| View previous topic :: View next topic |
| Author |
Message |
iandoug
Guru
Joined: 11 Feb 2005
Posts: 566
Location: Cape Town, South Africa
|
 Posted: Sun Jan 15, 2017 9:38 am Post subject: Firefox 45.6 [Kinda SOLVED -- WorkAround] |
 |
|
Hi all
Add me to the list of people with problems with Firefox.
IIRC the order of events was something like this:
1. restarted Firefox
2. could not access any Google sites. Other sites tested worked fine.
3. emerge -auvDN world
4. Firefox borked.
When started from icon, it launches and hangs.
When run from commandline as firefox --safe-mode it launches. I disabled all extentions and themes. Took a few attempts, as it crashed/hung while doing this.
Running as root (ie different profile) it appears towork okay, but no extentions or themes loaded in this profile.
After "cleanup" above it can run from commandline but hangs and is generally unusable.
Also complains about crashes in files in portage, which I find a bit odd, but I'm not a dev.
| Code: |
ian@trooper ~> firefox
Vector smash protection is enabled.
[NPAPI 2154] ###!!! ABORT: Aborting on channel error.: file /var/tmp/portage/www-client/firefox-45.6.0-r1/work/firefox-45.6.0esr/ipc/glue/MessageChannel.cpp, line 1861
[NPAPI 2154] ###!!! ABORT: Aborting on channel error.: file /var/tmp/portage/www-client/firefox-45.6.0-r1/work/firefox-45.6.0esr/ipc/glue/MessageChannel.cpp, line 1861
fish: “firefox” terminated by signal SIGKILL (Forced quit)
|
Can some kind soul please advise how to get back to normal behaviour? 
I rely on Firefox daily... also has a lot of username/passwords saved...
Thanks, Ian
_________________
Asus Sabertooth P990, AMD FX-8150, GeForce GTX 560, 16GB Ram | Asus M3A78 64, X2 6000+, PX9800 GT, 4GB Ram | Asus M4A77TD PRO, X2 245, HD4350, 4GB RAM
Last edited by iandoug on Sun Jan 15, 2017 2:59 pm; edited 1 time in total |
|
| Back to top |
|
 |
Anon-E-moose
Advocate
Joined: 23 May 2008
Posts: 4398
Location: Dallas area
|
| Posted: Sun Jan 15, 2017 10:24 am Post subject: |
|
|
https://forums.gentoo.org/viewtopic-t-1057470.html
_________________
Asus m5a99fx, FX 8320 - nouveau, oss4, rx550 for qemu passthrough
Acer laptop E5-575, i3-7100u - i965, alsa
---both---
5.0.13 zen kernel, profile 17.1 (no-pie & modified) amd64-no-multilib
gcc 8.2.0, eudev, openrc, openbox, palemoon |
|
| Back to top |
|
|
iandoug
Guru
Joined: 11 Feb 2005
Posts: 566
Location: Cape Town, South Africa
|
| Posted: Sun Jan 15, 2017 10:38 am Post subject: |
|
|
| Anon-E-moose wrote: | | https://forums.gentoo.org/viewtopic-t-1057470.html |
Thanks, I think I read that post. Google actually works at the moment, so I don't think it's THAT problem. I am also on 45-r1.
When I start it it's plain vanilla, I even have to add the Menu bar. (File //etc).
If I select File -> exit it fails to exit properly, looking at running processes at this point (ctrl-esc) shows Firefox chewing 12 or 13% CPU.
When I start up from command line I get dialog to start in safe mode or "refresh" but I found out the hard way on another box that "refresh" means "wipe all your stuff" which I can NOT afford to do on this installation.
| Code: |
~ $ firefox
Vector smash protection is enabled.
** (firefox:9180): WARNING **: Invalid borders specified for theme pixmap:
/usr/share/themes/Breeze/gtk-2.0/../assets/line-h.png,
borders don't fit within the image
[NPAPI 9249] ###!!! ABORT: Aborting on channel error.: file /var/tmp/portage/www-client/firefox-45.6.0-r1/work/firefox-45.6.0esr/ipc/glue/MessageChannel.cpp, line 1861
[NPAPI 9249] ###!!! ABORT: Aborting on channel error.: file /var/tmp/portage/www-client/firefox-45.6.0-r1/work/firefox-45.6.0esr/ipc/glue/MessageChannel.cpp, line 1861
Segmentation fault
|
I did consider downgrading FF but there are no older versions in portage ATM ...
Thanks, Ian
_________________
Asus Sabertooth P990, AMD FX-8150, GeForce GTX 560, 16GB Ram | Asus M3A78 64, X2 6000+, PX9800 GT, 4GB Ram | Asus M4A77TD PRO, X2 245, HD4350, 4GB RAM |
|
| Back to top |
|
|
Anon-E-moose
Advocate
Joined: 23 May 2008
Posts: 4398
Location: Dallas area
|
| Posted: Sun Jan 15, 2017 12:07 pm Post subject: |
|
|
I keep copies of old ebuild and the associated files, so I can always back down.
I had made a copy of 45.6.0 just before it was removed.
