Not sure what you mean by "It is not useful for that purpose" what do you mean
by "It" and "that" in this context.

With regard to complaints there are reports about this that have been active
nearly all the time since this was broken by 2.2.0:
https://wald.intevation.org/forum/message.php?msg_id=2569&group_id=11

This thread triggered your initial opening of this issue.

I agree that wildcard / multifile it is not very important. But I think we
should fix it as it is a small annoyance and it worked with gpg 2.1.0. That it
stopped working was due to our switch to mingw-w64 and not due to a decision
made by us to disable this. So I would see this as a Bug.

Some more threads about this:
https://wald.intevation.org/forum/forum.php?thread_id=1119&forum_id=21&group_id=11
https://wald.intevation.org/forum/message.php?msg_id=3404

P.S.
Yes I dislike the message boards too but they are surprisingly active.

It is not useful for that purpose. This is probably the reason why we received
no complaints about it unless someone uses --multifile with filenames. It is
almost always wrong to do this.

I disagree, all builtin windows tools do globbing so you expect it on the
command line. And as this is default for software that is built with visual
studio and with some versions of mingw it is more likely to see a software that
supports globbing then not.

With regards to quoting problems. I think that users would prefer some problems
instead of just "no". There is no xargs equivalent in Windows by default and so
you have to fall back to really ugly batch scripts that do this in for loops.

Yes one could argue that those users should use a "real shell" on a "real
system" ;-)

Well, it's part of 2.2.2 now, lets see if we get some feedback on this and maybe
for the next release we can revisit the question again if we should enable this
or not.

It would be rather surprising if just one tool does globbing on Windows but all
the otehrs don't do it. It may actually lead to bugs due to broken quoting in
mingw. I recall than many years ago we had a lot of problems with the ever
changing quoting in the globbing implementation.

In any case using globbing with --multifile is not a good idea at all. There is
always a limit on this and thus the proper way of handling this is to use xargs
or whatever the counterpart is on Windows.

Tested this myself again. And I feel comfortable that this is fixed. Still I do
not know why this works now.

Werner can you take a look at the patch I've added for this:

http://git.gnupg.org/cgi-bin/gitweb.cgi?p=gpg4win.git;a=blob;f=patches/gnupg2-2.0.26/0001-Enable-wildcard-expansion-with-mingw-w64.patch;h=6c5b9f8b0794cf32ab909c0a340f81377dc3d42f;hb=HEAD

Is this something you would consider including in some form in gnupg or should I
move the patch into the generic gnupg patch directory?

Now for 2.2.2 we've switched to a new dedicated buildsystem with a fully updated
debian wheezy whereas the old buildsystem was a debian unstable snapshot with
the same mingw Version: 2.0.3-1
I could not find an upstream bug or any indication that there was a fix for this.

But with beta-37 of 2.2.2 I was able to use wildcards on the command line.
I am out of my wits why this works now.

Emanuel: can you confirm that wildcards now work on the command line so that we
can mention this in the news and maybe look at integrating the patch into gnupg?
I'd like a different pair of eyes on that behavior. Maybe I've just messed up my
testsystem.

I've commited the patch to gpg4win so it will be part of the 2.2.2 release.

Thanks for summing up the other problems. I've added a reference to this issue
to the "Improve encoding handling" point in the backlog:
http://wiki.gnupg.org/Gpg4win/Wishlist

Thanks Andre for the patch!

I managed to build gpg4win with the patch added and I verified that it seems to
solve the problem reported by me and also in Issues 1373 and 1674!

But I'd like to summarize the problems related to the charset / codepage on MS
Windows of which I am aware, as a reminder:

1. incorrect display of GPG 2 output translated into another language (also

reported in Issue 1373 and Issue 1674): fixed by your patch;

2. passphrases (both for secret keys and symmetrical encryption) with non ASCII

characters set using GPG 1.4.18 are considered not valid using GPG 2.0.26 and
vice versa

3. incorrect display of filenames with non ASCII characters (also in Issue 1409)

4. GPG 2.0.26 and 1.4.18 ignore or weirdly comply with --utf8-strings, --no-

utf8-strings or --charset options for utf-8 encoding of encrypted filenames (see
Issue 1409)

5. charset weirdness searching keyserver for some non-ASCII user IDs under non-

UTF-8 locales (see Issue 1514 - although relates to Linux it seems to occur also
on Windows, both CLI and GPA but not Kleopatra)

Hope this will help to improve the great GnuPG :-)

This should be working with gpgOl 1.2.1

As I don't have a test set up I'm setting this to resolved until someone complains.

