open `%s': %s\n"), db_name, strerror(errno) );
register_secured_file (db_name);
/* check whether we need to do a version migration */
do
n = read (db_fd, buf, 5);
while (n==-1 && errno == EINTR);
if (n == 5 && !memcmp (buf, "\x01gpg\x02", 5))
{
migrate_from_v2 ();
}
/* read the version record */
if (tdbio_read_record (0, &rec, RECTYPE_VER ) )
log_fatal( _("%s: invalid trustdb\n"), db_name );
}
/****************
* Make a hashtable: type 0 = trust hash
*/
static void
create_hashtable( TRUSTREC *vr, int type )
{
TRUSTREC rec;
off_t offset;
ulong recnum;
int i, n, rc;
offset = lseek( db_fd, 0, SEEK_END );
if( offset == -1 )
log_fatal("trustdb: lseek to end failed: %s\n", strerror(errno) );
recnum = offset / TRUST_RECORD_LEN;
assert(recnum); /* this is will never be the first record */
if( !type )
vr->r.ver.trusthashtbl = recnum;
/* Now write the records */
n = (256+ITEMS_PER_HTBL_RECORD-1) / ITEMS_PER_HTBL_RECORD;
for(i=0; i < n; i++, recnum++ ) {
memset( &rec, 0, sizeof rec );
rec.rectype = RECTYPE_HTBL;
rec.recnum = recnum;
rc = tdbio_write_record( &rec );
if( rc )
log_fatal( _("%s: failed to create hashtable: %s\n"),
db_name, g10_errstr(rc));
}
/* update the version record */
rc = tdbio_write_record( vr );
if( !rc )
rc = tdbio_sync();
if( rc )
log_fatal( _("%s: error updating version record: %s\n"),
db_name, g10_errstr(rc));
}
int
tdbio_db_matches_options()
{
static int yes_no = -1;
if( yes_no == -1 )
{
TRUSTREC vr;
int rc;
rc = tdbio_read_record( 0, &vr, RECTYPE_VER );
if( rc )
log_fatal( _("%s: error reading version record: %s\n"),
db_name, g10_errstr(rc) );
yes_no = vr.r.ver.marginals == opt.marginals_needed
&& vr.r.ver.completes == opt.completes_needed
&& vr.r.ver.cert_depth == opt.max_cert_depth
&& vr.r.ver.trust_model == opt.trust_model;
}
return yes_no;
}
byte
tdbio_read_model(void)
{
TRUSTREC vr;
int rc;
rc = tdbio_read_record( 0, &vr, RECTYPE_VER );
if( rc )
log_fatal( _("%s: error reading version record: %s\n"),
db_name, g10_errstr(rc) );
return vr.r.ver.trust_model;
}
/****************
* Return the nextstamp value.
*/
ulong
tdbio_read_nextcheck ()
{
TRUSTREC vr;
int rc;
rc = tdbio_read_record( 0, &vr, RECTYPE_VER );
if( rc )
log_fatal( _("%s: error reading version record: %s\n"),
db_name, g10_errstr(rc) );
return vr.r.ver.nextcheck;
}
/* Return true when the stamp was actually changed. */
int
tdbio_write_nextcheck (ulong stamp)
{
TRUSTREC vr;
int rc;
rc = tdbio_read_record( 0, &vr, RECTYPE_VER );
if( rc )
log_fatal( _("%s: error reading version record: %s\n"),
db_name, g10_errstr(rc) );
if (vr.r.ver.nextcheck == stamp)
return 0;
vr.r.ver.nextcheck = stamp;
rc = tdbio_write_record( &vr );
if( rc )
log_fatal( _("%s: error writing version record: %s\n"),
db_name, g10_errstr(rc) );
return 1;
}
/****************
* Return the record number of the trusthash tbl or create a new one.
