GNOME Bugzilla – Bug 324327
Why does evolution try to guess account?
Last modified: 2013-09-13 00:57:08 UTC
This bug has been opened here: http://bugzilla.ubuntu.com/show_bug.cgi?id=20766 "I've been trying to use the ubuntu/breezy version of evolution (2.4.1-0ubuntu7) and evolution-exchange (2.4.1-0ubuntu1) for some time now. Sometimes, when I open mails that I have moved from exchange inbox to my local folders, evolution crashes. Backtrace shows that evo is trying to figure out whether it should do some receipt handling (dunno why this would be needed when opening a filed message) and ends up accessing a null pointer. The em_utils_handle_receipt() in em-composer-utils.c calls guess_account() which returns zero (dunno why the account needs to be "guessed") probably because the "url" seems to be a file name. I added checking for null before using the account var but something else is probably more appropriate.. (gdb) where
+ Trace 64660
This is 'Evolution' 'Mailer' issue, so changing component. Seb: Bug has been raised for 'Evolution' crashes ? if not please raise bug for that issue.
what do you mean? this bug is about the crash and has a backtrace
REOPENing.
Is it possible to attach the message here. It would just make it easier for me to write and test a fix. Probably just the null check might do. Please remove any confidential info in the mail and attach. Thanks
*** Bug 326801 has been marked as a duplicate of this bug. ***
*** Bug 326179 has been marked as a duplicate of this bug. ***
Confirming as per duplicates.
I really need to get a copy of the mail. Changing to NEEDINFO. Please reopen when you provide the necessary info
*** Bug 329387 has been marked as a duplicate of this bug. ***
Created attachment 58651 [details] This was requested by the developer (Susarla P.) for bug# 324327. Bug status was changed to NEEDINFO pending this document. the attachment is a zip archive but this option was not available. Hence selected tar. The archive consists of html & pdf files. The mail consisted of the following text and the attached archive.(I have foobarred my name) text start feestructure doc has detailed DD info -> important also reg form is included. Regards, foobar text end
(In reply to comment #8) > I really need to get a copy of the mail. Changing to NEEDINFO. Please reopen > when you provide the necessary info > Please change the status to REOPEN. I couldn't find how to do that.
This bug is fixed on head. Thanks for the info frodo. :)
*** Bug 331527 has been marked as a duplicate of this bug. ***
Bug is still here. The fix, introduced in em-composer-utils.c 1.50 (saps) is missing a one liner: --- em-composer-utils.c.orig 2006-03-03 17:28:25.000000000 +0100 +++ em-composer-utils.c 2006-03-03 15:04:14.000000000 +0100 @@ -1122,7 +1122,7 @@ } if (account && account->receipt_policy == E_ACCOUNT_RECEIPT_ALWAYS - || ( account->receipt_policy == E_ACCOUNT_RECEIPT_ASK + || ( account && account->receipt_policy == E_ACCOUNT_RECEIPT_ASK && e_error_run (NULL, "mail:ask-receipt", addr, camel_mime_message_get_subject(msg)) == GTK_RESPONSE_YES)) em_utils_send_receipt(folder, msg); }
Without the patch above, I am able to reproduce. The fix was testing for account not being null before dereferencing account->receipt_policy, but it was only done on one line of the test, not on the second line. Could you please reopen this bug till the fix is complete ? Thank you
reopening as per last comment. i guess that bug 333921 is a duplicate, not entirely sure though. targetting to 2.6.
Thank you for reopening. And yes bug 320661, bug 324602, and bug 333921 seem to be duplicates of this one (the crash descriptions and backtraces match). Since the fix is a quite simpe one line patch, maybe an earlier release could be targetted ?
thomas: there isn't anything earlier. we're currently under hard code freeze. partha: ping?
Created attachment 63492 [details] [review] 324327.patch
It would be great to mark bug 320661, bug 324602, and bug 333921 as duplicates of this one (which I would do, if my account had the permission). Thank you!
*** Bug 320661 has been marked as a duplicate of this bug. ***
*** Bug 324602 has been marked as a duplicate of this bug. ***
*** Bug 333921 has been marked as a duplicate of this bug. ***
committed to CVS
*** Bug 342653 has been marked as a duplicate of this bug. ***
*** Bug 339361 has been marked as a duplicate of this bug. ***
I don't hope we're facing a regression with bug 519683.