Good anlysis. Thanks.
Feel free to put it as a patch into gpg4win.
I need to look closer at it because we have have the gettext code also in
libgpg-error. You should also send a DCO for GnuPG.

Pretty picture.

D211: 465_0001-Let-wchar_to_native-convert-to-console-codepage.patch

I've taken a look at this. The problem is that the working conversion code in
jnlib/utf8conv.c is not used on Windows but instead jnlib/w32-gettext.c does
it's own conversion to wchar and then back from wchar to the native codepage
which is simpler and should work.

But the conversion back used the wrong codepage. CP_ACP instead of the codepage
retuned by GetConsoleOutputCP. jnlib/utf8conv.c actually had a comment
explaining why it is neccessary to use GetConsoleOutputCP.

With this changed (see attached patch) I get correct output and can verify /
sign files with non-ascii filenames.

I think gnupg master behaves differently though and I don't have a test setup
for this so the patch is only against STABLE.

Werner any objections into including this patch into GnuPG / Gpg4Win?

Thanks Werner. I had already done a search among the issues using the world "charset", but the 1373 issue was
not displayed.

However the bug occurs more problematically with passphrases: a secret key passphrase with non ASCII characters
set using GPG 1.4.18 for a given key is considered not valid using GPG v 2.0.26 to deal with the same key and
vice versa.
I think this problem may cause a lot of trouble to users.

The bug also occurs using filenames in which there are characters outside the ASCII block. This occurs also
setting LC_MESSAGES=C or LC_MESSAGES=en_US.UTF-8 (i.e. gpg -vvvv --verify C:\Users\Andrea\Downloads\gpg4win-
2.2.2-beta37_è_.exe.sig -> "gpg: assuming signed data in `C:\\Users\\Andrea\\Downloads\\gpg4win-2.2.2-
beta37_Þ_.exe'" in which are also wrongly displayed double backslashes instead of single) and also with gpg
1.4.18 (but without the double backslashes). Luckily this does not prevent the signature verification. It's
similar to T1409.

Some additional information:

the CHCP command executed at the Windows command prompt reports: "Tabella codici attiva: 850" (active codepage
850);

using the -vvvv option with the --verify command, gpg 2.0.26 and 1.4.18 report: "gpg: using character set
`CP850'".

This is known. See the other bug report.

Duplicate of T1373

Ok using the same compiler I've used a year ago I also get the same result I got
a year ago (phew I thought i screwed up the test).

But with the patch this should just work once we change the used mingw-version.
I could not find a bugreport or some other information where and when this is fixed.

I've tested this again with the mingw version that is part of ubuntu and it works.

I've commited a patch to gpg4win to enable this and will check if this also
works with the debian wheezy mingw-crt version.

Thanks for the explanations.

Further analysis showed that the second call to AllowSetForeground window was
blocked by the ForegroundWindowLockTimeout. If you set this timeout to zero
everything worked as expected.

There is a discussion on this under:
http://social.msdn.microsoft.com/Forums/vstudio/en-US/09fa16ba-c6ef-410d-bec2-a99a9b9a98d9/issue-with-foregroundlocktimeout-setforegroundwindow

Which suggests a solution of setting the foregroundlocktimeout to zero before
attempting to set the foreground window. I've found this solution not applicable
for our case as we run into the lock on "AllowSetForegroundWindow" and not when
actually setting the foreground Window.

Another workaround is to call AttachThreadInput on the current foreground thread
call SetForegroundWindow and then detach again. There might be problems with
this approach so it is just used as fallback. But it should resolve this problem
for now.

What I described is done for all pinentries.

Ok, thanks I'll check what happens if

5. The agent thinks that the passphrase entered is too weak
6. The agent starts another pinentry
7. The agent sends "getinfo pid" to the pinentry (this i can see)
8. ?

