I'm running Debian/Trixie on an AMD64 system, using the Plasma 5 over X desktop. Firefox 115.12.0esr is crashing multiple times per day. It frequently happens when page I'm transfers to another page that creates a > PDF or just has a complicated link. It's annoying.
I also have found that at least one site refuses to work with 115. That's been going on for a while. Again, I have to use Chromium for that site.
On Wed, 2024-07-17 at 22:17 -0400, eben@gmx.us wrote:
On 7/17/24 21:25, Gary Dale wrote:
I'm running Debian/Trixie on an AMD64 system, using the Plasma 5
over X
desktop. Firefox 115.12.0esr is crashing multiple times per day. It frequently happens when page I'm transfers to another page that
creates a > PDF or just has a complicated link. It's annoying.
I upgraded from Debian 10 to Debian 12.5 "Bookworm." The NVidia 390
driver no longer works, so I had software rendering because nouveau apparently can't do GPU rendering. Rather than crashing, the system essentially froze. After waiting for a VERY long time, I would give up
and cycle power, with my reboot set up to start an empty session, not
the one I had going at time of the power cycle. I replaced the graphics
card with a Quadro K2200, which works with the nvidia-drivers package
that's still part of the Debian 12.5 distro. With GPU rendering, I no
longer have the problem.
I can't do that with my old Dell Vostro 1700 because the NVidia
graphics chip is soldered to the motherboard, and it needs the 340
driver, which is also no longer available.
I tried several of the methods discussed in this thread to get the
drivers working, but had no success. Maybe I was just holding my mouth
wrong.
I have Firefox 115.13.0esr and it rarely crashes for me, and I have
dozens
of tabs open. I use straight XFCE, no Plasma. Could be it doesn't
do PDFs
well? I use Zathura to view PDFs. It's rather ... "feature free",
so I may
change.
Can you say what that site is?
--
An idea that is not dangerous is unworthy of
being called an idea at all. -- Oscar Wilde
I'm running Debian/Trixie on an AMD64 system, using the Plasma 5 over
X desktop. Firefox 115.12.0esr is crashing multiple times per day. It frequently happens when page I'm transfers to another page that
creates a PDF or just has a complicated link. It's annoying.
To visit some pages, I have to use Chromium instead. Earlier today I
had to rename a sessionstore-backups json file because Firefox got
caught in loop where it recognized it had a new tab open but the tab
caused it to crash.
I also have found that at least one site refuses to work with 115.
That's been going on for a while. Again, I have to use Chromium for
that site.
I'm running Debian/Trixie on an AMD64 system, using the Plasma 5 over
X desktop. Firefox 115.12.0esr is crashing multiple times per day. It frequently happens when page I'm transfers to another page that
creates a PDF or just has a complicated link. It's annoying.
To visit some pages, I have to use Chromium instead. Earlier today I
had to rename a sessionstore-backups json file because Firefox got
caught in loop where it recognized it had a new tab open but the tab
caused it to crash.
On 2024-07-17 21:25, Gary Dale wrote:
I'm running Debian/Trixie on an AMD64 system, using the Plasma 5 overThanks for the tips guys, but I'm not going to switch to XFCE, I'm
X desktop. Firefox 115.12.0esr is crashing multiple times per day. It
frequently happens when page I'm transfers to another page that
creates a PDF or just has a complicated link. It's annoying.
To visit some pages, I have to use Chromium instead. Earlier today I
had to rename a sessionstore-backups json file because Firefox got
caught in loop where it recognized it had a new tab open but the tab
caused it to crash.
I also have found that at least one site refuses to work with 115.
That's been going on for a while. Again, I have to use Chromium for
that site.
using an old AMD graphics card, it's a desktop machine, and the
problem isn't specific to PDFs - although that seems to be one of the
major triggers.
My system has been upgrading from earlier versions of Debian since
Potato. I've been on Trixie since it became the new testing. This
crashing of Firefox is a new issue - had few problems with Trixie
before that.
I'm beginning to suspect it may be related to my recent introduction
of a Pi-Hole into my network. Could it be a problem for Firefox when
it gets a 0.0.0.0 address returned on a DNS lookup?
On 2024-07-17 21:25, Gary Dale wrote:
I'm running Debian/Trixie on an AMD64 system, using the Plasma 5 overThanks for the tips guys, but I'm not going to switch to XFCE, I'm
X desktop. Firefox 115.12.0esr is crashing multiple times per day. It
frequently happens when page I'm transfers to another page that
creates a PDF or just has a complicated link. It's annoying.
To visit some pages, I have to use Chromium instead. Earlier today I
had to rename a sessionstore-backups json file because Firefox got
caught in loop where it recognized it had a new tab open but the tab
caused it to crash.
I also have found that at least one site refuses to work with 115.
That's been going on for a while. Again, I have to use Chromium for
that site.
using an old AMD graphics card, it's a desktop machine, and the
problem isn't specific to PDFs - although that seems to be one of the
major triggers.
