| View previous topic :: View next topic |
| Author |
Message |
gungholady
Guru
Joined: 19 Oct 2003
Posts: 392
|
 Posted: Wed Sep 08, 2004 10:55 pm Post subject: |
 |
|
| darkphader wrote: |
I ended up changing my method as well. Just copied /lib/libgcc_s-3.4.2.so.1 to /usr/lib/gcc/i686-pc-linux-gnu/3.4.2/libgcc_s.so.1 then symlinked /usr/lib/gcc/i686-pc-linux-gnu/3.4.2/libgcc_s.so to it. |
Thanks. This and I did the ldconfig seems to have solved the issue with qt. For the time being it is happily re-emerging. I'm just hoping that it solves the issue that caused to to try re-emerging qt. |
|
| Back to top |
|
 |
Lv
Retired Dev
Joined: 22 Jun 2003
Posts: 352
|
| Posted: Wed Sep 08, 2004 11:18 pm Post subject: |
|
|
| saber850 wrote: | | darkphader wrote: |
The problem appears to be that the ebuild, for some reason doesn't create the /usr/lib/gcc/i686-pc-linux-gnu/3.4.2/libgcc_s.so.1 file nor the /usr/lib/gcc/i686-pc-linux-gnu/3.4.2/libgcc_s.so link to it. As a workaround I placed links in /usr/lib/gcc/i686-pc-linux-gnu/3.4.2/ to the /lib/libgcc_s* files and added the libgcc_s.so link as well.
|
I too have this problem, except my binaries are linking against libgcc_s* from my gcc 3.3.4 build. Where are the libgcc_s* files for 3.4.2? I searched /usr/lib and didn't find them. |
i made so many changes to 3.4.2 that i'm amazed it isnt more broken than it already is, hence why it's masked -*. that said, this bug never shows up on multilib archs (like my own), and a fix for non-multilib archs is now in the tree.
for those of you who already compiled gcc:
| Code: |
cd /lib
ln -s libgcc_s.so.1 libgcc_s.so
|
if that doesnt work for you, -please- let me know at https://bugs.gentoo.org/show_bug.cgi?id=63333
I dont always pay attention to the forums. |
|
| Back to top |
|
|
manywele
l33t
Joined: 12 Jul 2003
Posts: 739
Location: Inside
|
| Posted: Thu Sep 09, 2004 3:40 am Post subject: |
|
|
Well I decided to try out gcc-3.4 so I emerged gcc-3.4.1-r2 and did
| Code: | emerge -e system
emerge -e system
emerge -e world |
but I'm left with 3 packages that won't compile.
openquicktime:
| Code: | gcc -I/usr/include/glib-1.2 -I/usr/lib/glib/include -fpic -O3 -march=athlon-xp -momit-leaf-frame-pointer -fomit-frame-pointer -ftracer -pipe -O3 -ffast-math -funroll-loops -fstrength-reduce -march=i586 -DHAVE_CONFIG_H -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -DUSE_MMX -DPENTIUM_OPT -I. -DBUILDDATE="\"Wed Sep 8 20:19:32 PDT 2004\"" -I../.. -I. -c audioencoder/lame3.70/quantize-pvt.c -o audioencoder/lame3.70/quantize-pvt.o
audioencoder/lame3.70/quantize-pvt.c: In function `quantize_xrpow':
audioencoder/lame3.70/quantize-pvt.c:978: error: PIC register `%ebx' clobbered in `asm'
make[1]: *** [audioencoder/lame3.70/quantize-pvt.o] Error 1
make[1]: Leaving directory `/var/tmp/portage/openquicktime-1.0-r1/work/openquicktime-1.0-src/audioplugin/MP3'
make: *** [mp3] Error 2
|
x11-misc/xaos:
lots of | Code: | formulas.c:257: error: conflicting types for 'truecolor_output'
formulas.c:253: error: previous declaration of 'truecolor_output' was here
....
formulas.c:1007: warning: initialization from incompatible pointer type
....
docalc.c:129: warning: 'mand_calc' used but never defined
|
and it eventually dies with
| Code: | gcc: lib/libui.a: No such file or directory
gcc: lib/libui-hlp.a: No such file or directory
gcc: lib/libengine.a: No such file or directory
gcc: lib/libutil.a: No such file or directory
make[1]: *** [/var/tmp/portage/xaos-3.1/work/XaoS-3.1/bin/xaos] Error 1
make[1]: Leaving directory `/var/tmp/portage/xaos-3.1/work/XaoS-3.1/src'
make: *** [default] Error 2
|
games-action/armagetron: lots of stuff like
| Code: | tLocale.h:112: error: expected `0' before ""
....
