Red Hat Bugzilla – Bug 453421
Evolution fails to start on x86_64
Last modified: 2008-07-09 04:55:43 EDT
Description of problem:
Evolution is unable to start due to a glibc problem
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. start evolution from either the command line or desktop icon
with thread apply all bt using gdb enabled
[Thread debugging using libthread_db enabled]
[New Thread 0x7ffff7f9e7a0 (LWP 8146)]
CalDAV Eplugin starting up ...
** (evolution:8146): DEBUG: mailto URL command: evolution %s
** (evolution:8146): DEBUG: mailto URL program: evolution
GLib-ERROR **: gmem.c:136: failed to allocate 31740376080 bytes
Program received signal SIGTRAP, Trace/breakpoint trap
IA__g_logv (log domain=<value optimized out>, log_level=<value optimized out>,
format=<value optimized out>, args1=<value optimized out>) at gmessages.c:503
503 g_private_set(g_log_depth, GUINT_TO_POINTER(depth));
Evolution should start
Untested on x86
You probably meant glib2 here.
Created attachment 310649 [details]
Output from valgrind
It can be also because of broken summary file in one of your local folders. The
place where it crashed is in time of creating the folder structure and loading
message from the summary. If you try to recreate index/summary files from your
~/.evolution/mail/local directory, (and subdirectories), then there's some
chance to get this fixed. More info how to recreate summary/index files can be
Please do backup of the 'local' folder before doing anything, to be able to
return to some data when something goes wrong.
Also, the above link describes the error on an Inbox folder, but it can be in
any folder you have, from the valgrind output is not obvious which folder is in
I don't think this is a glib problem, really.
Moving back to evolution.
I don't think it's a glibc problem either. Looks like another 64-bit
miscalculation in libcamel. We've seen this before with libcamel.
Paul, can you get a backtrace for this?
I've done as in comment #3 which seems to have sorted the problem, however,
there is already a valgrind output attached to this bug, though I suspect you
need a libcamel one...
Okay, so are you still unable to start then or can the bug be closed?
I'll close it.