HELP: Compiler crashes while emerging Gnome (Mozilla)

[theosib]

When I 'emerge gnome', and it gets to the point where it's building Mozilla, the build dies.

The information on the screen isn't very informative to me, but this is what I see:

[It's compiling nsContentIterator.cpp]
The bug is not reproducable, so it is likely a hardware or OS problem
... [leaving directories] ...

!!! Error: net-www/mozilla-1.6-r1 failed
!!! Function src_compile, Line 378, Exitcode 2
!!! (no error message)



Looks to me like this is a GCC problem. Anyone have suggestions for getting past this? "The bug" isn't very specific.


Oh, specifics on my system:

Athlon XP 2800+ (compiling with -march=athlon-xp -O2 -fomit-frame-pointer -pipe)
ABIT KD7 motherboard (VIA KT400 chipset)
3ware RAID in RAID1 with two WD1200JB drives
1 gig of RAM
I built a gentoo kernel (2.4.22-gentoo-r7)
I built the system using stage1 and using the 2004.0 livecd.

I'm not sure what other important information you may need to know about my build environment. But something's broken. PLEASE HELP!

Thanks.

[iten]

Strange. I actually had the same problem when I was doing an upgrade. I'm pretty sure it was on the same file, too.
I have an Athlon Duron processor. Is this some problem with AMDs?

[superjaded]

FWIW, I compiled mozilla 1.6-r1 on my machine last night just fine.
An AMD DUron 900mhz.

What CFLAGS are you using?
Mine are: -Os -march=i686 -funroll-loops -pipe

As far as the compiletime error.. that's the first time I've ever seen something like that.

[d_f0rce]

[theosib]

Ok, some good advice from others here. I was using '-j2', so the error was hidden higher up. Turning that off, this is the tail end of the compile:

It's some kind of assember error.... Does this help with the troubleshooting?


nsChromeProtocolHandler.cpp
g++ -o nsChromeProtocolHandler.o -c -DOSTYPE=\"Linux2.4\" -DOSARCH=\"Linux\" -I../../../dist/include/xpcom -I../../../dist/include/string -I../../../dist/include/rdf -I../../../dist/include/xuldoc -I../../../dist/include/gfx -I../../../dist/include/layout -I../../../dist/include/content -I../../../dist/include/necko -I../../../dist/include/dom -I../../../dist/include/intl -I../../../dist/include/widget -I../../../dist/include/js -I../../../dist/include/appshell -I../../../dist/include/caps -I../../../dist/include/pref -I../../../dist/include/docshell -I../../../dist/include/imglib2 -I../../../dist/include/xpconnect -I../../../dist/include/jar -I../../../dist/include/chrome -I../../../dist/include -I/var/tmp/portage/mozilla-1.6-r1/work/mozilla/dist/include/nspr -I/usr/X11R6/include -fPIC -I/usr/X11R6/include -fno-rtti -fno-exceptions -Wall -Wconversion -Wpointer-arith -Wcast-align -Woverloaded-virtual -Wsynth -Wno-ctor-dtor-privacy -Wno-long-long -mcpu=athlon-xp -march=athlon-xp -pipe -s -fforce-addr -Wno-deprecated -fshort-wchar -pthread -pipe -DNDEBUG -DTRIMMED -ffunction-sections -O2 -I/usr/X11R6/include -DMOZILLA_CLIENT -include ../../../mozilla-config.h -Wp,-MD,.deps/nsChromeProtocolHandler.pp nsChromeProtocolHandler.cpp
{standard input}: Assembler messages:
{standard input}:2062: Warning: rest of line ignored; first ignored character is `,'
{standard input}:2063: Error: unknown pseudo-op: `.p2ovign'
{standard input}:2064: Error: unknown pseudo-op: `.gloca'
make[4]: *** [nsChromeProtocolHandler.o] Error 1
make[4]: Leaving directory `/var/tmp/portage/mozilla-1.6-r1/work/mozilla/rdf/chrome/src'
make[3]: *** [libs] Error 2
make[3]: Leaving directory `/var/tmp/portage/mozilla-1.6-r1/work/mozilla/rdf/chrome'
make[2]: *** [libs] Error 2
make[2]: Leaving directory `/var/tmp/portage/mozilla-1.6-r1/work/mozilla/rdf'
make[1]: *** [tier_9] Error 2
make[1]: Leaving directory `/var/tmp/portage/mozilla-1.6-r1/work/mozilla'
make: *** [default] Error 2

!!! ERROR: net-www/mozilla-1.6-r1 failed.
!!! Function src_compile, Line 378, Exitcode 2
!!! (no error message)

[simulacrum]

I just tried emerging mozilla-1.6-r1 and am receiving the same error myself. I just submitted a bug report on this. Man, that was my first experience reporting a bug. I'll be right back after I have a beer. Oh, it's Bug 44213.

