Firefox on XP - frequent hangs/crashes

Is this just me or does everyone have problems with Firefox (V42.0 for me)
hanging or crashing regularly? It seems like it happens at least once or twice a
week.

In message , Bob F
writes:
Is this just me or does everyone have problems with Firefox (V42.0 for me)
hanging or crashing regularly? It seems like it happens at least once
or twice a
week.

No; my FF 26 does occasionally crash, but I think more rarely than that
(and nearly always recoverably).

--
J. P. Gilliver. UMRA: 1960/1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

It is dangerous to be sincere, unless you are also stupid. - George Bernard
Shaw

Bob F wrote:
Is this just me or does everyone have problems with Firefox (V42.0 for me)
hanging or crashing regularly? It seems like it happens at least once or twice a
week.

My XP machine uses FF 27. It crashes once every few months.

Bob F wrote:
Is this just me or does everyone have problems with Firefox (V42.0 for me)
hanging or crashing regularly? It seems like it happens at least once or twice a
week.



I'm working on something similar. The details are quite different,
so about all I can do, is point in one particular direction.

The only progress I've made so far, is to do a debug build,
which enables console messages after Firefox starts. And
I see Firefox go into a loop on a particular ASSERT() statement.

When you don't use a debug build, with console output, then
the thing will loop silently and appear hung.

Firefox has "fatal" ASSERTs in the JavaScript engine,
and "warning" ASSERTs in the display rendering code.
And the problem I'm working on is in the display
rendering code. This message prints in a tight loop
to my debug console.

[31718] ###!!! ASSERTION: available space should never grow:
'aOldAvailableSpace.IStart(aWM) = aNewAvailableSpace.IStart(aWM)
&& aOldAvailableSpace.IEnd(aWM) = aNewAvailableSpace.IEnd(aWM)',
file nsBlockFrame.cpp, line 922

There are no debug builds in the tree, just
release ones, so you have to build your own.

http://releases.mozilla.org/pub/fire...leases/43.0.4/

A rough recipe would be:

1) Acquire Visual Studio Express (Community Edition)
2) The Firefox build environment just runs the compiler and
linker from a shell environment. The VS IDE is not used.
Getting Visual Studio Express is just to get compiler and
linker. You might need some SDK for header files, libraries
and so on.
3) Problems include, not enough address space on a 32 bit OS
to complete the build any more. I had to move the kernel/userspace
split, to make 3GB available the last time, and it just barely fit.

Which means some careful consideration has to go into your
build environment, before you start. The purpose of all this,
is so you can report some message the browser spits out,
rather than a generic "it be loopin".

And if you need a link to hang a recent Firefox, you
can try this. This is my test case. For most people,
DON'T CLICK THIS. This is only for people who need to
test their browser, and it's already doing the recovery
thing at the beginning. An older browser will likely
not have a problem.

https://www.links.hr/hr/adapter-poe-...nog-kabla-0535

"Are we having fun yet ?"

Paul

Paul wrote:
Bob F wrote:
Is this just me or does everyone have problems with Firefox (V42.0
for me) hanging or crashing regularly? It seems like it happens at
least once or twice a week.



I'm working on something similar. The details are quite different,
so about all I can do, is point in one particular direction.

The only progress I've made so far, is to do a debug build,
which enables console messages after Firefox starts. And
I see Firefox go into a loop on a particular ASSERT() statement.

When you don't use a debug build, with console output, then
the thing will loop silently and appear hung.

Firefox has "fatal" ASSERTs in the JavaScript engine,
and "warning" ASSERTs in the display rendering code.
And the problem I'm working on is in the display
rendering code. This message prints in a tight loop
to my debug console.

[31718] ###!!! ASSERTION: available space should never grow:
'aOldAvailableSpace.IStart(aWM) =
aNewAvailableSpace.IStart(aWM) &&
aOldAvailableSpace.IEnd(aWM) =
aNewAvailableSpace.IEnd(aWM)', file nsBlockFrame.cpp, line 922
There are no debug builds in the tree, just
release ones, so you have to build your own.

http://releases.mozilla.org/pub/fire...leases/43.0.4/

A rough recipe would be:

1) Acquire Visual Studio Express (Community Edition)
2) The Firefox build environment just runs the compiler and
linker from a shell environment. The VS IDE is not used.
Getting Visual Studio Express is just to get compiler and
linker. You might need some SDK for header files, libraries
and so on.
3) Problems include, not enough address space on a 32 bit OS
to complete the build any more. I had to move the kernel/userspace
split, to make 3GB available the last time, and it just barely fit.

