Bug 129527
Summary: | Evolution crashes when receiving mail | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Thomas J. Baker <tjb> |
Component: | evolution | Assignee: | Matthew Barnes <mbarnes> |
Status: | CLOSED NEXTRELEASE | QA Contact: | |
Severity: | high | Docs Contact: | |
Priority: | high | ||
Version: | 3 | CC: | dave, dmalcolm, gerardo, kain, kiosso, mattdm, temlakos |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | i686 | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2007-04-06 15:43:01 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: | |||
Attachments: |
Description
Thomas J. Baker
2004-08-10 01:45:37 UTC
Created attachment 102554 [details]
bug buddy stack trace
FWIW, the ximian bug is http://bugzilla.ximian.com/show_bug.cgi?id=62693 The latest evo doesn't fix the problem (evolution-1.5.93-2). If I disable the single comcast pop account, the other local pop and remote imap accounts work perfectly for days at a time. Just for kicks last night, I enabled the comcast account and evo crashed within 5 minutes. Thanks. Looking at the stack trace it looks like a crash inside camel_gethostbyname; might be related to bug #127225 (perhaps a general problem with gethostbyname); I'll investigate further... I no longer have this problem. I turned back on the troublesome account and it has been working for a few weeks now. I don't recall if it was a glibc update or an evolution update that fixed it but it definitely seems fixed. Please can you specify which versions you're running in which it seems to work: rpm -q evolution rpm -q glibc Thanks! evolution-1.5.94.1-1 and glibc-2.3.3-47 work now but I know it was working with the prior version of Evolution and glibc too since glibc-2.3.3-47 was installed on 9/1/2004 and evolution on 9/6/2004. Created attachment 106598 [details]
BugBuddy-generated report on my latest crash of Evolution
To reproduce this on a Fedora Core 2 system after the update of the glibc
packages distributed 11/11/2004:
Start Evolution, and either:
Hit the Send/Receive button, or else:
Wait for Evolution to try to send/receive mail in the background.
This is a total blocker--I had to switch to Mozilla Mail just to keep on
transacting business. I cannot, therefore, recommend Evolution to a client if
it is subject to such breakage. The correspondence I have received before
states that glibc might be responsible in some way. Neither Ximian nor Gnome
want to admit that the problem is with their code.
*** Bug 139573 has been marked as a duplicate of this bug. *** Although the original report is against Evolution 1.5, this looks like a duplicate of this massively-duplicated bug against Evolution 1.4 in Ximian's bugzilla: http://bugzilla.ximian.com/show_bug.cgi?id=69431 I've added a link to this bug into that bug Regarding comment #8: I'm trying to narrow down what changes occurred to glibc to cuase the problem. Can you tell me when was the last time you updated glibc before 11/11/2004 ? (and ideally, exactly which version of the package you had installed when it was working) What version of the package do you now have installed? Thank you. Regarding Comment #11: I don't remember when the last time was that I updated before November 11, 2004. I *can* tell you that, in general, I update whenever up2date says that I should. Therefore, I had the *very last version before the update*. Before I updated, Evolution never gave me a lick of trouble. Once I updated, Evo 1.4.6 no longer worked. (I now use Evo 2.0.2, which was the only way I could restore Evolution functionality.) The output of command string "rpm -q glibc" (issued as root) is: glibc-2.3.3-27.1 The output of command string "rpm -q glibc-common" is: glibc-common-2.3.3-27.1 As you can see, both those version numbers are consistent. I also have the packages glibc-devel and glibc-headers installed, and the version number is the same. (I just checked that out with Synaptic.) Thanks for the clarification; this is an FC2 box, right? the last update of glibc for FC2 was in May 2004 (glibc-2.3.3-27); the recent update 2.3.3-27.1 contained a large number of fixes; I'm looking at them now to see what might have caused this breakage. Re Comment #14, yes, that is correct: this is an FC2 box. I never had to update glibc until recently--and that's when the trouble started. I also have been experiencing this since the last update to FC2. I also have Comcast as my ISP, which seems like it might be relevant. Until it is fixed, could it make sense to retrograde glibc? Re Comment #16: Either roll back your glibc *or* manage to upgrade to Evolution 2.0.2. You'll need the following as a special devel.conf file: [main] #cachedir=/var/cache/yum debuglevel=2 logfile=/var/log/yum.log pkgpolicy=newest distroverpkg=fedora-release gpgcheck=1 tolerant=1 exactarch=1 [freshrpm - devel] name=Fedora Linux $releasever - $basearch - fedora - development baseurl=http://ayo.freshrpms.net/fedora/linux/development/i386/core Paste the above into a file called devel.conf. Either do this as root or else run: chown root /etc/devel.conf Then run (as root): yum -c /etc/devel.conf update evolution Thirty-four RPM's later, you'll at least be able to use Evo 2.0.2 and get past this problem. Or, you could fall back on Mozilla Mail, which has *almost* all of Evo's features. Just FYI, I upgraded two FC2 systems to glibc and glibc-common 2.3.3-27.1 this week. On the AMD/ACPI machine evolution broke in the manner described above, but no bad behavior on the Intel/AMP machine. I rolled back to plain 2.3.3-27 (no .1) by downgrading to the original glibc* rpms off of CD1, and now evolution works again on the AMD machine. Yeah, I tried to upgrade using the yum config in Comment #17 and, after 1500+ header updates, got the following: Resolving dependencies .............Unable to satisfy dependencies Package openoffice.org-libs needs libdb_cxx-4.2.so, this is not available. As a final effort, I tried to do an RPM update on evo 2.0.2 and still got dependency failures (which I expected.) I'm rolling glibc* back to 2.3.3-27 until I do a fork-lift upgrade to FC3. Please, if you're seeing this bug, and haven't yet upgraded, can you install these packages: - evolution-debuginfo - glibc-debuginfo - glibc-debuginfo-common ...and then generate a stack backtrace, and post it here. See this link for more information about how to do this: http://fedora.linux.duke.edu/wiki/index.cgi/StackTraces This would be very useful; so far the backtraces I've seen aren't with the -debuginfo packages; with the debuginfo they should yield much more information on what's gone wrong. Thanks. Created attachment 107084 [details]
backtrace w/debuginfo
here is a back/stacktrace. Dave, please let me know if it doesn't meet your
needs.
Created attachment 107673 [details]
evolution crash debug info after installing debug packages
I've added an attachment of my debug info after installing the debug packages.
Created attachment 107726 [details]
Backtrace as requested on #evolution
From ximian bug: After reading comments about this bug for about 2 hours, I've concluded that there is a bug in Fedora's latest (2.3.3-27.1) upgrade of glibc libraries. (Please read bug # 129527.) I've upgraded the glibc, glibc-common, and glibc-headers to v2.3.3-74 from Fedora Core 3, and the bug's gone! For added confirmation, I've re-installed the 2.3.3-27.1 version of the above libraries, and the bug's back. Changing version to correct one. (test1 -> fc3test1, and some were filed as test3 accidentally instead; but clearly must be fc3test1 given the date of filing.) If this was really a glibc bug I imagine the problem has long since been fixed, but is anyone still encountering this crash in Fedora Core 6 or later? Fedora Core 3 and Fedora Core 4 are no longer supported. If you could retest this issue on a current release or on the latest development / test version, we would appreciate that. Otherwise, this bug will be marked as CANTFIX one month from now. Thanks for your help and for your patience. As I understand it, this is not an issue in current releases. Marking fixed now. Please reopen if I'm wrong. |