FireFox 24.0,1

rtwingfield

rtwingfield

I have installed FireFox v24.0.1 from the ports. After starting the application (in an X11 window), it does start but is unstable, crashes almost immediately and issues the following diagnostic message:
Code: 
# firefox

(process:1754): GLib-CRITICAL **: g_slice_set_config: assertion `sys_page_size == 0' failed
Segmentation fault (core dumped)
#

. . .what to do?
 
rtwingfield said:
Code: 
(process:1754): GLib-CRITICAL **: g_slice_set_config: assertion `sys_page_size == 0' failed
Click to expand...

I get this messages on all FreeBSD boxes I have Firefox installed on and it runs fine.

Is this system running 9-STABLE i386 with the source synced and built from a few months ago? I'm neither sure if the i386 is relevant nor the specific dates when the problems were introduced, but I had segmentation faults as soon as I browsed to anything other than very simple pages. Unfortunately, the only solution I found was updating the OS to a more recent STABLE. You could also use freebsd-update() to update to 9.2-RC4 and then 9.2-RELEASE when it's out.
 
Overnight, I upgraded from FreeBSD v9.0 to 9.1-RELEASE-p7. Unfortunately, the problem persists.

From the /usr/ports/www/firefox/pkg-message instructions, the maintainers request the following:
Code: 
. . .bug reports should be addressed to the maintainers at:
        gecko@FreeBSD.org
You may also Cc: freebsd-ports@FreeBSD.org.

Please include the following information with any bug report:
* Output from 'uname -a'.
* Output from 'ident /usr/ports/www/firefox/Makefile'
* Where/when did the problem occur: configuring, building, or
        running firefox
* How can you reproduce the problem?

. . .to wit as follows:
Code: 
[CMD]# uname -a[/CMD]
FreeBSD arwingcap.org 9.1-RELEASE-p7 FreeBSD 9.1-RELEASE-p7 #0: Mon Sep  9 21:23
:37 UTC 2013     [email]root@amd64-builder.daemonology.net[/email]:/usr/obj/usr/src/sys/GENERIC
  i386

[CMD]# ident /usr/ports/www/firefox/Makefile[/CMD]
/usr/ports/www/firefox/Makefile:
     $FreeBSD: www/firefox/Makefile 327776 2013-09-20 23:36:50Z bapt $

The problem is [B]repeatable[/B] and occurs when starting FireFox in an Xwindow.  
Before the FireFox Xwindow opens, the parent Xwindow displays the GLib-CRITICAL message 
on the command line, then the first panel of FireFox will display in a new child Xwindow
and apparently remains stable, but any attempt to browse to another webpage causes a
segmentation fault followed by a core dump.

QUESTION: I'm running this on an older Compaq Proliant ML370 as described below:
. . .from dmesg:
Code: 
CPU: Intel Pentium III (598.12-MHz 686-class CPU)
  Origin = "GenuineIntel"  Id = 0x683  Family = 6  Model = 8  Stepping = 3
  Features=0x383fbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CM
OV,PAT,PSE36,MMX,FXSR,SSE>
real memory  = 4294967296 (4096 MB)
avail memory = 4031164416 (3844 MB)
Event timer "LAPIC" quality 400
ACPI APIC Table: <COMPAQ 00000083>
FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs
FreeBSD/SMP: 2 package(s) x 1 core(s)
 cpu0 (BSP): APIC ID:  1
 cpu1 (AP): APIC ID:  0

I've upgraded the SDRAM-PC133 CL3 ECC compliment to 4GB. With the additional RAM, problem still persists. I should think that 4GB would be enough with little else running . . .?

I'm trying to find minimum RAM requirements for FireFox and a GB is what I'm finding as suggested from the searches. It would be interesting to know how much RAM is installed regarding anyone successfully running FF 24.0.1. Suggestions will be appreciated!
 
No. A gigabyte is not a minimum. I have machines running FF 24 that would never be able to do anything else if they needed to give up a gigabyte to Firefox.

As I was typing I just noticed you're running a Pentium III! There is your problem. I'd have to look it up but I don't think ANY major browser supports the Pentium III.