*/
static ulong
get_trusthashrec(void)
{
static ulong trusthashtbl; /* record number of the trust hashtable */
if( !trusthashtbl ) {
TRUSTREC vr;
int rc;
rc = tdbio_read_record( 0, &vr, RECTYPE_VER );
if( rc )
log_fatal( _("%s: error reading version record: %s\n"),
db_name, g10_errstr(rc) );
if( !vr.r.ver.trusthashtbl )
create_hashtable( &vr, 0 );
trusthashtbl = vr.r.ver.trusthashtbl;
}
return trusthashtbl;
}
/****************
* Update a hashtable.
* table gives the start of the table, key and keylen is the key,
* newrecnum is the record number to insert.
*/
static int
upd_hashtable( ulong table, byte *key, int keylen, ulong newrecnum )
{
TRUSTREC lastrec, rec;
ulong hashrec, item;
int msb;
int level=0;
int rc, i;
hashrec = table;
next_level:
msb = key[level];
hashrec += msb / ITEMS_PER_HTBL_RECORD;
rc = tdbio_read_record( hashrec, &rec, RECTYPE_HTBL );
if( rc ) {
log_error("upd_hashtable: read failed: %s\n", g10_errstr(rc) );
return rc;
}
item = rec.r.htbl.item[msb % ITEMS_PER_HTBL_RECORD];
if( !item ) { /* insert a new item into the hash table */
rec.r.htbl.item[msb % ITEMS_PER_HTBL_RECORD] = newrecnum;
rc = tdbio_write_record( &rec );
if( rc ) {
log_error("upd_hashtable: write htbl failed: %s\n",
g10_errstr(rc) );
return rc;
}
}
else if( item != newrecnum ) { /* must do an update */
lastrec = rec;
rc = tdbio_read_record( item, &rec, 0 );
if( rc ) {
log_error( "upd_hashtable: read item failed: %s\n",
g10_errstr(rc) );
return rc;
}
if( rec.rectype == RECTYPE_HTBL ) {
hashrec = item;
level++;
if( level >= keylen ) {
log_error( "hashtable has invalid indirections.\n");
return G10ERR_TRUSTDB;
}
goto next_level;
}
else if( rec.rectype == RECTYPE_HLST ) { /* extend list */
/* see whether the key is already in this list */
for(;;) {
for(i=0; i < ITEMS_PER_HLST_RECORD; i++ ) {
if( rec.r.hlst.rnum[i] == newrecnum ) {
return 0; /* okay, already in the list */
}
}
if( rec.r.hlst.next ) {
rc = tdbio_read_record( rec.r.hlst.next,
&rec, RECTYPE_HLST);
if( rc ) {
log_error( "upd_hashtable: read hlst failed: %s\n",
g10_errstr(rc) );
return rc;
}
}
else
break; /* not there */
}
/* find the next free entry and put it in */
for(;;) {
for(i=0; i < ITEMS_PER_HLST_RECORD; i++ ) {
if( !rec.r.hlst.rnum[i] ) {
rec.r.hlst.rnum[i] = newrecnum;
rc = tdbio_write_record( &rec );
if( rc )
log_error( "upd_hashtable: write hlst failed: %s\n",
g10_errstr(rc) );
return rc; /* done */
}
}
if( rec.r.hlst.next ) {
rc = tdbio_read_record( rec.r.hlst.next,
&rec, RECTYPE_HLST );
if( rc ) {
log_error( "upd_hashtable: read hlst failed: %s\n",
g10_errstr(rc) );
return rc;
}
}
else { /* add a new list record */
rec.r.hlst.next = item = tdbio_new_recnum();
rc = tdbio_write_record( &rec );
if( rc ) {
log_error( "upd_hashtable: write hlst failed: %s\n",
g10_errstr(rc) );
return rc;
}
memset( &rec, 0, sizeof rec );
rec.rectype = RECTYPE_HLST;
rec.recnum = item;
rec.r.hlst.rnum[0] = newrecnum;
rc = tdbio_write_record( &rec );
if( rc )
log_error( "upd_hashtable: write ext hlst failed: %s\n",