tList.h:143: warning: there are no arguments to `Len' that depend on a template parameter, so a declaration of `Len' must be available
tList.h:145: warning: there are no arguments to `operator()' that depend on a template parameter, so a declaration of `operator()' must be available
|
CFLAGS="-O3 -march=athlon-xp -fprefetch-loop-arrays -momit-leaf-frame-pointer -fomit-frame-pointer -ftracer -pipe"
ACCEPT_KEYWORDS="x86 ~x86"
Athlon-XP 2800+
They all compile fine with gcc-3.3.4 btw.
Any suggestions? |
|
| Back to top |
|
|
PrakashP
Veteran
Joined: 27 Oct 2003
Posts: 1249
Location: C.C.A.A., Germania
|
| Posted: Thu Sep 09, 2004 6:38 am Post subject: |
|
|
Try Using saner CFLAGS. This is the general rule. I something breaks, lower your flage.
In the second case, I think the package has a dependency which the ebuild forgot...
Perhaps the third one isn't yet gcc-3.4 compatible. Take a look into the sources... |
|
| Back to top |
|
|
manywele
l33t
Joined: 12 Jul 2003
Posts: 739
Location: Inside
|
| Posted: Thu Sep 09, 2004 7:00 am Post subject: |
|
|
| PrakashKC wrote: | | Try Using saner CFLAGS. This is the general rule. If something breaks, lower your flags. |
I do know that rule and I did try them all with CFLAGS=-O3 -march=athlon-xp -pipe and they still all failed in the same way. Are there saner flags I can use?
| Quote: | | In the second case, I think the package has a dependency which the ebuild forgot... |
Nope. As I said it compiles fine under gcc-3.3.4. Those missing libs get built in a succesful compile.
| Quote: | | Perhaps the third one isn't yet gcc-3.4 compatible. Take a look into the sources... |
I guess that's the kind of info I'm looking for. What exactly am I looking for to be 3.4 compatible? Can you link to some reading perhaps? |
|
| Back to top |
|
|
PrakashP
Veteran
Joined: 27 Oct 2003
Posts: 1249
Location: C.C.A.A., Germania
|
| Posted: Thu Sep 09, 2004 7:09 am Post subject: |
|
|
| Quote: | | I do know that rule and I did try them all with CFLAGS=-O3 -march=athlon-xp -pipe and they still all failed in the same way. Are there saner flags I can use? |
Try -march=i686 -O2 -pipe, for example. (This is more for testing.)
| Quote: |
I guess that's the kind of info I'm looking for. What exactly am I looking for to be 3.4 compatible? Can you link to some reading perhaps?
|
No link, but I rember the gcc 3.4 release notes showing the changes esp considering syntax which now leads to errors but was accepted before. |
|
| Back to top |
|
|
manywele
l33t
Joined: 12 Jul 2003
Posts: 739
Location: Inside
|
| Posted: Thu Sep 09, 2004 8:14 am Post subject: |
|
|
| PrakashKC wrote: |
Try -march=i686 -O2 -pipe, for example. (This is more for testing.)
|
No love. They still fail in the same way.
| Quote: |
No link, but I rember the gcc 3.4 release notes showing the changes esp considering syntax which now leads to errors but was accepted before. |
Thanks, I'll look into it. |
|
| Back to top |
|
|
spb
Retired Dev
Joined: 02 Jan 2004
Posts: 2135
Location: Cambridge, UK
|
| Posted: Thu Sep 09, 2004 10:01 am Post subject: |
|
|
| black hole sun wrote: | | New gcc release, eh, this one fixes alot of bugs. Does this mean it is finally stable enough to be unmasked in portage? Not that it matters to me, I can always get it through package.keywords, but when's it expected to be available to the generally conservative masses? |
Well, Lv for one seems to be pushing for 3.4 to be the default in the next release. It's already default for amd64, and there's interest in getting it marked stable on x86. Only problem is that some programs (eg openoffice) still don't compile with it. |
|
| Back to top |
|
|
Lv
Retired Dev
Joined: 22 Jun 2003
Posts: 352
|
| Posted: Thu Sep 09, 2004 3:59 pm Post subject: |
|
|
| thebell wrote: | | black hole sun wrote: | | New gcc release, eh, this one fixes alot of bugs. Does this mean it is finally stable enough to be unmasked in portage? Not that it matters to me, I can always get it through package.keywords, but when's it expected to be available to the generally conservative masses? |
Well, Lv for one seems to be pushing for 3.4 to be the default in the next release. It's already default for amd64, and there's interest in getting it marked stable on x86. Only problem is that some programs (eg openoffice) still don't compile with it. |
i'm not pushing anything for x86. i dont use it, so i'm not motivated to push.  -morfic- is the dev doing the pushing on x86. that and robmoss. i'm looking to make gcc 3.4.2, when stable after all those changes i made to it, the default for amd64's 2004.3 release. |
|
| Back to top |
|
|
geki
Advocate
Joined: 13 May 2004
Posts: 2387
Location: Germania
|
| Posted: Thu Sep 09, 2004 4:12 pm Post subject: |
|
|
After switching from 3.4.1-r2 to 3.4.2 doing
shows on last line only "GCC version 3.4.2" and not anymore "GCC version 3.4.1 (pie-8.7.6.5 etc... bla bla)".
