Red Hat Bugzilla – Bug 439114
Closes when hitting the print button
Last modified: 2008-04-10 07:32:18 EDT
Description of problem:While highlighting an email, and then hitting the print
button, the program closes.
Version-Release number of selected component (if applicable):
How reproducible:Almost always
Steps to Reproduce:
2.highlight an email but don't open
3.Click on the print button
Actual results:The program closes/exits/crashes
Expected results:Should print the email/message
Additional info:After today's updates, it doesn't seem to do it all the time
now. Just on occasion, so I don't know if a particular type code in the email or
whatever else might be causing it.
When Evolution closes do you get a Bug Buddy dialog? If so, can you post the
Bug Buddy report here (with debugging symbols)?
Also, have you noticed whether this occurs for text/html messages or text/plain
messages or both?
No dialog, it just closes with nothing left.
It does it on both I guess. I tested on few emails to the list, and it did it 3
times out of 10. Going back to those same emails after reopening evo, and it
didn't close this time, as it gave me the print dialog. I did try it on a
couple of html ones, from espn and Dallas News sites, and they didn't close it.
So right now, no idea what is causing it.
Maybe try running Evolution from a terminal and look for clues on stdout/stderr.
What you're describing sounds like it might be a missing library symbol.
I'll try reproducing this for myself later today.
Opened it via gnome-terminal and finally got it to close on me, but no debug
messages/errors/nothing at all. See below..
[mike@scrappy ~]$ evolution
CalDAV Eplugin starting up ...
** (evolution:5066): DEBUG: mailto URL command: evolution --component=mail %s
** (evolution:5066): DEBUG: mailto URL program: evolution
libnm_glib_nm_state_cb: dbus returned an error.
(org.freedesktop.DBus.Error.ServiceUnknown) The name
org.freedesktop.NetworkManager was not provided by any .service files
Everything above cept my last prompt was there while it was running. Went
straight to command prompt once it closed with no messages.
Huh, that's bizarre.
Not suprisingly I haven't been able to reproduce this for myself yet, but I'll
scout around for other similar sounding bug reports.
HAHAHAHAHA, kinda ironic, but when I clicked on this email, I accidentally hit
the print button, and evo closed.
Anyway, I just kept going through emails and randomly highlighting and hitting
print to try n get it to close.
Maybe when you've running evolution through gdb, then in time of crash it will
stop and show what's going wrong. You'll be also able to type "thread apply all
bt" and we will see where it get crashed, and maybe why too.
Ok, after yesterday's updates, and today again accidentally clicking on print
button, got a close/crash, but this time bug buddy was used. Will attach it's
report, as well as the url to it being pointed upstread too.
Created attachment 299625 [details]
Print error via bug buddy
Awesome, thanks for the stack trace. The crash is in the CUPS backend of
GtkPrint, so reassigning to gtk2. Do you see these crashes when trying to print
from other GNOME applications like GEdit or Evince?
When printing pdf files, I sometimes get the program to do that. Mostly adobe
reader as I use that.
I have been trying to reproduce this bug but with no success. Could you install
gtk with debuginfo and try to run evolution inside of gdb? And could you send me
"thread apply all bt" result of crash?
You can install gtk with debuginfo via "debuginfo-install gtk2" command.
And again. Could you do that for evolution too? I see that the log shows number
of lines but not for all libraries.
I reproduced the bug and I fixed it. There was problem with function
cups_parse_user_default_printer() (64bit specific). This function calls
g_fopen() but there was no declaration of the g_fopen() (missing #define
I committed changes into cvs and I built it with koji.
Thank you for reporting this bug.
Small correction of previous comment. There should be #include <glib/gstdio.h>
of course. :-)
I'm closing this bug with resolution RAWHIDE.