g10_errstr(rc) );
return rc; /* done */
}
} /* end loop over hlst slots */
}
else if( rec.rectype == RECTYPE_TRUST ) { /* insert a list record */
if( rec.recnum == newrecnum ) {
return 0;
}
item = rec.recnum; /* save number of key record */
memset( &rec, 0, sizeof rec );
rec.rectype = RECTYPE_HLST;
rec.recnum = tdbio_new_recnum();
rec.r.hlst.rnum[0] = item; /* old keyrecord */
rec.r.hlst.rnum[1] = newrecnum; /* and new one */
rc = tdbio_write_record( &rec );
if( rc ) {
log_error( "upd_hashtable: write new hlst failed: %s\n",
g10_errstr(rc) );
return rc;
}
/* update the hashtable record */
lastrec.r.htbl.item[msb % ITEMS_PER_HTBL_RECORD] = rec.recnum;
rc = tdbio_write_record( &lastrec );
if( rc )
log_error( "upd_hashtable: update htbl failed: %s\n",
g10_errstr(rc) );
return rc; /* ready */
}
else {
log_error( "hashtbl %lu: %lu/%d points to an invalid record %lu\n",
table, hashrec, (msb % ITEMS_PER_HTBL_RECORD), item);
list_trustdb(NULL);
return G10ERR_TRUSTDB;
}
}
return 0;
}
/****************
* Drop an entry from a hashtable
* table gives the start of the table, key and keylen is the key,
*/
static int
drop_from_hashtable( ulong table, byte *key, int keylen, ulong recnum )
{
TRUSTREC rec;
ulong hashrec, item;
int msb;
int level=0;
int rc, i;
hashrec = table;
next_level:
msb = key[level];
hashrec += msb / ITEMS_PER_HTBL_RECORD;
rc = tdbio_read_record( hashrec, &rec, RECTYPE_HTBL );
if( rc ) {
log_error("drop_from_hashtable: read failed: %s\n",
g10_errstr(rc) );
return rc;
}
item = rec.r.htbl.item[msb % ITEMS_PER_HTBL_RECORD];
if( !item ) /* not found - forget about it */
return 0;
if( item == recnum ) { /* tables points direct to the record */
rec.r.htbl.item[msb % ITEMS_PER_HTBL_RECORD] = 0;
rc = tdbio_write_record( &rec );
if( rc )
log_error("drop_from_hashtable: write htbl failed: %s\n",
g10_errstr(rc) );
return rc;
}
rc = tdbio_read_record( item, &rec, 0 );
if( rc ) {
log_error( "drop_from_hashtable: read item failed: %s\n",
g10_errstr(rc) );
return rc;
}
if( rec.rectype == RECTYPE_HTBL ) {
hashrec = item;
level++;
if( level >= keylen ) {
log_error( "hashtable has invalid indirections.\n");
return G10ERR_TRUSTDB;
}
goto next_level;
}
if( rec.rectype == RECTYPE_HLST ) {
for(;;) {
for(i=0; i < ITEMS_PER_HLST_RECORD; i++ ) {
if( rec.r.hlst.rnum[i] == recnum ) {
rec.r.hlst.rnum[i] = 0; /* drop */
rc = tdbio_write_record( &rec );
if( rc )
log_error("drop_from_hashtable: write htbl failed: %s\n",
g10_errstr(rc) );
return rc;
}
}
if( rec.r.hlst.next ) {
rc = tdbio_read_record( rec.r.hlst.next,
&rec, RECTYPE_HLST);
if( rc ) {
log_error( "drop_from_hashtable: read hlst failed: %s\n",
g10_errstr(rc) );
return rc;
}
}
else
return 0; /* key not in table */
}
}
log_error( "hashtbl %lu: %lu/%d points to wrong record %lu\n",
table, hashrec, (msb % ITEMS_PER_HTBL_RECORD), item);
return G10ERR_TRUSTDB;
}
/****************
* Lookup a record via the hashtable tablewith key/keylen and return the
* result in rec. cmp() should return if the record is the desired one.