1. The agent starts pinentry.
2. The agent sends "getinfo pid" to the pinentry"
3. The agents sends an "INQUIRE PINENTRY_LAUNCHED <pid>" to the caller.

4a. If the caller is gpg, gpg calls AllowSetForegroundWindow for the received
pid; gpg has been started via gpgme and gpgme has called ASFW for gpg.
4b. If the caller is gpgsm, it proxies the PINENTRY_LAUNCHED to gpgme which then
calls ASFG for the pid.

The different methods are required because gpg is a one-off process while gpgsm
may be used several times.

Yes Kleopatra uses gpgme for key generation. Still I don't see how gpgme figures
into this sorry. Should gpgme start gpg-agent and ensure that it has the
setforeground window access right?

1. The gpg-agent asks for a passphrase with pinentry -> set foreground window is

allowed.

2. If that passphrase is weak it launches a new pinentry with a confirm dialog

-> set foreground window is not allowed.

And the agent does not have the right to call allowsetforegroundwindow. In the
first case I think this might be because another component gets the PID of the
pinentry and calls allowsetforeground window (have to do further debugging to
check this)

In the second case afaik only gpg-agent is involved and no one sets
allowsetforeground window on the pid of the second pinentry.

BTW, gpgme 1.5.1 has a spawn interface which is better than to code your own in Qt.

It goes all the way back to GPGME via the assuan interface. grep for
_gpgme_allow_set_foreground_window. For GPG the assuan interface is not used.
We do it by passing our internal IOSPAWN_FLAG_ALLOW_SET_FG flag to the
gpgme-w32spawn.c helper which then uses this to call set foreground API.

Does Kleopatra always use gpgme?

I won't call that leaking memory.

With your fix it no longer crashes. But you are leaking memory in case
default_homedir returns allocated memory from read_w32_registry_string or
standard_homedir with a successful call to w32_shgetfolderpath.

As the result should be cached in the static variable this is probably not an
issue..

I just pushed a fix.

The problem with gpgme is that we need to stop the Explorer if we want to update
or uninstall gpgme. Given tha we can't use gpgex with the portable version
anyway, I think it is better to keep the existing code and don't switch to gpgme.

Using gpgme would be nice of course to avoid duplicated functionality (this same
stuff is also implemented in gpgol although without the bug) but remember that
we would then also need a x64 version of gpgme.

The bug is pretty obvious. I consider to rewrite it by calling gpgconf
--list-dirs to get the home directory. This would be simliar to what we do in
gpgme. Or well, we could link to to gpgme to and make use of its higher level
interface.

Thanks for looking at this. If you like I could test your fixes. I currently
have a build setup and a test setup where I can reproduce the crash.

Btw. Maybe I don't understand c++ enough but afaik this just makes no sense and
is broken:

  try { name = ((string) dir) + "\\S.uiserver"; } catch (...) {}

should be rather something like:
if (dir)
name = string(dir) + "\\S.uiserver";

I have some fixes sitting here in my local repo. I need to check them.

sorry can't delete

Mmmh when I run:

#include <stdio.h>
int _dowildcard = -1;
int main (int argc, char **argv)
{

int i = 0;
for ( i = 0; i< argc; i++) {
    printf("%s\n", argv[i]);
}
return 0;

}

With the command line *.* it just prints "*.*". Linking with CRT_glob.o directly
also leads to the same result.

Hi,
there is apparently a compile option "enable-wildcard" for the mingw-w64 c runtime.

But from:
http://mingw-w64.sourcearchive.com/documentation/0~20100125-3/wildcard_8c-source.html

I take it that you can also enable it in the program by defining int _dowildcard

-1;

I'll test it and let you know if this worked.

works with Gpg4win 2.2.0. But there is a failure after decrypting folder with
umlauts: see gpg4win wald issue #6451
(https://wald.intevation.org/tracker/index.php?func=detail&aid=6451&group_id=11&atid=126)

Fixed in d3f20e7. Applied patch to gpg4win as well.

