GNOME Bugzilla – Bug 272632
crash when accepting meeting invitation
Last modified: 2013-09-13 00:50:36 UTC
This bug has been opened here: https://bugzilla.ubuntu.com/6604 "I have a ms outlook meeting invitation (mail is from exchange account, trying to send to exchange calendar) that crashes everytime I try to accept it. I can accept it fine in the evolution in debian unstable, but it crashes in hoary (evo 2.1.5). The email has a word document attachment in addition to the meeting invitation. Backtrace was generated from '/usr/bin/evolution' Using host libthread_db library "/lib/tls/i686/cmov/libthread_db.so.1". `system-supplied DSO at 0xffffe000' has disappeared; keeping its symbols. [Thread debugging using libthread_db enabled] [New Thread -1232513504 (LWP 31909)] [New Thread -1303348304 (LWP 32665)] [New Thread -1293108304 (LWP 32657)] [New Thread -1248117840 (LWP 32537)] [New Thread -1282364496 (LWP 32455)] [New Thread -1273877584 (LWP 32422)] [New Thread -1265484880 (LWP 32421)] [New Thread -1257092176 (LWP 32420)] [New Thread -1239725136 (LWP 32416)] 0xffffe410 in __kernel_vsyscall ()
+ Trace 55752
please attach an example message. not likely to be fixed otherwise.
i can't attach the email here, but who should I forward it to if I want somebody to look at this?
send it to me
I'm seeing similar crash with 2.2.1.1. I can't provide an example, but debugging the issue a little I see that em_utils_temp_save_part() is called with part = NULL
Created attachment 45137 [details] [review] proposed patch The crash appears to go away with this patch, now camel_mime_message_get_part_by_content_id() returns non-NULL value. A better fix could possibly be to make camel_mime_message_get_part_by_content_id() accept cids including the 'cid:' prefix as well, but throw out the prefix before processing.
hmm, the patch should work, get_part_by_content_id() should NOT perform any processing on the string at all. the code should also check the part isn't null before using it - there is no guarantee the message will contain the part anyway. that function also looks somewhat questionable, it should probably using the em_format functions to look for parts, as they should be goverened (i guess?) by the multipart/related rules. i.e. this code will not properly handle multiple attachments with the same cid, in a different sub-tree of the message. which is possible, if not a good case. PS i never got the test message
forwarding test message now. (lost the old one, but just got a new one from the same source that displays the same behavior)
This looks like harish's attachment code, moving to him.
*** Bug 300255 has been marked as a duplicate of this bug. ***
*** Bug 300712 has been marked as a duplicate of this bug. ***
*** Bug 302060 has been marked as a duplicate of this bug. ***
*** Bug 304418 has been marked as a duplicate of this bug. ***
*** Bug 307122 has been marked as a duplicate of this bug. ***
*** Bug 306427 has been marked as a duplicate of this bug. ***
For anyone interested in this bug, in April, I had applied the patch above on my Debian system. I have been using it ever since, with this bug never reoccurring. I'm curious if/when this patch (or other fix) is planned to make it into a release of Evolution to close this bug?
I am able to reproduce this with evolution 2.3.x using exchange account. Debugging Info:
+ Trace 61941
Thread 1 (Thread 1097772672 (LWP 23517))
Created attachment 49560 [details] attached the example meeting request that crashes evolution
Fixed comitted to CVS Head.
*** Bug 313151 has been marked as a duplicate of this bug. ***
*** Bug 320885 has been marked as a duplicate of this bug. ***
*** Bug 319345 has been marked as a duplicate of this bug. ***
*** Bug 303958 has been marked as a duplicate of this bug. ***
*** Bug 326355 has been marked as a duplicate of this bug. ***
*** Bug 326894 has been marked as a duplicate of this bug. ***
*** Bug 327883 has been marked as a duplicate of this bug. ***
*** Bug 316973 has been marked as a duplicate of this bug. ***
*** Bug 335449 has been marked as a duplicate of this bug. ***