* Returns -1 if not found, 0 if found or another errocode
*/
static int
lookup_hashtable( ulong table, const byte *key, size_t keylen,
int (*cmpfnc)(const void*, const TRUSTREC *),
const void *cmpdata, TRUSTREC *rec )
{
int rc;
ulong hashrec, item;
int msb;
int level=0;
hashrec = table;
next_level:
msb = key[level];
hashrec += msb / ITEMS_PER_HTBL_RECORD;
rc = tdbio_read_record( hashrec, rec, RECTYPE_HTBL );
if( rc ) {
log_error("lookup_hashtable failed: %s\n", g10_errstr(rc) );
return rc;
}
item = rec->r.htbl.item[msb % ITEMS_PER_HTBL_RECORD];
if( !item )
return -1; /* not found */
rc = tdbio_read_record( item, rec, 0 );
if( rc ) {
log_error( "hashtable read failed: %s\n", g10_errstr(rc) );
return rc;
}
if( rec->rectype == RECTYPE_HTBL ) {
hashrec = item;
level++;
if( level >= keylen ) {
log_error("hashtable has invalid indirections\n");
return G10ERR_TRUSTDB;
}
goto next_level;
}
else if( rec->rectype == RECTYPE_HLST ) {
for(;;) {
int i;
for(i=0; i < ITEMS_PER_HLST_RECORD; i++ ) {
if( rec->r.hlst.rnum[i] ) {
TRUSTREC tmp;
rc = tdbio_read_record( rec->r.hlst.rnum[i], &tmp, 0 );
if( rc ) {
log_error( "lookup_hashtable: read item failed: %s\n",
g10_errstr(rc) );
return rc;
}
if( (*cmpfnc)( cmpdata, &tmp ) ) {
*rec = tmp;
return 0;
}
}
}
if( rec->r.hlst.next ) {
rc = tdbio_read_record( rec->r.hlst.next, rec, RECTYPE_HLST );
if( rc ) {
log_error( "lookup_hashtable: read hlst failed: %s\n",
g10_errstr(rc) );
return rc;
}
}
else
return -1; /* not found */
}
}
if( (*cmpfnc)( cmpdata, rec ) )
return 0; /* really found */
return -1; /* no: not found */
}
/****************
* Update the trust hashtbl or create the table if it does not exist
*/
static int
update_trusthashtbl( TRUSTREC *tr )
{
return upd_hashtable( get_trusthashrec(),
tr->r.trust.fingerprint, 20, tr->recnum );
}
void
tdbio_dump_record( TRUSTREC *rec, FILE *fp )
{
int i;
ulong rnum = rec->recnum;
fprintf(fp, "rec %5lu, ", rnum );
switch( rec->rectype ) {
case 0: fprintf(fp, "blank\n");
break;
case RECTYPE_VER: fprintf(fp,
"version, td=%lu, f=%lu, m/c/d=%d/%d/%d tm=%d nc=%lu (%s)\n",
rec->r.ver.trusthashtbl,
rec->r.ver.firstfree,
rec->r.ver.marginals,
rec->r.ver.completes,
rec->r.ver.cert_depth,
rec->r.ver.trust_model,
rec->r.ver.nextcheck,
strtimestamp(rec->r.ver.nextcheck)
);
break;
case RECTYPE_FREE: fprintf(fp, "free, next=%lu\n", rec->r.free.next );
break;
case RECTYPE_HTBL:
fprintf(fp, "htbl,");
for(i=0; i < ITEMS_PER_HTBL_RECORD; i++ )
fprintf(fp, " %lu", rec->r.htbl.item[i] );
putc('\n', fp);
break;
case RECTYPE_HLST:
fprintf(fp, "hlst, next=%lu,", rec->r.hlst.next );
for(i=0; i < ITEMS_PER_HLST_RECORD; i++ )
fprintf(fp, " %lu", rec->r.hlst.rnum[i] );
putc('\n', fp);
break;
case RECTYPE_TRUST:
fprintf(fp, "trust ");
for(i=0; i < 20; i++ )
fprintf(fp, "%02X", rec->r.trust.fingerprint[i] );
fprintf (fp, ", ot=%d, d=%d, vl=%lu\n", rec->r.trust.ownertrust,
rec->r.trust.depth, rec->r.trust.validlist);
break;
case RECTYPE_VALID:
fprintf(fp, "valid ");
for(i=0; i < 20; i++ )