Which means some careful consideration has to go into your
build environment, before you start. The purpose of all this,
is so you can report some message the browser spits out,
rather than a generic "it be loopin".

And if you need a link to hang a recent Firefox, you
can try this. This is my test case. For most people,
DON'T CLICK THIS. This is only for people who need to
test their browser, and it's already doing the recovery
thing at the beginning. An older browser will likely
not have a problem.

https://www.links.hr/hr/adapter-poe-...nog-kabla-0535


Sounds a lot like my problem. Sometimes, the machine will just hang for awhile
(minutes or more), and sometimes it will not recover, and I have to close
firefox and re-open it. Occasionally, the whole computer hangs (no cursor
movement) and I have to restart, but that is much less often.

Bob F wrote on 2016/01/17:

Is this just me or does everyone have problems with Firefox (V42.0 for me)
hanging or crashing regularly? It seems like it happens at least once or twice a
week.

Review what add-ons that you installed to Firefox. Disable them all for
long enough to be sure whether the problem continues or went away. Then
reenable one at a time and retest just that one. Enable only one at a
time, not enable one, enable another, etc. Add-ons can conflict with
each other so just test one at a time.

Hanging can be caused by bad scripts at the sites that you visit.
Disable Javascripting when visiting those pages that cause frequent
hangs in Firefox. Without the scripts, perhaps those pages won't hang
anymore and maybe you really don't need scripting for those pages to
make them usable. Crashes should be recorded in Firefox
(about:crashes).

On Sun, 17 Jan 2016 14:51:11 -0600, VanguardLH wrote:

Bob F wrote on 2016/01/17:

Is this just me or does everyone have problems with Firefox (V42.0 for me)
hanging or crashing regularly? It seems like it happens at least once or twice a
week.

Review what add-ons that you installed to Firefox. Disable them all for
long enough to be sure whether the problem continues or went away. Then
reenable one at a time and retest just that one. Enable only one at a
time, not enable one, enable another, etc. Add-ons can conflict with
each other so just test one at a time.

Hanging can be caused by bad scripts at the sites that you visit.
Disable Javascripting when visiting those pages that cause frequent
hangs in Firefox. Without the scripts, perhaps those pages won't hang
anymore and maybe you really don't need scripting for those pages to
make them usable. Crashes should be recorded in Firefox
(about:crashes).

Every time I have a freeze up I believe it was a scripting bug. I have
suspended Flash (get the dialog box) and I really don't miss it if I
am not intentionally streaming video.

VanguardLH wrote:
Bob F wrote on 2016/01/17:

Is this just me or does everyone have problems with Firefox (V42.0
for me) hanging or crashing regularly? It seems like it happens at
least once or twice a week.

Review what add-ons that you installed to Firefox. Disable them all
for long enough to be sure whether the problem continues or went
away. Then reenable one at a time and retest just that one. Enable
only one at a time, not enable one, enable another, etc. Add-ons can
conflict with each other so just test one at a time.

Hanging can be caused by bad scripts at the sites that you visit.
Disable Javascripting when visiting those pages that cause frequent
hangs in Firefox. Without the scripts, perhaps those pages won't hang
anymore and maybe you really don't need scripting for those pages to
make them usable. Crashes should be recorded in Firefox
(about:crashes).\

No add-ons, and flash has to ask me to run.

In message , VanguardLH
writes:
[]
Review what add-ons that you installed to Firefox. Disable them all for
long enough to be sure whether the problem continues or went away. Then
reenable one at a time and retest just that one. Enable only one at a
time, not enable one, enable another, etc. Add-ons can conflict with
each other so just test one at a time.
[]
No, enable half of them, then (if no failure) half of what's left or (if
failure) half of those you enabled, and so on. OK, because of
interactions, you _may_ have to retrace and try combinations, but a
binary search is a lot quicker than one-at-a-time ... (-:
--
J. P. Gilliver. UMRA: 1960/1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

Insane people are crazy, not stupid. - kenckar, 2015-7-23