[theosib]

I though I should point out, if anyone is listening, that this has me completely hosed here. I'm stuck, and I cannot complete the setup of my gentoo system. Designers of the Gentoo distribution seem to have an interest in having other people USE Gentoo, so I would hope that they could help me resolve this problem before I have to give up and try something else.

Thanks!

[simulacrum]

Out of curiosity, how is being unable to emerge mozilla preventing you from setting up the rest of the system? If access to the web is critical for your configuration, there are other browsers you can use in the interim, such as Konqueror or Galeon.

On an off note, I'd like to point out that the Gentoo distribution doesn't cost any of us a dime, and the majority of us are Gentoo fanatics. The Gentoo developers have done a wonderful job thus far and put forth a good deal of effort. It pains me when someone takes such a negative attitude and act as if the develops owe them something. Please be patient, I'm sure this issue will be resolved as quickly as possible.

[theosib]

First of all, the reason that it's preventing me from setting up is that the gnome build seems to depend on mozilla, although now that I think about it, it may be that's because I put "mozilla" in my "USE" string.

Also, I don't mean to insult Gentoo developers, and what I said was indeed out of place and inappropriate. Really, they should take pride in how well things have worked so far. Indeed, their online docs are EXCELLENT, and it's wonderful to just "emerge" a program and have it WORK. I was simply frustrated.

I'll remove the 'mozilla' dependency for now.

[theosib]

Removing "mozilla" from my USE parameter in make.conf does not seem to have removed the gnome dependency on mozilla. THAT is indeed preventing me from building my system.

And of course, being new to all of this, I have no idea how to remove that dependency.

I beg of all those who know to please help me to solve this!

Thanks!

[superjaded]

You can't.
Epiphany is part of the GNOME desktop and it depends on Gecko (read: Mozilla) for its rendering.

The only option you really have is to edit the gnome-base/gnome ebuild and take Epiphany out of the dependancies.

But GNOME isn't required for desktop usage -- there are hosts of other WMs/DEs that don't require Mozilla at all: KDE, Xfce4, the *boxes, to name a few. You can use those until people figure out what's going wrong with your compile.

[theosib]

Apparently, there is the command "emerge -i" which I can use to trick the system into thinking something is installed even when it's not. Is there anything which depends on epiphany that'll break if I 'inject' it?

In the mean time, I'll emerge kde. I've been meaning to do that anyhow.

Thanks!

[moocha]

Please post the output of

[theosib]

Here is my emerge info:

Gentoo Base System version 1.4.3.13
Portage 2.0.50-r1 (default-x86-2004.0, gcc-3.3.2, glibc-2.3.2-r9, 2.4.22-gentoo-r7)
=================================================================
System uname: 2.4.22-gentoo-r7 i686 AMD Athlon(tm) XP 2800+
Autoconf: sys-devel/autoconf-2.58-r1
Automake: sys-devel/automake-1.7.7
ACCEPT_KEYWORDS="x86"
AUTOCLEAN="yes"
CFLAGS="-O2 -mcpu=athlon-xp -march=athlon-xp -fomit-frame-pointer -pipe"
CHOST="i686-pc-linux-gnu"
COMPILER="gcc3"
CONFIG_PROTECT="/etc /usr/X11R6/lib/X11/xkb /usr/kde/2/share/config /usr/kde/3/share/config /usr/share/config /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d"
CXXFLAGS="-O2 -mcpu=athlon-xp -march=athlon-xp -fomit-frame-pointer -pipe"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoaddcvs ccache sandbox"
GENTOO_MIRRORS="http://www.gtlib.cc.gatech.edu/pub/gentoo ftp://gentoo.mirrors.pair.com/ http://mirrors.tds.net/gentoo http://gentoo.mirrors.pair.com/"
MAKEOPTS="-j2"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
PORTDIR_OVERLAY=""
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="3dnow X Xaw3d acpi aim apache2 apm arts avi berkdb bonobo cdr crypt cups dga directfb emacs emacs-w3 encode esd evo fbcon foomaticdb gb gd gdbm gif gnome gpm gtk gtk2 gtkhtml imagemagick imlib innodb java jpeg kde libg++ libgda libwww mad mikmod mmx motif mozilla mpeg msn mysql ncurses nls oggvorbis opengl oss pam pdflib perl plotutils png postgres python qt quicktime readline samba sasl scanner sdl slang spell sse ssl svga tcltk tcpd theora tiff truetype unicode usb wxwindows x86 xml xml2 xmms xv zlib"



BTW, there is something very odd. I removed MAKEOPS="-j2" from my /etc/make.conf file, and in fact, that does seem to have turned off -j2, so why does my "emerge info" say that I have -j2?

Anyhow, does this tell you anything useful?