fprintf(fp, "%02X", rec->r.valid.namehash[i] );
fprintf (fp, ", v=%d, next=%lu\n", rec->r.valid.validity,
rec->r.valid.next);
break;
default:
fprintf(fp, "unknown type %d\n", rec->rectype );
break;
}
}
/****************
* read the record with number recnum
* returns: -1 on error, 0 on success
*/
int
tdbio_read_record( ulong recnum, TRUSTREC *rec, int expected )
{
byte readbuf[TRUST_RECORD_LEN];
const byte *buf, *p;
int rc = 0;
int n, i;
if( db_fd == -1 )
open_db();
buf = get_record_from_cache( recnum );
if( !buf ) {
if( lseek( db_fd, recnum * TRUST_RECORD_LEN, SEEK_SET ) == -1 ) {
log_error(_("trustdb: lseek failed: %s\n"), strerror(errno) );
return G10ERR_READ_FILE;
}
n = read( db_fd, readbuf, TRUST_RECORD_LEN);
if( !n ) {
return -1; /* eof */
}
else if( n != TRUST_RECORD_LEN ) {
log_error(_("trustdb: read failed (n=%d): %s\n"), n,
strerror(errno) );
return G10ERR_READ_FILE;
}
buf = readbuf;
}
rec->recnum = recnum;
rec->dirty = 0;
p = buf;
rec->rectype = *p++;
if( expected && rec->rectype != expected ) {
log_error("%lu: read expected rec type %d, got %d\n",
recnum, expected, rec->rectype );
return G10ERR_TRUSTDB;
}
p++; /* skip reserved byte */
switch( rec->rectype ) {
case 0: /* unused (free) record */
break;
case RECTYPE_VER: /* version record */
if( memcmp(buf+1, "gpg", 3 ) ) {
log_error( _("%s: not a trustdb file\n"), db_name );
rc = G10ERR_TRUSTDB;
}
p += 2; /* skip "gpg" */
rec->r.ver.version = *p++;
rec->r.ver.marginals = *p++;
rec->r.ver.completes = *p++;
rec->r.ver.cert_depth = *p++;
rec->r.ver.trust_model = *p++;
p += 3;
rec->r.ver.created = buftoulong(p); p += 4;
rec->r.ver.nextcheck = buftoulong(p); p += 4;
p += 4;
p += 4;
rec->r.ver.firstfree =buftoulong(p); p += 4;
p += 4;
rec->r.ver.trusthashtbl =buftoulong(p); p += 4;
if( recnum ) {
log_error( _("%s: version record with recnum %lu\n"), db_name,
(ulong)recnum );
rc = G10ERR_TRUSTDB;
}
else if( rec->r.ver.version != 3 ) {
log_error( _("%s: invalid file version %d\n"), db_name,
rec->r.ver.version );
rc = G10ERR_TRUSTDB;
}
break;
case RECTYPE_FREE:
rec->r.free.next = buftoulong(p); p += 4;
break;
case RECTYPE_HTBL:
for(i=0; i < ITEMS_PER_HTBL_RECORD; i++ ) {
rec->r.htbl.item[i] = buftoulong(p); p += 4;
}
break;
case RECTYPE_HLST:
rec->r.hlst.next = buftoulong(p); p += 4;
for(i=0; i < ITEMS_PER_HLST_RECORD; i++ ) {
rec->r.hlst.rnum[i] = buftoulong(p); p += 4;
}
break;
case RECTYPE_TRUST:
memcpy( rec->r.trust.fingerprint, p, 20); p+=20;
rec->r.trust.ownertrust = *p++;
rec->r.trust.depth = *p++;
rec->r.trust.min_ownertrust = *p++;
p++;
rec->r.trust.validlist = buftoulong(p); p += 4;
break;
case RECTYPE_VALID:
memcpy( rec->r.valid.namehash, p, 20); p+=20;
rec->r.valid.validity = *p++;
rec->r.valid.next = buftoulong(p); p += 4;
rec->r.valid.full_count = *p++;
rec->r.valid.marginal_count = *p++;
break;
default:
log_error( "%s: invalid record type %d at recnum %lu\n",
db_name, rec->rectype, (ulong)recnum );
rc = G10ERR_TRUSTDB;
break;
}
return rc;
}
/****************
* Write the record at RECNUM
