Bug 538892 – mail-client/mutt-1.5.23-r5 - mutt when sending encrypted messages: segmentation fault in ? at ?

• Product: Gentoo Linux
• Status: RESOLVED FIXED

[Description slash 2015-02-05 08:15:48 UTC]

Hi,

I have upgraded mutt few weeks ago to mail-client/mutt-1.5.23-r5. It appears that it systematically crashes (segfault) when I want to send an encrypted email. I can reproduce it at will: I write the email, then I choose encrypt or sign+encrypt, and it segfaults just after I push enter. This bug occurs when I try to encrypt emails, but I can still sign them without any segfault.

I downgraded Mutt and now I can encrypt my emails without any problem.

For the moment, I did not take the time to hunt the bug and to identify it. Therefore I do not have any additional detail and I did not report it upstream.


FYI, the USE flags used for mutt:
USE:       berkdb crypt doc gdbm gnutls gpg idn imap mbox nls pop sasl sidebar smime smtp ssl -debug -kerberos -nntp -qdbm -selinux -slang -tokyocabinet

I tried it on my 64bit computer, and on a 32bit hardened one. (same problem)

[Comment 1 Jeroen Roovers (RETIRED) 2015-02-05 09:48:04 UTC]

1) Please post your `emerge --info mail-client/mutt' output in a comment,
2) Please generate a gdb backtrace and attach its output.

[Comment 2 Michal 'vorner' Vaner 2015-02-05 14:24:44 UTC]

It happens to me too. It does not happen when I use „set crypt_use_gpgme“ in configuration. The culprit is selection of the encryption key, when parsing the output of „gpg --no-verbose --batch --with-colons --list-keys email@address“. The -r4 version is fine.

Backtrace:
(gdb) bt
#0  is_email_wsp (c=<error reading variable: Cannot access memory at address 0x0>) at lib.h:115
#1  rfc822_parse_adrlist (top=top@entry=0x0, s=0x0) at rfc822.c:387
#2  0x000000000047f062 in pgp_getkeybyaddr (a=a@entry=0xbd1ce0, abilities=abilities@entry=2, keyring=keyring@entry=PGP_PUBRING) at pgpkey.c:864
#3  0x000000000047c0db in pgp_findKeys (to=<optimized out>, cc=<optimized out>, bcc=<optimized out>) at pgp.c:1236
#4  0x0000000000411c15 in crypt_get_keys (msg=0xbb2a10, keylist=keylist@entry=0x7fff461fd908) at crypt.c:729
#5  0x000000000045eabb in ci_send_message (flags=<optimized out>, flags@entry=0, msg=msg@entry=0xbb2a10, tempfile=<optimized out>, ctx=ctx@entry=0x0, 
    cur=cur@entry=0x0) at send.c:1781
#6  0x0000000000406ba2 in main (argc=2, argv=<optimized out>) at main.c:1037

I'll put the emerge info into an attachment.

[Comment 3 Michal 'vorner' Vaner 2015-02-05 14:26:39 UTC]

Created attachment 395600 [details]
emerge --info output

[Comment 4 Fabian Groffen 2015-02-05 16:56:23 UTC]

A scan through the commits applied to -r5 after -r4 doesn't show an immediate commit to blame.

[Comment 5 Oleh 2015-02-06 16:41:36 UTC]

this happens to me aswell. downgrade solves the segfault

[Comment 6 Fabian Groffen 2015-02-11 08:24:46 UTC]

http://thread.gmane.org/gmane.mail.mutt.devel/21950

I'll apply the patch asap and bump.

[Comment 7 Fabian Groffen 2015-02-11 20:42:33 UTC]

bump should be in -r6

[Comment 8 Fabian Groffen 2015-02-24 10:08:02 UTC]

Patch in -r6