EDIT: Firefox24 system requirements
They "recommend" a Pentium4 as a minimum. Chromium says Pentium III is a minimum requirement.

I retired two boxes last year that were Pentium IIIs that ran at 550 MHz and had 192 MB of RAM. They ran Firefox 4 and Chrome but were very sluggish.
 
FireFox 24.0.1 Compile Failure

After an exhaustive 9-day-week (over one hundred hours) invested in upgrading all ports and wrestling with the differences between the new pgk2ng vs pkg_* scenarios, I have once again tried to compile FireFox 24.0.1 and the compile crashes with the following as of 30 DEC 2013:

Code: 
Stack dump:
0.      Program arguments: /usr/bin/clang++ -cc1 -triple i386-unknown-freebsd9.0 -emit-obj -disable-free -main-file-name SharedMemory.cpp -pic-level 2 -mdisable-fp-elim -relaxed-aliasing -masm-verbose -mconstructor-aliases -target-cpu i486 -momit-leaf-frame-pointer -g -ffunction-sections -fdata-sections -coverage-file SharedMemory.o -resource-dir /usr/bin/../lib/clang/3.1 -dependency-file .deps/SharedMemory.o.pp -MT SharedMemory.o -sys-header-deps -MP -include ../../../config/gcc_hidden.h -isystem /usr/local/include -isystem /usr/local/include -include ../../mozilla-config.h -D MOZ_CHILD_PROCESS_NAME="plugin-container" -D MOZ_CHILD_PROCESS_BUNDLE="plugin-container.app/Contents/MacOS/" -D MOZ_GLUE_IN_PROGRAM -D MOZILLA_INTERNAL_API -D _IMPL_NS_COM -D EXPORT_XPT_API -D EXPORT_XPTC_API -D _IMPL_NS_GFX -D _IMPL_NS_WIDGET -D IMPL_XREAPI -D IMPL_NS_NET -D IMPL_THEBES -D STATIC_EXPORTABLE_JS_API -D NO_NSPR_10_SUPPORT -D OS_POSIX=1 -D OS_FREEBSD=1 -D OS_BSD=1 -D DEBUG -D _DEBUG -D TRACING -D MOZILLA_CLIENT -I ../../dist/stl_wrappers -I ../../dist/system_wrappers -I ../../../toolkit/crashreporter -I ../../../ipc/chromium/src -I ../../../ipc/glue -I ../../ipc/ipdl/_ipdlheaders -I ../../../ipc/glue -I . -I ../../dist/include -I /usr/local/include/nspr -I /usr/local/include/nss -I /usr/local/include/nss/nss -I /usr/local/include -I /usr/local/include -I /usr/local/include -I /usr/local/include -fmodule-cache-path /tmp/clang-module-cache -O2 -Wall -Wpointer-arith -Woverloaded-virtual -Werror=return-type -Wtype-limits -Wempty-body -Wsign-compare -Wno-invalid-offsetof -Wno-c++0x-extensions -Wno-extended-offsetof -Wno-unknown-warning-option -Wno-return-type-c-linkage -Wno-mismatched-tags -std=gnu++0x -fdeprecated-macro -fdebug-compilation-dir /usr/ports/www/firefox/work/mozilla-esr24/obj-i386-portbld-freebsd9.1 -ferror-limit 19 -fmessage-length 97 -mstackrealign -fno-rtti -fgnu-runtime -fobjc-runtime-has-arc -fobjc-runtime-has-weak -fobjc-dispatch-method=non-legacy -fdiagnostics-show-option -fcolor-diagnostics -o SharedMemory.o -x c++ /usr/ports/www/firefox/work/mozilla-esr24/ipc/glue/SharedMemory.cpp
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'Function Pass Manager' on module '/usr/ports/www/firefox/work/mozilla-esr24/ipc/glue/SharedMemory.cpp'.
[B]4.      Running pass 'X86 DAG->DAG Instruction Selection' on function '@_ZN7mozilla3ipc12SharedMemory8UnmappedEv'
clang++: error: unable to execute command: Abort trap: 6 (core dumped)[/B]
clang++: error: clang frontend command failed due to signal (use -v to see invocation)
clang++: note: diagnostic msg: Please submit a bug report to http://llvm.org/bugs/ and include command line arguments and all diagnostic information.
clang++: note: diagnostic msg: Preprocessed source(s) and associated run script(s) are located at:
clang++: note: diagnostic msg: /tmp/SharedMemory-cg1Pga.ii
clang++: note: diagnostic msg: /tmp/SharedMemory-cg1Pga.sh
gmake[4]: *** [SharedMemory.o] Error 254
gmake[4]: Leaving directory `/usr/ports/www/firefox/work/mozilla-esr24/obj-i386-portbld-freebsd9.1/ipc/glue'
gmake[3]: *** [libs] Error 2
gmake[3]: Leaving directory `/usr/ports/www/firefox/work/mozilla-esr24/obj-i386-portbld-freebsd9.1/ipc'
gmake[2]: *** [libs_tier_platform] Error 2
gmake[2]: Leaving directory `/usr/ports/www/firefox/work/mozilla-esr24/obj-i386-portbld-freebsd9.1'
gmake[1]: *** [tier_platform] Error 2
gmake[1]: Leaving directory `/usr/ports/www/firefox/work/mozilla-esr24/obj-i386-portbld-freebsd9.1'
gmake: *** [default] Error 2
*** [do-build] Error code 1