*/
int
tdbio_write_record( TRUSTREC *rec )
{
byte buf[TRUST_RECORD_LEN], *p;
int rc = 0;
int i;
ulong recnum = rec->recnum;
if( db_fd == -1 )
open_db();
memset(buf, 0, TRUST_RECORD_LEN);
p = buf;
*p++ = rec->rectype; p++;
switch( rec->rectype ) {
case 0: /* unused record */
break;
case RECTYPE_VER: /* version record */
if( recnum )
BUG();
memcpy(p-1, "gpg", 3 ); p += 2;
*p++ = rec->r.ver.version;
*p++ = rec->r.ver.marginals;
*p++ = rec->r.ver.completes;
*p++ = rec->r.ver.cert_depth;
*p++ = rec->r.ver.trust_model;
p += 3;
ulongtobuf(p, rec->r.ver.created); p += 4;
ulongtobuf(p, rec->r.ver.nextcheck); p += 4;
p += 4;
p += 4;
ulongtobuf(p, rec->r.ver.firstfree ); p += 4;
p += 4;
ulongtobuf(p, rec->r.ver.trusthashtbl ); p += 4;
break;
case RECTYPE_FREE:
ulongtobuf(p, rec->r.free.next); p += 4;
break;
case RECTYPE_HTBL:
for(i=0; i < ITEMS_PER_HTBL_RECORD; i++ ) {
ulongtobuf( p, rec->r.htbl.item[i]); p += 4;
}
break;
case RECTYPE_HLST:
ulongtobuf( p, rec->r.hlst.next); p += 4;
for(i=0; i < ITEMS_PER_HLST_RECORD; i++ ) {
ulongtobuf( p, rec->r.hlst.rnum[i]); p += 4;
}
break;
case RECTYPE_TRUST:
memcpy( p, rec->r.trust.fingerprint, 20); p += 20;
*p++ = rec->r.trust.ownertrust;
*p++ = rec->r.trust.depth;
*p++ = rec->r.trust.min_ownertrust;
p++;
ulongtobuf( p, rec->r.trust.validlist); p += 4;
break;
case RECTYPE_VALID:
memcpy( p, rec->r.valid.namehash, 20); p += 20;
*p++ = rec->r.valid.validity;
ulongtobuf( p, rec->r.valid.next); p += 4;
*p++ = rec->r.valid.full_count;
*p++ = rec->r.valid.marginal_count;
break;
default:
BUG();
}
rc = put_record_into_cache( recnum, buf );
if( rc )
;
else if( rec->rectype == RECTYPE_TRUST )
rc = update_trusthashtbl( rec );
return rc;
}
int
tdbio_delete_record( ulong recnum )
{
TRUSTREC vr, rec;
int rc;
/* Must read the record fist, so we can drop it from the hash tables */
rc = tdbio_read_record( recnum, &rec, 0 );
if( rc )
;
else if( rec.rectype == RECTYPE_TRUST ) {
rc = drop_from_hashtable( get_trusthashrec(),
rec.r.trust.fingerprint, 20, rec.recnum );
}
if( rc )
return rc;
/* now we can chnage it to a free record */
rc = tdbio_read_record( 0, &vr, RECTYPE_VER );
if( rc )
log_fatal( _("%s: error reading version record: %s\n"),
db_name, g10_errstr(rc) );
rec.recnum = recnum;
rec.rectype = RECTYPE_FREE;
rec.r.free.next = vr.r.ver.firstfree;
vr.r.ver.firstfree = recnum;
rc = tdbio_write_record( &rec );
if( !rc )
rc = tdbio_write_record( &vr );
return rc;
}
/****************
* create a new record and return its record number
*/
ulong
tdbio_new_recnum()
{
off_t offset;
ulong recnum;
TRUSTREC vr, rec;
int rc;
/* look for unused records */
rc = tdbio_read_record( 0, &vr, RECTYPE_VER );
if( rc )
log_fatal( _("%s: error reading version record: %s\n"),
db_name, g10_errstr(rc) );
if( vr.r.ver.firstfree ) {
recnum = vr.r.ver.firstfree;
rc = tdbio_read_record( recnum, &rec, RECTYPE_FREE );
if( rc ) {
log_error( _("%s: error reading free record: %s\n"),
db_name, g10_errstr(rc) );
return rc;
}
/* update dir record */