THANKS!

[moocha]

Sadly no... I had suspected you had used some of the more aggressive optimization flags (-fprefetch-loop-arrays being the cause of a lot of problems), but that's not the case, and I'm stumped on this... The emerge works just fine for me.
Incidentally, the "-mcpu=athlon-xp" in your CFLAGS setting is redundant - the gcc manualdirectly states that "-march=athlon-xp" implies "-mcpu=athlon-xp".
As to the MAKEOPTS problem - my /etc/make.conf explicitely sets

[theosib]

Thanks for responding.

Just to let you know, I had since learned that -mcpu was redundant, but I just hadn't bothered to fix it. Also, just a moment ago, I was doing some research and found out about /etc/make.globals, which has a -j2 in it, so commenting out the MAKEOPTS line in /etc/make.conf was the wrong thing to do.

Have you tried emerging specifically "net-www/mozilla-1.6-r1"? That is the one which is causing the compiler crash.

[theosib]

I'm getting the same compiler crash with postgresql.

gcc -O2 -mcpu=athlon-xp -march=athlon-xp -fomit-frame-pointer -pipe -Wall -Wmissing-prototypes -Wmissing-declarations -I. -I../../../src/include -c -o gram.o
gram.c -MMD
y.tab.c: In function `yyparse':
y.tab.c:17410: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:https://bugs.gentoo.org/> for instructions.
The bug is not reproduceable, so it is likely a hardware or OS problem
make[3]: *** [gram.o] Error 1
make[3]: Leaving directory `/var/tmp/portage/postgresql-7.3.5/work/postgresql-7.3.5/src/backend/parser'
make[2]: *** [parser-recursive] Error 2
make[2]: Leaving directory `/var/tmp/portage/postgresql-7.3.5/work/postgresql-7.3.5/src/backend'
make[1]: *** [all] Error 2
make[1]: Leaving directory `/var/tmp/portage/postgresql-7.3.5/work/postgresql-7.3.5/src'
make: *** [all] Error 2

!!! ERROR: dev-db/postgresql-7.3.5 failed.
!!! Function src_compile, Line 120, Exitcode 2
!!! (no error message)

[moocha]

[theosib]

I think I've have some additional clues about where the crash is coming from.

I believe my GCC is broken. Not all, but most packages I try to emerge cause GCC to crash. Even GCC itself. "emerge gcc" crashes while compiling "fold-const.c". This happens if I tell GCC to compile with NO optimization at all: CFLAGS=""

At this point, I can only assume that everything that has been built by GCC so far is also suspect, because whatever is causing GCC to crash may also have caused GCC to produce incorrect object code.

I can only assume that one of my optimization options has caused the problem, probably "-march=athlon-xp". GCC probably can't produce code correctly optimized for Athlon XP. That guess is based on the fact that others are also using the "-pipe -fomit-frame-pointer -O2" options successfully.

So much for Gentoo's claim that you can fully optimize for your own system.

I'm going to wait around a bit longer to see if anyone can give me some advice. Actually, if I can get a GCC binary package that isn't broken, and I can figure out a way to make portage recompile EVERYTHING in the system, then I won't have to do a completely fresh install (which too forever the first time I did it, but doing a reemerge should just run unattended).

So, if "athlon-xp" is broken, how much less performance will I get out of "athlon"? What are the chances someone will look at the bugzilla report I made and figure out why GCC is producing bad code? [So, the bootstrap GCC didn't crash, but when it compiled the work GCC, it produced incorrect code that causes the work GCC to crash.]

[moocha]

I highly doubt it's -march=athlon-xp. I've been using that architecture flag for months now, with gcc 3.2.whatever (can't remember), with gcc 3.3.2, and now with gcc 3.3.3 - all worked like a charm...
Here's the compiler banner:

[theosib]

I'm not home right now, but the version of GCC that is being used is 3.3.2-r5, I believe. All later versions are "masked".

Some research has suggested that -fomit-frame-pointer is a really bad thing to use.

Also, if anyone can tell me where I can get a binary package of gcc, please do, because I can't even get GCC to recompile itself! GCC crashes while compiling GCC.

[jra822]

My configuration is very similar to yours except that i use the 2.4 linux-headers (with the 2.6 kernel). I think (not sure) that I read somewhere that the 2.6 headers would cause problems with some current software (or at least some unmasked portage packages), but the 2.4 headers seem to work fine (at least for me).

If you switched to the 2.4 linux-headers it may fix your problem, but I would expect that you would still need to replace your gcc.

You may be able to install the GRP build of gcc, but I have never used GRP and would have no idea of how to do this.

Good Luck

[theosib]

I haven't yet gotten quite adventurous to try the 2.6 kernel. I'm still using "2.4.22-gentoo-r7".

[moocha]

[langthang]