As far as errors, I don't see anything from what I'm using, but I didn't upgrade to the r1 version
and I kept the older versions of nss/nspr active and newer ones masked.
So I can't help with your problems. Have you checked bugs
I have a tendency to keep firefox stuck at one version for a while, before I updated to 45.6 I was using 38.8 for a while.
Once I find something that works stably I like to stick with it.
Edit to add: It looks like some of the problems might be conflicts with the theme you're using.
_________________
Asus m5a99fx, FX 8320 - nouveau, oss4, rx550 for qemu passthrough
Acer laptop E5-575, i3-7100u - i965, alsa
---both---
5.0.13 zen kernel, profile 17.1 (no-pie & modified) amd64-no-multilib
gcc 8.2.0, eudev, openrc, openbox, palemoon |
|
| Back to top |
|
|
iandoug
Guru
Joined: 11 Feb 2005
Posts: 566
Location: Cape Town, South Africa
|
| Posted: Sun Jan 15, 2017 2:58 pm Post subject: |
|
|
Running from the command line into Profile Select, and creating and using a new profile allows FF to run.
So I guess there's something somewhere in the old profile that the new version does not like.
Now to figure out how to transfer relevant bits from old profile to new profile without borking it.
FF generates a multitude of these messages :
| Code: |
(firefox:13686): dconf-CRITICAL **: unable to create file '/var/run/user/1000/dconf/user': Permission denied. dconf will not work properly.
|
Thanks, Ian
_________________
Asus Sabertooth P990, AMD FX-8150, GeForce GTX 560, 16GB Ram | Asus M3A78 64, X2 6000+, PX9800 GT, 4GB Ram | Asus M4A77TD PRO, X2 245, HD4350, 4GB RAM |
|
| Back to top |
|
|
Hu
Moderator
Joined: 06 Mar 2007
Posts: 14956
|
| Posted: Sun Jan 15, 2017 5:08 pm Post subject: Re: Firefox 45.6 [Kinda SOLVED -- WorkAround] |
|
|
| iandoug wrote: | Also complains about crashes in files in portage, which I find a bit odd, but I'm not a dev.
| Code: | ian@trooper ~> firefox
Vector smash protection is enabled.
[NPAPI 2154] ###!!! ABORT: Aborting on channel error.: file /var/tmp/portage/www-client/firefox-45.6.0-r1/work/firefox-45.6.0esr/ipc/glue/MessageChannel.cpp, line 1861
[NPAPI 2154] ###!!! ABORT: Aborting on channel error.: file /var/tmp/portage/www-client/firefox-45.6.0-r1/work/firefox-45.6.0esr/ipc/glue/MessageChannel.cpp, line 1861
fish: “firefox” terminated by signal SIGKILL (Forced quit)
|
|
Odd, but not unusual. That looks like a combination of two factors. First, someone used __FILE__ in their source file to get a string constant referring to the file name. This is fairly common for debug messages, since it lets the programmer find the right file and keeps the string accurate if the code is moved to a new file. Second, someone wrote the build system to pass absolute paths to gcc. This is a bit less common, and personally I think that is bad practice because it produces uglier messages when handling __FILE__. Many build systems do it though, so my opinion is not universal and maybe not even in the majority. You can see the effect like so: | Code: | cd /tmp
echo __FILE__ > file.cpp
gcc -E file.cpp
gcc -E $PWD/file.cpp
cd /
gcc -E tmp/file.cpp
gcc -E tmp/../tmp/file.cpp |
For me, the output is: | Code: | /tmp$ echo __FILE__ > file.cpp
/tmp$ gcc -E file.cpp
# 1 "file.cpp"
# 1 "<built-in>"
# 1 "<command-line>"
# 1 "/usr/include/stdc-predef.h" 1 3 4
# 1 "<command-line>" 2
# 1 "file.cpp"
"file.cpp"
/tmp$ gcc -E $PWD/file.cpp
# 1 "/tmp/file.cpp"
# 1 "<built-in>"
# 1 "<command-line>"
# 1 "/usr/include/stdc-predef.h" 1 3 4
# 1 "<command-line>" 2
# 1 "/tmp/file.cpp"
"/tmp/file.cpp"
/tmp$ cd /
/$ gcc -E tmp/file.cpp
# 1 "tmp/file.cpp"
# 1 "<built-in>"
# 1 "<command-line>"
# 1 "/usr/include/stdc-predef.h" 1 3 4
# 1 "<command-line>" 2
# 1 "tmp/file.cpp"
"tmp/file.cpp"
/$ gcc -E tmp/../tmp/file.cpp
# 1 "tmp/../tmp/file.cpp"
# 1 "<built-in>"
# 1 "<command-line>"
# 1 "/usr/include/stdc-predef.h" 1 3 4
# 1 "<command-line>" 2
# 1 "tmp/../tmp/file.cpp"
"tmp/../tmp/file.cpp"
|
Observe how, in each case, the expansion of __FILE__ is exactly the argument I passed to gcc. It is not canonicalized in any way.
The messages you quoted would use a different path if you had changed $PORTAGE_TMPDIR. Their use of that particular path is harmless. The error that Firefox is trying to report by using that path may or may not be harmless. |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Desktop Environments