vr.r.ver.firstfree = rec.r.free.next;
rc = tdbio_write_record( &vr );
if( rc ) {
log_error( _("%s: error writing dir record: %s\n"),
db_name, g10_errstr(rc) );
return rc;
}
/*zero out the new record */
memset( &rec, 0, sizeof rec );
rec.rectype = 0; /* unused record */
rec.recnum = recnum;
rc = tdbio_write_record( &rec );
if( rc )
log_fatal(_("%s: failed to zero a record: %s\n"),
db_name, g10_errstr(rc));
}
else { /* not found, append a new record */
offset = lseek( db_fd, 0, SEEK_END );
if( offset == -1 )
log_fatal("trustdb: lseek to end failed: %s\n", strerror(errno) );
recnum = offset / TRUST_RECORD_LEN;
assert(recnum); /* this is will never be the first record */
/* we must write a record, so that the next call to this function
* returns another recnum */
memset( &rec, 0, sizeof rec );
rec.rectype = 0; /* unused record */
rec.recnum = recnum;
rc = 0;
if( lseek( db_fd, recnum * TRUST_RECORD_LEN, SEEK_SET ) == -1 ) {
log_error(_("trustdb rec %lu: lseek failed: %s\n"),
recnum, strerror(errno) );
rc = G10ERR_WRITE_FILE;
}
else {
int n = write( db_fd, &rec, TRUST_RECORD_LEN);
if( n != TRUST_RECORD_LEN ) {
log_error(_("trustdb rec %lu: write failed (n=%d): %s\n"),
recnum, n, strerror(errno) );
rc = G10ERR_WRITE_FILE;
}
}
if( rc )
log_fatal(_("%s: failed to append a record: %s\n"),
db_name, g10_errstr(rc));
}
return recnum ;
}
static int
cmp_trec_fpr (const void *fpr, const TRUSTREC *rec )
{
return (rec->rectype == RECTYPE_TRUST
&& !memcmp( rec->r.trust.fingerprint, fpr, 20));
}
int
tdbio_search_trust_byfpr( const byte *fingerprint, TRUSTREC *rec )
{
int rc;
/* Locate the trust record using the hash table. */
rc = lookup_hashtable (get_trusthashrec(), fingerprint, 20,
cmp_trec_fpr, fingerprint, rec);
return rc;
}
int
tdbio_search_trust_bypk (PKT_public_key *pk, TRUSTREC *rec)
{
byte fingerprint[MAX_FINGERPRINT_LEN];
size_t fingerlen;
fingerprint_from_pk( pk, fingerprint, &fingerlen );
for (; fingerlen < 20; fingerlen++ )
fingerprint[fingerlen] = 0;
return tdbio_search_trust_byfpr (fingerprint, rec);
}
void
tdbio_invalid(void)
{
log_error (_("Error: The trustdb is corrupted.\n"));
how_to_fix_the_trustdb ();
g10_exit (2);
}
/*
* Migrate the trustdb as just up to gpg 1.0.6 (trustdb version 2)
* to the 2.1 version as used with 1.0.6b - This is pretty trivial as needs
* only to scan the tdb and insert new the new trust records. The old ones are
* obsolte from now on
*/
static void
migrate_from_v2 ()
{
TRUSTREC rec;
int i, n;
struct {
ulong keyrecno;
byte ot;
byte okay;
byte fpr[20];
} *ottable;
int ottable_size, ottable_used;
byte oldbuf[40];
ulong recno;
int rc, count;
ottable_size = 5;
ottable = xmalloc (ottable_size * sizeof *ottable);
ottable_used = 0;
/* We have some restrictions here. We can't use the version record
* and we can't use any of the old hashtables because we dropped the
* code. So we first collect all ownertrusts and then use a second
* pass fo find the associated keys. We have to do this all without using
* the regular record read functions.