My system has been upgrading from earlier versions of Debian since
Potato. I've been on Trixie since it became the new testing. This
crashing of Firefox is a new issue - had few problems with Trixie
before that.
I'm beginning to suspect it may be related to my recent introduction
of a Pi-Hole into my network. Could it be a problem for Firefox when
it gets a 0.0.0.0 address returned on a DNS lookup?
On 2024-07-18 09:52, Gary Dale wrote:
Thanks for the tips guys, but I'm not going to switch to XFCE, I'm
using an old AMD graphics card, it's a desktop machine, and the
problem isn't specific to PDFs - although that seems to be one of
the major triggers.
My system has been upgrading from earlier versions of Debian since
Potato. I've been on Trixie since it became the new testing. This
crashing of Firefox is a new issue - had few problems with Trixie
before that.
I'm beginning to suspect it may be related to my recent
introduction of a Pi-Hole into my network. Could it be a problem
for Firefox when it gets a 0.0.0.0 address returned on a DNS
lookup?
Well, I can confirm it's not the Pi-Hole. Took it out of the DNS
chain and Firefox is still crashing frequently. In fact, it's worse
today. Now it crashes when I'm on Facebook and scrolling down using
the mouse wheel.
On 2024-07-19 at 10:34, Gary Dale wrote:
On 2024-07-18 09:52, Gary Dale wrote:What kind of crashes are we talking about? I think there may be an 'about:crashes' or similar type of page built in to Firefox, which could
Thanks for the tips guys, but I'm not going to switch to XFCE, I'mWell, I can confirm it's not the Pi-Hole. Took it out of the DNS
using an old AMD graphics card, it's a desktop machine, and the
problem isn't specific to PDFs - although that seems to be one of
the major triggers.
My system has been upgrading from earlier versions of Debian since
Potato. I've been on Trixie since it became the new testing. This
crashing of Firefox is a new issue - had few problems with Trixie
before that.
I'm beginning to suspect it may be related to my recent
introduction of a Pi-Hole into my network. Could it be a problem
for Firefox when it gets a 0.0.0.0 address returned on a DNS
lookup?
chain and Firefox is still crashing frequently. In fact, it's worse
today. Now it crashes when I'm on Facebook and scrolling down using
the mouse wheel.
give information about what it's seen happen.
If it's memory-access-related, there might be benefit to trying to e.g.
run under valgrind, intentionally reproduce a crash, and see what that
tool reports. Then again, Firefox is a sufficiently complex app that
that might not be fruitful.
Have you tried running any hardware-error checking tools, e.g. one of
the memtest suites? Crashes that frequent (with software, versions, and
data which other people do not reproduce the problem with) suggest a
possible hardware issue to my mind, although if nothing else is
exhibiting visible issues that makes the hardware a less likely culprit.
Is there any possibility that something in the library/etc. stack which Firefox sits on top of may be unreliable, or at least be of different versions from those which the people not observing the problem are
using? One obvious candidate would probably be the graphics stack
(driver, firmware, etc.), but that's not necessarily the only possibility.
On 2024-07-19 10:42, The Wanderer wrote:
On 2024-07-19 at 10:34, Gary Dale wrote:Looking at the submitted and unsubmitted reports, it seems the
On 2024-07-18 09:52, Gary Dale wrote:What kind of crashes are we talking about? I think there may be an
Thanks for the tips guys, but I'm not going to switch to XFCE, I'mWell, I can confirm it's not the Pi-Hole. Took it out of the DNS
using an old AMD graphics card, it's a desktop machine, and the
problem isn't specific to PDFs - although that seems to be one of
the major triggers.
My system has been upgrading from earlier versions of Debian since
Potato. I've been on Trixie since it became the new testing. This
crashing of Firefox is a new issue - had few problems with Trixie
before that.
I'm beginning to suspect it may be related to my recent
introduction of a Pi-Hole into my network. Could it be a problem
for Firefox when it gets a 0.0.0.0 address returned on a DNS
lookup?
chain and Firefox is still crashing frequently. In fact, it's worse
today. Now it crashes when I'm on Facebook and scrolling down using
the mouse wheel.
'about:crashes' or similar type of page built in to Firefox, which could
give information about what it's seen happen.
If it's memory-access-related, there might be benefit to trying to e.g.
run under valgrind, intentionally reproduce a crash, and see what that
tool reports. Then again, Firefox is a sufficiently complex app that
that might not be fruitful.
Have you tried running any hardware-error checking tools, e.g. one of
the memtest suites? Crashes that frequent (with software, versions, and
data which other people do not reproduce the problem with) suggest a
possible hardware issue to my mind, although if nothing else is
exhibiting visible issues that makes the hardware a less likely culprit.
Is there any possibility that something in the library/etc. stack which
Firefox sits on top of may be unreliable, or at least be of different
versions from those which the people not observing the problem are
using? One obvious candidate would probably be the graphics stack
(driver, firmware, etc.), but that's not necessarily the only
possibility.
crashing started on July 10. It always seems to be "CanvasRenderer" as
the culprit with libxul.so as the guilty module. Firefox was
reportedly installed 32 days ago.
