View previous topic :: View next topic |
Author |
Message |
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Mon Aug 30, 2004 7:03 pm Post subject: Evolution 1.4.6 Freezes with no network |
|
|
Hi all,
Probably a strange question for a program that's made for the internet, but anyways...
I'm wondering if anyone else has had the problem of Evolution freezing only when there's no network available. I'm using a laptop, and wanted to reply to e-mails offline while on the train to work, to have them be sent from the Outbox when I reconnect to the network. But when I get in (if I get in), and it auto-checks mail (or I click send/receive), I get the "unexpectedly crashed" error.
Running evolution from the command line doesn't really provide any feedback. I do get this occasionally:
Code: |
(evolution:21919) GLib-GObject-WARNING **: g_object_set_valist: construct property "type" for object 'ESplash' can't be set after construction.
(evolution:21919): GLib-GObject-WARNING **: gsignal.c:1738: instance '0x837a940' has no handler with id '2619'
|
Any thoughts?
Thanks,
Brian |
|
Back to top |
|
|
nightblade Guru
Joined: 20 Jul 2004 Posts: 368 Location: back from SE Asia
|
Posted: Thu Sep 02, 2004 10:38 am Post subject: |
|
|
When working offline, you should click on the icon on the bottom left (sort of representing two plugs attached), in order to tell Evolution not to check for e-mails, thus not trying to access network devices that are not ready to handle the request. _________________ In God we trust. All the others must provide a valid X.509 certificate |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Thu Sep 02, 2004 12:56 pm Post subject: |
|
|
Usually that works no problem (going to 'Work Offline'), but there are times that I can't even get Evolution to start without having the network connected. Is there any way to just have it time out instead of crashing? Would even crash when the network is disconnected even when I'm at home... doesn't seem normal.
Thanks,
Brian |
|
Back to top |
|
|
nightblade Guru
Joined: 20 Jul 2004 Posts: 368 Location: back from SE Asia
|
Posted: Fri Sep 03, 2004 9:07 am Post subject: |
|
|
that actually seems to be a bug then... have you tried to perform a 'strace' when you launch evolution in order to figure out at what exact point it crashes ? _________________ In God we trust. All the others must provide a valid X.509 certificate |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Sun Sep 05, 2004 12:31 pm Post subject: |
|
|
Here's the last chunk of strace:
Code: |
write(3, "5\20\4\0\342\6 \1&\3 \1\352\1\32\0;\3\5\0\252\0 \1\0\0"..., 536) = 536ioctl(3, FIONREAD, [0]) = 0
gettimeofday({1094390798, 289302}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=8, events=POLLIN|POLLPRI}, {fd=10, events=POLLIN}, {fd=13, events=POLLIN|POLLPRI}, {fd=14, events=POLLIN|POLLPRI}, {fd=15, events=POLLIN|POLLPRI}, {fd=12, events=POLLIN|POLLPRI}, {fd=17, events=POLLIN|POLLPRI}, {fd=16, events=POLLIN|POLLPRI}, {fd=22, events=POLLIN}, {fd=37, events=POLLIN|POLLPRI}, {fd=18, events=POLLIN}, {fd=20, events=POLLIN}], 14, 596185) = -1 EINTR (Interrupted system call)
--- SIGSEGV (Segmentation fault) @ 0 (0) ---
write(3, "\33\20\2\0\0\0\0\0", 8) = 8
write(3, " \20\2\0\0\0\0\0", 8) = 8
write(3, "+\20\1\0", 4) = 4
read(3, "\1\2\0225\0\0\0\0009\1 \1\0\0\0\0\1\0\0\0\20\0\0\0\210"..., 32) = 32
fork() = 12398
waitpid(12398, [WIFEXITED(s) && WEXITSTATUS(s) == 0], 0) = 12398
--- SIGCHLD (Child exited) @ 0 (0) ---
sigreturn() = ? (mask now [SEGV RTMIN])
exit_group(1) = ?
|
Any clues? I tried to keep mouse movements to a minimum to avoid all the poll() calls, but it still picked some up. If you'd like the entire strace return let me know... it's quite long.
Thanks,
Brian |
|
Back to top |
|
|
nightblade Guru
Joined: 20 Jul 2004 Posts: 368 Location: back from SE Asia
|
Posted: Mon Sep 06, 2004 3:37 pm Post subject: |
|
|
That doesn't say much to me, but I am not a guru when it comes to debugging...
Anyway, feel free to send me the whole trace (if I am not able to figure out what's wrong, somebody else in my office might).
Another thing: you used particularly aggressive settings in your CFLAGS ? _________________ In God we trust. All the others must provide a valid X.509 certificate |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Mon Sep 06, 2004 4:01 pm Post subject: |
|
|
My current USE flags are:
USE="X mysql gtk gnome -kde -qt alsa oss opengl scanner usb mozilla mpeg mad jpeg java gtk2 avi mmx sasl ssl cups -esound imap maildir libwww"
The strace file is quite large, and pretty repetitive... might be better for e-mail if you'd like to look at it.
Is there a way to try another version of evolution? If I remember correctly 1.4.6 isn't the newest version, and downgrading to a previous version might help with the stability... |
|
Back to top |
|
|
nightblade Guru
Joined: 20 Jul 2004 Posts: 368 Location: back from SE Asia
|
Posted: Mon Sep 06, 2004 9:13 pm Post subject: |
|
|
The 1.4.6 is the latest stable. There are also the 1.5.92.1 and 1.5.93 which are marked unstable, tho.
I use the 1.4.6 with no problem. What about your CFLAGS, however ? You can find them in the file make.conf. I use these:
Code: |
CFLAGS="-O3 -march=pentium4 -pipe -fomit-frame-pointer -funroll-loops"
|
and evolution is rock stable. Maybe you compiled it with different (and more aggressive) flags. _________________ In God we trust. All the others must provide a valid X.509 certificate |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Tue Sep 07, 2004 1:21 am Post subject: |
|
|
No luck... the only thing I had different was the lack of a -funroll_loops, everything else was the same. Recompiled and still crashes when internet connectivity is lost.
Any other ideas/suggestions? |
|
Back to top |
|
|
nightblade Guru
Joined: 20 Jul 2004 Posts: 368 Location: back from SE Asia
|
Posted: Tue Sep 07, 2004 1:39 am Post subject: |
|
|
..running our of ideas, man...
what version of gcc are you compiling it with ?
On the ebuild file we have this:
Code: |
# problems with -O3 on gcc-3.3.1
replace-flags -O3 -O2
|
Try using -O2, and maybe also avoiding -funroll-loops and -fomit-frame-pointer.
Have you also tried emerging the binary version and checking if the problem still occurs ? _________________ In God we trust. All the others must provide a valid X.509 certificate |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Tue Sep 07, 2004 1:53 am Post subject: |
|
|
I'll try that... though I noticed -O2 appeared when gcc was called anyways.
How do I grab the binary instead? |
|
Back to top |
|
|
nightblade Guru
Joined: 20 Jul 2004 Posts: 368 Location: back from SE Asia
|
Posted: Tue Sep 07, 2004 2:06 am Post subject: |
|
|
bhogg wrote: |
How do I grab the binary instead? |
the option "--usepkg" of emerge should do the trick _________________ In God we trust. All the others must provide a valid X.509 certificate |
|
Back to top |
|
|
|