Stop in /usr/ports/www/firefox.
*** [build] Error code 1

Stop in /usr/ports/www/firefox.

===>>> make failed for www/firefox
===>>> Aborting update
===>>> Killing background jobs
Terminated
===>>> Installation of accessibility/atk (atk-2.8.0) complete
===>>> You can restart from the point of failure with this command line:
       portmaster <flags> www/firefox
===>>> Exiting

I have no civil words to describe my frustration with this port. The platform is FreeBSD 9.1-RELEASE-p7 #0: Mon Sep 9 21:23:37 UTC 2013 root@amd64-builder.daemonology.net:/usr/obj/usr/src/sys/GENERIC.

The process ran for over four hours before crashing with this complaint:
Code: 
4.      Running pass 'X86 DAG->DAG Instruction Selection' on function '@_ZN7mozilla3ipc12SharedMemory8UnmappedEv'
clang++: error: unable to execute command: Abort trap: 6 (core dumped)
This apparently is a problem with the clang compiler. I've eMail the port maintainer(s) flo@smeets.im and cc'd gecko@FreeBSD.org and freebsd-ports@FreeBSD.org) and have received no followup response to:

I'm getting this message too and firefox does not crash. The crash has
to be something else. Please build firefox with DEBUG, load the core
file in gdb and show us the output.

Just read the thread on the forums, did you rebuild firefox after
upgrading from 9.0 to 9.1? If you didn't please do, there were problems
in 9.0 that could explain this.

Florian
Click to expand...
I did attach clang.out.gz and make-build.out.gz files to my response regarding their request for additional information via eMail.

The problem now is that I cannot complete the compile(s) on FreeBSD v9.1; therefore, I cannot produce a core dump.

What to do?
 
The new Firefox version has hit the ports tree and it needs using gcc instead of clang in x86 platforms. If I were you I would try updating my ports tree and installing this new version. It works perfectly here.
 
Regarding the suggestion:
Juanitou said:
The new Firefox version has hit the ports tree and it needs using gcc instead of clang in x86 platforms. If I were you I would try updating my ports tree and installing this new version. It works perfectly here.
Click to expand...
1) how would one know this? 2) I've just invested over 100 man-hours
FireFox 24.0.1 Compile Failure
After an exhaustive 9-day-week (over one hundred hours) invested in upgrading all ports and wrestling with the differences between the new pgk2ng vs pkg_* scenarios
Click to expand...
updating the ports tree, learning the nuances of pkg2ng vs. pkg_*, etc. and I know that the tree is now up to date; still, the FireFox 24.0.1 compile crashes.

How would one force the use of the gcc compiler vs. clang?
 
RE:
morbit said:
Do you have sound in HTML5 content in Firefox 25?
Click to expand...
. . .what do you mean by sound? I have have the alsa-plugins port installed.

