Change of gpg-agent for ECC-SOS

I agree.
It (only) fixed a regression where a user can specify a fingerprint to select a card (rarely used feature in the scdaemon protocol).

Perhaps, no change would be required.
My major concern is that: the data object for fingerprints C5 and C6 were defined as fixed-size 60-byte objects (and actually _is_ defined still in the current specification of 3.4), but it's 80-byte (newer Yubikey), which might cause problem(s).

Hand parsing the data object content:

fa 82 01 e2
   c1 06
      010800001100
   c1 06
      010c00001100
   c1 06
      011000001100
   c1 09
      132a8648ce3d030107
   c1 06
      132b81040022
   c1 06
      132b81040023
   c1 06
      132b8104000a
   c1 0a
      132b2403030208010107
   c1 0a
      132b240303020801010b
   c1 0a
      132b240303020801010d
   c1 0a
      162b06010401da470f01
   c1 0b
      162b060104019755010501
   c2 06
      010800001100
   c2 06
      010c00001100
   c2 06
      011000001100
   c2 09
      122a8648ce3d030107
   c2 06
      122b81040022
   c2 06
      122b81040023
   c2 06
      122b8104000a
   c2 0a
      122b2403030208010107
   c2 0a
      122b240303020801010b
   c2 0a
      122b240303020801010d
   c2 0a
      162b06010401da470f01
   c2 0b
      162b060104019755010501
   c3 06
      010800001100
   c3 06
      010c00001100
   c3 06
      011000001100
   c3 09
      132a8648ce3d030107
   c3 06
      132b81040022
   c3 06
      132b81040023
   c3 06
      132b8104000a
   c3 0a
      132b2403030208010107
   c3 0a
      132b240303020801010b
   c3 0a
      132b240303020801010d
   c3 0a
      162b06010401da470f01
   c3 0b
      162b060104019755010501
   da 06
      010800001100
   da 06
      010c00001100
   da 06
      011000001100
   da 09
      132a8648ce3d030107
   da 06
      132b81040022
   da 06
      132b81040023
   da 06
      132b8104000a
   da 0a
      132b2403030208010107
   da 0a
      132b240303020801010b
   da 0a
      132b240303020801010d
   da 0a
      162b06010401da470f01
   da 0b
      162b060104019755010501

And here is (raw) dump of the data object FA:

Here is the dump of "Application Related Data" (6E):

6e 82 01 47
   4f 10
      d2760001240103040006106160490000
   5f 52 08
      00730000e0059000
   7f 74 03
      810120
   73 82 01 20
      c0 0a
         7d000bfe080000ff0000
      c1 0b
         162b06010401da470f0100
      c2 0c
         122b06010401975501050100
      c3 0b
         162b06010401da470f0100
      da 06  <-------------------------------------- This is algorithm attributes for Attestation key (Yubikey specific)
         010800001100
      c4 07
         ff7f7f7f030003
      c5 50
         eeeed1b50b1b1d9c669033fe019e94a27992b44c
         d00b630fdcb5c4397d5ffbd69aa68a3ff9f8ed10
         1b2a3d46f4f0c5afd0115e7eb858d476daf64cdb
         0000000000000000000000000000000000000000  <--- This appears to be fingerprint of Attestation key
      c6 50
         0000000000000000000000000000000000000000
         0000000000000000000000000000000000000000
         0000000000000000000000000000000000000000
         0000000000000000000000000000000000000000  <--- This appears to be fingerprint of some key related to Attestation key???
      cd 10
         5e58b1e65e58b1c55e58b1f900000000
      de 08
         0102020203028102
   7f 66 08
      02020bfe02020bfe
   d6 02
      0020
   d7 02
      0020
   d8 02
      0020
   d9 02
      0020

In the SOS branch, rG1c4291c3951d: ecc-sos: Add special leading zero octet removal. should be reverted.
Instead, the S_KEY should be fixed up in read_key_file in findkey.c,
and merge_lists in protect.c.
(Then, no need to be fixed up in extract_private_key.)

Exactly same problem is there in libgcrypt.
In the definitions of curves, it uses negative constant internally in some specific places, but for other parts, we have same problems.

I should concentrate the case of ECC, in particular, ECC with modern curves.
Removing leading zero from RSA/ECC/ELGamal assuming unsigned integer would result more work.

In libgcrypt, we have another problem of GCRYSEXP_FMT_ADVANCED formatting, which is used by gpg-agent of GnuPG 2.3 with name-value list.

Confusingly, in the SSH specification, it is signed MPI.
See RFC4251, for the definition of "mpint": https://tools.ietf.org/html/rfc4251#page-8

There are more places for clean up in GnuPG.
While "MPI" in OpenPGP specification is based on unsigned integer, the default "MPI" handling of GnuPG/Libgcrypt is signed. This difference matters internally.
Formatting by "%m" with libgcrypt, it may result prefixed by 0x00 (so that it represents unsigned value, even if scanned as signed).
And because of this, existing private keys in private-keys-v1.d may have this leading zero-byte.
But the counting bits don't count this byte.

Fixed in master and applied to 2.2 branch too.

Important interoperability issue:
OpenPGP implementations should implement:

• Recovery of leading zero octets for Ed25519 key handling (secret part) and Ed25519 signature

Better to paste directly:

# SOS representation
#
# Initially, it was intended as "Simply, Octet String", but 
# it is actually "Strange" Octet String.
#

I wrote this:

libgpg-error used to be blamed because of this kind of architectural support in earlier stage of building operating system.
T4774 is my try to fix the problem.

Thank you for your work. Please go ahead.

Sorry, I was reading the next commit (libdns: Avoid using compound literals (3)).

Possibly, it would be dns_p_init which was caught. If so, it's false positive; It returns a pointer given to the function (which is automatic variable of parent function), but it is valid within the scope of parent function.

Could you please show more information, a specific point of the bug?
I can't locate any place where a function returns a pointer to automatic buffer.

How does it show when you specify --full-time-strings (in UTC by ISO time format)?
I wonder if it is valid as data, but there is a problem of showing key(s).

I tested with this patch (which changes use of constant-time routine when it's secure memory):