*/
/* get all the ownertrusts */
if (lseek (db_fd, 0, SEEK_SET ) == -1 )
log_fatal ("migrate_from_v2: lseek failed: %s\n", strerror (errno));
for (recno=0;;recno++)
{
do
n = read (db_fd, oldbuf, 40);
while (n==-1 && errno == EINTR);
if (!n)
break; /* eof */
if (n != 40)
log_fatal ("migrate_vfrom_v2: read error or short read\n");
if (*oldbuf != 2)
continue;
/* v2 dir record */
if (ottable_used == ottable_size)
{
ottable_size += 1000;
ottable = xrealloc (ottable, ottable_size * sizeof *ottable);
}
ottable[ottable_used].keyrecno = buftoulong (oldbuf+6);
ottable[ottable_used].ot = oldbuf[18];
ottable[ottable_used].okay = 0;
memset (ottable[ottable_used].fpr,0, 20);
if (ottable[ottable_used].keyrecno && ottable[ottable_used].ot)
ottable_used++;
}
log_info ("found %d ownertrust records\n", ottable_used);
/* Read again and find the fingerprints */
if (lseek (db_fd, 0, SEEK_SET ) == -1 )
log_fatal ("migrate_from_v2: lseek failed: %s\n", strerror (errno));
for (recno=0;;recno++)
{
do
n = read (db_fd, oldbuf, 40);
while (n==-1 && errno == EINTR);
if (!n)
break; /* eof */
if (n != 40)
log_fatal ("migrate_from_v2: read error or short read\n");
if (*oldbuf != 3)
continue;
/* v2 key record */
for (i=0; i < ottable_used; i++)
{
if (ottable[i].keyrecno == recno)
{
memcpy (ottable[i].fpr, oldbuf+20, 20);
ottable[i].okay = 1;
break;
}
}
}
/* got everything - create the v3 trustdb */
if (ftruncate (db_fd, 0))
log_fatal ("can't truncate `%s': %s\n", db_name, strerror (errno) );
if (create_version_record ())
log_fatal ("failed to recreate version record of `%s'\n", db_name);
/* access the hash table, so it is store just after the version record,
* this is not needed put a dump is more pretty */
get_trusthashrec ();
/* And insert the old ownertrust values */
count = 0;
for (i=0; i < ottable_used; i++)
{
if (!ottable[i].okay)
continue;
memset (&rec, 0, sizeof rec);
rec.recnum = tdbio_new_recnum ();
rec.rectype = RECTYPE_TRUST;
memcpy(rec.r.trust.fingerprint, ottable[i].fpr, 20);
rec.r.trust.ownertrust = ottable[i].ot;
if (tdbio_write_record (&rec))
log_fatal ("failed to write trust record of `%s'\n", db_name);
count++;
}
revalidation_mark ();
rc = tdbio_sync ();
if (rc)
log_fatal ("failed to sync `%s'\n", db_name);
log_info ("migrated %d version 2 ownertrusts\n", count);
xfree (ottable);
}

File Metadata

Mime Type
text/x-diff
Expires
Thu, Dec 4, 1:02 PM (1 d, 14 h)
Storage Engine
local-disk
Storage Format
Raw Data
Storage Handle
c0/ad/54aedf8e6151c19063da97ba5c60

Event Timeline