Anyway, the Firefox developers have received dozens of automated crash reports from me over the past10 days.
I do have a rather old graphics card, but it's an AMD one so the
drivers should be OK. I doubt it's anything else hardware related as
I'm not having problems with other programs.
The only other things of note:
1) my screen does briefly go blank sometimes while doing something
involving windows.
2) Plasma 5 isn't saving my desktop when I reboot.
On 2024-07-19 11:09, Gary Dale wrote:
On 2024-07-19 10:42, The Wanderer wrote:It's not X11 either. It's happening when I use Plasma 5 on Wayland.
On 2024-07-19 at 10:34, Gary Dale wrote:Looking at the submitted and unsubmitted reports, it seems the
On 2024-07-18 09:52, Gary Dale wrote:What kind of crashes are we talking about? I think there may be an
Thanks for the tips guys, but I'm not going to switch to XFCE, I'mWell, I can confirm it's not the Pi-Hole. Took it out of the DNS
using an old AMD graphics card, it's a desktop machine, and the
problem isn't specific to PDFs - although that seems to be one of
the major triggers.
My system has been upgrading from earlier versions of Debian since
Potato. I've been on Trixie since it became the new testing. This
crashing of Firefox is a new issue - had few problems with Trixie
before that.
I'm beginning to suspect it may be related to my recent
introduction of a Pi-Hole into my network. Could it be a problem
for Firefox when it gets a 0.0.0.0 address returned on a DNS
lookup?
chain and Firefox is still crashing frequently. In fact, it's worse
today. Now it crashes when I'm on Facebook and scrolling down using
the mouse wheel.
'about:crashes' or similar type of page built in to Firefox, which
could
give information about what it's seen happen.
If it's memory-access-related, there might be benefit to trying to e.g.
run under valgrind, intentionally reproduce a crash, and see what that
tool reports. Then again, Firefox is a sufficiently complex app that
that might not be fruitful.
Have you tried running any hardware-error checking tools, e.g. one of
the memtest suites? Crashes that frequent (with software, versions, and
data which other people do not reproduce the problem with) suggest a
possible hardware issue to my mind, although if nothing else is
exhibiting visible issues that makes the hardware a less likely
culprit.
Is there any possibility that something in the library/etc. stack which
Firefox sits on top of may be unreliable, or at least be of different
versions from those which the people not observing the problem are
using? One obvious candidate would probably be the graphics stack
(driver, firmware, etc.), but that's not necessarily the only
possibility.
crashing started on July 10. It always seems to be "CanvasRenderer"
as the culprit with libxul.so as the guilty module. Firefox was
reportedly installed 32 days ago.
Anyway, the Firefox developers have received dozens of automated
crash reports from me over the past10 days.
I do have a rather old graphics card, but it's an AMD one so the
drivers should be OK. I doubt it's anything else hardware related as
I'm not having problems with other programs.
The only other things of note:
1) my screen does briefly go blank sometimes while doing something
involving windows.
2) Plasma 5 isn't saving my desktop when I reboot.
Firefox ESR is still crashing intermittently. Again, I can trigger it fairly consistently just by visiting some pages (usually ones that try to generate a PDF, for example). At other times it just crashes for no apparent reason.
I'm running Debian/Trixie on an AMD64 system, using the Plasma 5 over X desktop. Firefox 115.12.0esr is crashing multiple times per day. It frequently happens when page I'm transfers to another page that creates
a PDF or just has a complicated link. It's annoying.
To visit some pages, I have to use Chromium instead. Earlier today I
had to rename a sessionstore-backups json file because Firefox got
caught in loop where it recognized it had a new tab open but the tab
caused it to crash.
I also have found that at least one site refuses to work with 115.
That's been going on for a while. Again, I have to use Chromium for that site.
On Sat, Aug 17, 2024 at 9:07 PM Gary Dale <gary@extremeground.com> wrote:
[...]This thread is kind of long, and I did not read through all the responses.
Out of frustration with this and another problem, I did a complete fresh
install yesterday - first to Bookworm then a full-upgrade to Trixie. I
started with a new profile for Firefox then synced it to restore my
passwords and bookmarks.
Firefox ESR is still crashing intermittently. Again, I can trigger it
fairly consistently just by visiting some pages (usually ones that try
to generate a PDF, for example). At other times it just crashes for no
apparent reason.
Did you disable hardware acceleration and observe the crashes? See <https://support.mozilla.org/en-US/kb/upgrade-graphics-drivers-use-hardware-acceleration#w_turning-off-hardware-acceleration>.
Jeff
Sysop: | Keyop |
---|---|
Location: | Huddersfield, West Yorkshire, UK |
Users: | 546 |
Nodes: | 16 (0 / 16) |
Uptime: | 165:53:17 |
Calls: | 10,385 |
Calls today: | 2 |
Files: | 14,057 |
Messages: | 6,416,528 |