#768001 icedove segfaults in mime_decode_qp_buffer when trying to forward a specific message

#768001#5
Date:
2014-11-04 00:33:49 UTC
From:
To:
Dear Maintainer,

when I try to forward a specific message, icedove crashes reproducably with a
segmentation fault. The crash happens every time with that message, directly
after I have clicked the "Forward" button, also when I save and open the
offending message from disk and then try to forward it. The behaviour does not
show with some other messages in my inbox.

Appended are the offending mail, and the terminal output of my debugging
session.

(I guess the crash happens because the message says it is in
Content-Transfer-Encoding: quoted-printable, but there are umlauts in the
message body. But this is just a quick guess.)

Cheers,
 - Roland

#768001#12
Date:
2014-11-04 02:22:44 UTC
From:
To:
There seems to be some magic going on between reportbug and the Debian
bugtracker, which caused the offending message to be re-encoded so
Icedove no longer shows the described behaviour. Please see *this*
attachment instead, which is the offending message again, but explicitly
piped through base64(1) by me to (hopefully) prevent any re-encoding.

Cheers,
 - Roland

#768001#17
Date:
2014-11-04 06:10:20 UTC
From:
To:
Hello Roland,

as you tagged this issue is found in the current version in testing and
experimental, what about this bug in earlier versions?

And can you please open up or search a similiar bug in the Mozilla
bugtracker and give us the forward address? I found one related bug
entry in the bugtracker there.
https://bugzilla.mozilla.org/show_bug.cgi?id=469087

But the bug isn't activatly worked on. And the reporter was reporting a
windows version before TB3. There are no crash reports for such a errors.
https://crash-stats.mozilla.com/report/list?product=Thunderbird&query_search=signature&query_type=exact&query=mime_decode_qp_buffer&date=12%2F30%2F2010%2006%3A14%3A17&range_value=4&range_unit=weeks&hang_type=any&process_type=all&plugin_field=&plugin_query_type=&plugin_query=&do_query=1&admin=&signature=mime_decode_qp_buffer#tab-correlations
This makes it hard to reproduce the issue for any developer.

Regards
Carsten

#768001#22
Date:
2015-01-15 03:04:23 UTC
From:
To:
Control: found 768001 icedove/3.0.11-1+squeeze15

I could reproduce it on a squeeze VM running icedove-3.0.11-1+squeeze15. The GDB
output is attached below, and it looks similar. The backtrace even contains more
information (see stack #3, the "line" contains an \n and is strangely off
characters...)

I didn't find anything reated in the upstream tracker, besides that.

OK, so what do I need to do to get a crash report for this kind of behaviour?

 - Roland

#768001#29
Date:
2015-01-15 07:15:34 UTC
From:
To:
Hello Roland,

It's strange and interessing that you see this issue backt to this old
version of Icedove, so possible it's a real old bug that happen not very
often.
But we from the Debian packaging side haven't enough man power to fix
such bugs explicitly in such old versions.

[...]

O.k. then I missed the real point and picked up a wrong upstream bug.

[...]

The right place for fixing such issues is the bugtracker at Mozilla, as
ist's a problem within the upstream source and not related to the Debian
packaging.
So I suggest to open up a new bug entry in the Mozilla bugtracker. But
the developers request mostly a reproducable way to catch the issue, so
please provide as much as possible the informations you can give. Append
your GDB sessions. But please expand before your GDB command to

    'thread apply all bt'

Icedove is a multithreaded application so you want propably see all
informations from all threads.
Please see https://wiki.debian.org/Icedove#Starting_Debugging for more
informations.

If you open up the entry on Mozilla please give back the forwarding
adress and append this to this bug here. It's easiest to do with the
'bts' command. Also set a tag 'upstream'.

    $ bts forwarded 768001 http:/bugzilla.mozilla.org/foo/...
    $ bts tag 768001 upstream

Regards
Carsten

#768001#34
Date:
2015-01-18 16:51:17 UTC
From:
To:
Hello Roland,

can you please attach this email that produces the segfault?
If we can reproduce the segfault it's much easier to catch the real
issue.

Regards
Carsten

#768001#39
Date:
2015-01-20 21:19:59 UTC
From:
To:
Hi,

I had attached the respective mail in my second comment, base64-encoded so the
bugtracker does not convert it in any kind:

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=768001#12

 - Roland