Just wondering if the pie patches (and the others i do not remember) are applied or they just do not show up there anymore?!
no linux box available here for correct output  |
|
| Back to top |
|
|
teutzz
Guru
Joined: 22 Apr 2004
Posts: 333
Location: .ro
|
| Posted: Thu Sep 09, 2004 4:27 pm Post subject: |
|
|
| geki wrote: | After switching from 3.4.1-r2 to 3.4.2 doing
shows on last line only "GCC version 3.4.2" and not anymore "GCC version 3.4.1 (pie-8.7.6.5 etc... bla bla)".
Just wondering if the pie patches (and the others i do not remember) are applied or they just do not show up there anymore?!
no linux box available here for correct output |
same here
_________________
Cand nu stii ce sa raspunzi sau ce sa spui un simplu BLA ajunge... lolz |
|
| Back to top |
|
|
black hole sun
Apprentice
Joined: 14 Aug 2004
Posts: 245
|
| Posted: Thu Sep 09, 2004 10:36 pm Post subject: |
|
|
| Lv wrote: | | i'm not pushing anything for x86. i dont use it, so i'm not motivated to push. -morfic- is the dev doing the pushing on x86. that and robmoss. i'm looking to make gcc 3.4.2, when stable after all those changes i made to it, the default for amd64's 2004.3 release. |
Ok.
BTW, when can be expect your ebuild to go ~arch? I'm eager to compile the new gcc...without it totally breaking my system of course!  |
|
| Back to top |
|
|
Lv
Retired Dev
Joined: 22 Jun 2003
Posts: 352
|
| Posted: Fri Sep 10, 2004 12:17 am Post subject: |
|
|
| black hole sun wrote: |
BTW, when can be expect your ebuild to go ~arch? I'm eager to compile the new gcc...without it totally breaking my system of course! |
when it's ready. obviously.
people keep finding random bugs. plus embedded hasnt ported the uclibc patches. that and i'd like to move more of the ebuild into an eclass for easier maintenance. oh yeah, and i havent heard from ppc/ppc64... but i guess that shouldnt block unmasking it on other archs.
i dont -expect- it to break anything at the moment, but there are more updates being put in to the ebuild. for example, the version number patch thingie broke at some random point and i just recently fixed that. the version number should now be displayed correctly:
| Code: |
ayanami root # gcc --version
gcc (GCC) 3.4.2 (Gentoo Linux 3.4.2, ssp-3.4-2, pie-8.7.6.5)
Copyright (C) 2004 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
|
...oh yeah. that reminds me. there's a new version of propolice/ssp out but the tarball on ibm's site is corrupt. as soon as that's fixed i want to update to the new SSP... it contains some changes we've been pushing for.
hmm... what else... *shrug*. stuff.  |
|
| Back to top |
|
|
darkphader
Veteran
Joined: 09 May 2002
Posts: 1217
Location: Motown
|
| Posted: Fri Sep 10, 2004 2:28 am Post subject: xorg still doesn't like 3.4 |
|
|
I was hoping for success for gcc-3.4.2 and xorg-x11-6.8.0 but had to fall back to gcc-3.3.x. Xorg compiles and installs fine with 3.4.2 but it wont run long. Kde will start but then any attempt to run anything (even konsole) crashes it instantly. One bug report mentiones it doesn't like "pentium4" which I use, yet it all works fine with gcc-3.3.x.