In the backtrace gpa_backtrace2.png you can see that the crash happens:
https://bugzilla.gnome.org/show_bug.cgi?id=696232

If we apply the patch it no longer crashes, but as the error handling in that
patch is just to abort, the text is not copied into the clipboard and we do not
get an error in GPA.

So still some more debugging why invalid utf-8 data is sent from gpa.
As the data worked on comes from GPA and "should" be valid utf-8

Backtrace attached. Sorry that it is an image but windows was blocking me from
copying the gdb output to the clipboard.

Let's assume that the problem was in the glib library and has been fixed in the
meantime. I will soon release a new Gpg4win version with the latest stable glib
version. Feel free to reopen if you still encounter this problem.

Yeah sure, I meant "NOT a problem".

Yes I know what you mean. But without locking you will never be able
to get it right.

As you noted there is actually a problem with Libgcrypt under
Windows. In Libgcrypt we lock the seed file and thus the fatal error
is the right thing to do. But.....

  #ifdef __GCC__
  #warning Check whether we can lock on Windows.
  #endif
  #if LOCK_SEED_FILE

Thus we should implement locking for Windows. The problem here is
that there is no portable advisory locking in Windows. And frankly our
fcntl locking approach does not work on all Unices either and worse it
does not work if the home partition is on certain remote file systems.

The only solution I see is to employ the new dotlock code from GnuPG
here. It is slower than fcntl locking but very portable.

I don't think we will fix this in gpg 1.4.

This is known but a problem from a security POV.

I suppose you mean "NOT a problem"? I think it might be a problem in
opportunistic encryption scenarios if gpg encryption failures caused by
random_seed access conflicts are ignored like failures caused by missing keys.
But usually it's just a nuisance like any other randomly failing program.

The non-locking read is on purpose - if it works: okay. Otherwise we
re-generate a seed file.

I see that the code tries to tolerate access conflicts, but there's still a race
condition if the random_seed file is truncated between fstat() and read(). The
read() error handling is incomplete. Maybe this pseudo-patch explains best what
I mean:

diff -ru gnupg-1.4.12/cipher/random.c gnupg-1.4.12/cipher/random.c

• gnupg-1.4.12/cipher/random.c 2012-01-24 09:45:41.000000000 +0100

+++ gnupg-1.4.12/cipher/random.c 2012-09-18 19:47:54.449578800 +0200
@@ -492,11 +492,17 @@

    do {
	n = read( fd, buffer, POOLSIZE );
    } while( n == -1 && errno == EINTR );

• if( n != POOLSIZE ) {

+ if( n == -1 ) {

	log_fatal(_("can't read `%s': %s\n"), seed_file_name,strerror(errno) );
	close(fd);
	return 0;
    }

+ else if ( n == 0 ) {
+ ... handle like sb.st_size == 0
+ }
+ else if ( n != POOLSIZE ) {
+ ... handle like sb.st_size != POOLSIZ
+ }

     close(fd);

Yes, we could do the file locking

Proper locking would be the ideal solution, but a better read() error handling
would already be sufficient to avoid the sporadic fatal errors on random_seed
accesses.

and iirc, we do this in libgcrypt (GnuPG-2).

I'm just checking this and ... sorry, no. The gpg2.exe from Gpg4win 2.1.0 shows
the very same error:

note: random_seed file is empty
note: random_seed file is empty
Fatal: can't read `C:/Dokumente und
Einstellungen/jechternach/Anwendungsdaten/GnuPG /random_seed': No such file or
directory

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
note: random_seed file is empty

This is known but a problem from a security POV. The seed file is a cache. If
the seed file can't be read a new one is created. Right, there might be a
performance issue but at least on Windows this is not as severe as on certain
Linux systems.

The non-locking read is on purpose - if it works: okay. Otherwise we
re-generate a seed file. Yes, we could do the file locking and iirc, we do this
in libgcrypt (GnuPG-2).

The error on short reads is also on purpose. We want those 600 bytes at once
and nothing else. If another process is writing the seed file we may see the
short reads. But in this case there is no clear answer waht to do - thus we
assume the "no seed file" case.

Thanks, Werner, for your reply and your suggestions. I'll have a look at the
GPGME solution.