Also, you refer to Firefox 25 -- I'm not aware of Firefox 25.

Code: 
# [CMD]portsnap update[/CMD]
Ports tree is already up to date.


/usr/ports/www/firefox/distinfo contains the following:
Code: 
SHA256 (firefox-24.0esr.source.tar.bz2) = 3872f7c4c7b3f1f3f06dd3e4f7cc97b59a6857085af8f2df56820185e3d5cadb
SIZE (firefox-24.0esr.source.tar.bz2) = 119722621
 
rtwingfield said:
How would one force the use of the gcc compiler vs. clang?
Click to expand...

Try # cd /usr/ports/www/firefox/ && make USE_GCC=any install clean
 
www/firefox was updated to 25; unfortunately for me it's now mute on soundcloud (mp3), bandcamp (mp3), youtube (webm) etc.
 
rtwingfield said:
Regarding the suggestion:

1) how would one know this?
Click to expand...

It happened just yesterday, good timing.

rtwingfield said:
RE:
Code: 
# [CMD]portsnap update[/CMD]
Ports tree is already up to date.
Click to expand...
I think you missed the fetch part: portsnap fetch update
 
morbit said:
www/firefox was updated to 25; unfortunately for me it's now mute on soundcloud (mp3), bandcamp (mp3), youtube (webm) etc.
Click to expand...
It works here. :\ Maybe check build options with make showconfig, there has been some changes in the multimedia defaults as seen in the ports history, but all this is way over my head.
 
RE: gcc compiler:
cpm said:
Try # cd /usr/ports/www/firefox/ && make USE_GCC=any install clean
Click to expand...
. . .inconsideration that FireFox 25 has "hit the market", I've decided to abandon the FireFox 24 installation.

Interesting that FireFox-25.0,1 compile now "asks" for gcc46, and was not found. In the process of attempting the compile of FireFox-25.0,1, the make process recognized that
Code: 
firefox-25.0,1 depends on executable: gcc46 - not found
and launched into a compile of the gcc-4.6.3 compiler . . .and it has been running for six, going on sever hours now, with no indication that it is anywhere near done x(

Interesting to note that the FreeBSD Developers' Handbook, Chapter 2. Programming Tools, 2.4. Compiling with cc, Revision: 43074, Copyright © 2000-2013 The FreeBSD Documentation Project,

. . .comments that "Starting with FreeBSD 10.X clang is installed as cc."

. . .but the port for FireFox-25.0,1 is "looking for" the gcc-4+ compiler!

. . .so why am I installing the gcc-4.6.3 compiler???? In consideration that it (is) going to be depreciated x( (Probably a rhetorical question.)

From this point on, I'm going to start a thread for FireFox-25.0,1.
 
GCC isn't going to deprecated completely, only the really old 4.2.1 version that has been shipping in the base system is. The port versions of GCC will stay and are needed to build ports that can't be yet built with CLANG. If the port isn't pulling in the proper GCC port as a build time dependency, for example lang/gcc, there is bug somewhere that needs fixing.
 
FireFox-25.0,1 Built Successfully

I have successfully built and installed FireFox-25.0,1. Interesting to note that make build needed, i.e., ===> firefox-25.0,1 depends on executable: gcc46 - not found and installed
Code: 
gcc-4.6.3_1                    GNU Compiler Collection 4.6
gcc-ecj-4.5                    Eclipse Java Compiler used to build GCC Java
. . .and installed the compilers.

It would be interesting to know if FireFox-24.0,1 would have compiled successfully if the make file contained the override to use the gcc compiler, but when v25 was announced, I didn't want to devote the hours and hours to test, and moved on to v25.

See this thread for more info: FireFox-25.0,1
 
rtwingfield said:
It would be interesting to know if FireFox-24.0,1 would have compiled successfully if the make file contained the override to use the gcc compiler, but when v25 was announced, I didn't want to devote the hours and hours to test, and moved on to v25.

See this thread for more info: FireFox-25.0,1
Click to expand...

9.2-RELEASE needed to override buggy clang on i386.
 
You must log in or register to reply here.
Back
Top