_________________
WYSIWYG - What You See Is What You Grep |
|
| Back to top |
|
|
Xake
Guru
Joined: 11 Feb 2004
Posts: 588
Location: Göteborg, the rainy part of scandinavia
|
| Posted: Fri Sep 10, 2004 10:57 am Post subject: |
|
|
darkphader: Using -arch=pentium4?
gcc 3.4.2 && xorg6.7/xorg6.8 && -arch=pentium4 does not make sweet love on this system.
If I run that my X (using Gnome) restarts after max 15 seconds no matter what.
However if I compile it using -arch=pentium3 it works like it should.
A bit annoying.
But the pentium4- flag: is it a "combination" of other optimazionflags and is there any way to find out what in pentium4 xorg does not like? |
|
| Back to top |
|
|
Ssl
Apprentice
Joined: 21 Feb 2003
Posts: 178
Location: Serbia
|
| Posted: Fri Sep 10, 2004 12:02 pm Post subject: |
|
|
Hi,
Clean install with gcc 3.4.1. Emerge system gives:
| Code: | make[1]: Leaving directory `/var/tmp/portage/bash-3.0-r5/work/bash-3.0/lib/glob'
rm -f bash
gcc -L./builtins -L./lib/readline -L./lib/readline -L./lib/glob -L./lib/tilde - L./lib/sh -rdynamic -march=pentium4 -O3 -pipe -fomit-frame-pointer -o bash she ll.o eval.o y.tab.o general.o make_cmd.o print_cmd.o dispose_cmd.o execute_cmd. o variables.o copy_cmd.o error.o expr.o flags.o jobs.o subst.o hashcmd.o hashlib .o mailcheck.o trap.o input.o unwind_prot.o pathexp.o sig.o test.o version.o ali as.o array.o arrayfunc.o braces.o bracecomp.o bashhist.o bashline.o list.o stri nglib.o locale.o findcmd.o redir.o pcomplete.o pcomplib.o syntax.o xmalloc.o -lb uiltins -lsh -lreadline -lhistory //usr/lib/libcurses.a -lglob -ltilde -ldl
gcc: //usr/lib/libcurses.a: No such file or directory
make: *** [bash] Error 1
!!! ERROR: app-shells/bash-3.0-r5 failed.
!!! Function src_compile, Line 122, Exitcode 2
!!! make failed |
When switch to gcc 3.3.3 I get same error. Someone had this one?
Cheers,Ssl |
|
| Back to top |
|
|
spb
Retired Dev
Joined: 02 Jan 2004
Posts: 2135
Location: Cambridge, UK
|
| Posted: Fri Sep 10, 2004 12:29 pm Post subject: |
|
|
| Lv wrote: | | i'm not pushing anything for x86. i dont use it, so i'm not motivated to push. -morfic- is the dev doing the pushing on x86. that and robmoss. i'm looking to make gcc 3.4.2, when stable after all those changes i made to it, the default for amd64's 2004.3 release. |
Ok, I misspoke myself a bit. Lv, for one, seems to want gcc3.4 to be default in 2004.3 on the platform he uses. It's looking like it'll be the default in the next amd64 release, and there's interest from other devs to get it on x86 for 2004.3. That better? |
|
| Back to top |
|
|
darkphader
Veteran
Joined: 09 May 2002
Posts: 1217
Location: Motown
|
| Posted: Fri Sep 10, 2004 2:56 pm Post subject: |
|
|
| Xake wrote: | darkphader: Using -arch=pentium4?
gcc 3.4.2 && xorg6.7/xorg6.8 && -arch=pentium4 does not make sweet love on this system. |
Yes. But like you I would like to know why.
| Xake wrote: | | However if I compile it using -arch=pentium3 it works like it should. A bit annoying. |
And I wonder, for those of us who have both installed, which is actually better, simply compiling with -arch=pentium4 and using gcc-3.3.x or changing to -arch-pentium3 and using gcc-3.4.x.
| Xake wrote: | | But the pentium4- flag: is it a "combination" of other optimazionflags and is there any way to find out what in pentium4 xorg does not like? |
Maybe it's as simple as setting a particular CFLAG to turn off some optimization that gcc-3.4.x does by default that gcc-3.3.x doesn't to make the combination of -arch=pentium4, gcc-3.4.x, and xorg-x11 all work properly. But I don't know enough about it to hazard a guess.
_________________
WYSIWYG - What You See Is What You Grep |
|
| Back to top |
|
|
teutzz
Guru
Joined: 22 Apr 2004
Posts: 333
Location: .ro
|
| Posted: Fri Sep 10, 2004 4:06 pm Post subject: |
|
|
using gcc-3.4.2 final (not the the prerelease) i'm getting this error (never happend when using gcc-3.4.1-r2/r1): | Code: | checking for c++... g++
checking whether the C++ compiler (g++ -O2 -march=athlon-tbird -pipe -fomit-frame-pointer -ftracer -ffast-math -fforce-addr -fprefetch-loop-arrays -falign-functions=64 -fvisibility-inlines-hidden ) works... no
configure: error: installation or configuration problem: C++ compiler cannot create executables. |
_________________
Cand nu stii ce sa raspunzi sau ce sa spui un simplu BLA ajunge... lolz |
|
| Back to top |
|
|
Exci
Apprentice
Joined: 12 Jul 2002
Posts: 265
Location: The Netherlands, Zoetermeer
|
| Posted: Fri Sep 10, 2004 5:11 pm Post subject: |
|
|
I can't compile glibc-2.3.4.20040808 with gcc-3.4.1-r2
error:
| Code: |
checking for C cleanup handling... no
configure: error: the compiler must support C cleanup handling
!!! ERROR: sys-libs/glibc-2.3.4.20040808 failed.
!!! Function src_compile, Line 555, Exitcode 1
!!! (no error message)
|
emerge info
| Code: |
Portage 2.0.50-r11 (gcc34-x86-2004.2, gcc-3.4.1, glibc-2.3.4.20040808-r0, 2.6.7)
=================================================================
System uname: 2.6.7 i686 Intel(R) Pentium(R) M processor 1.70GHz
Gentoo Base System version 1.5.3
Autoconf: sys-devel/autoconf-2.59-r4
Automake: sys-devel/automake-1.8.5-r1
ACCEPT_KEYWORDS="x86 ~x86"
AUTOCLEAN="yes"
CFLAGS="-O3 -march=pentium-m -fomit-frame-pointer -fforce-addr -fprefetch-loop-arrays -falign-functions=64 -ftracer -mfpmath=sse"
CHOST="i686-pc-linux-gnu"
COMPILER=""
CONFIG_PROTECT="/etc /usr/X11R6/lib/X11/xkb /usr/kde/2/share/config /usr/kde/3/share/config /usr/share/config /usr/share/texmf/dvipdfm/config/ /usr/share/texmf/dvips/config/ /usr/share/texmf/tex/generic/config/ /usr/share/texmf/tex/platex/config/ /usr/share/texmf/xdvi/ /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d"
CXXFLAGS="-O3 -march=pentium-m -fomit-frame-pointer -fforce-addr -fprefetch-loop-arrays -falign-functions=64 -ftracer -mfpmath=sse"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoaddcvs ccache sandbox"
GENTOO_MIRRORS="http://gentoo.osuosl.org http://distro.ibiblio.org/pub/Linux/distributions/gentoo"
MAKEOPTS="-j2"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
PORTDIR_OVERLAY="/usr/local/portage"
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="X aalib acpi alsa audiofile avi bzlib cdr crypt dga directfbx divx4linux dvd encode extensions faac faad fbcon flac freetype ftp gcj gd gif gtk gtk2 imagemagick imlib ipv6 java javascript jp2 jpeg lcms libcaca live lzw mad mikmod mmx motif mpeg msn ncurses nptl oggvorbis opengl pam pcmcia pdflib perl pic png pnp python qt quicktime radeon readline sdl sockets spell sse ssl stroke svga tcltk tcpd tetex tiff truetype trusted usb video_cards_radeon wmf wxwindows x86 xml xml2 xmms xosd xv xvid zlib"
|
Does it matter if I keep using gcc compiled with gcc3.3 and the rest of the system compiled with gcc3.4?
Anybody any thoughts about the error and maybe a fix?
(I tried searching for this, but no hits on gcc3.4  ) |
|
| Back to top |
|
|
thubble
Tux's lil' helper
Joined: 16 May 2003
Posts: 119
Location: Winnipeg, Canada
|
| Posted: Sat Sep 11, 2004 8:07 pm Post subject: |
|
|
| Xake wrote: | | But the pentium4- flag: is it a "combination" of other optimazionflags and is there any way to find out what in pentium4 xorg does not like? |
The problem is with "-msse2". Compile xorg with "-march=pentium4 -mno-sse2" and it should work fine, at least it does for me. |
|
| Back to top |
|
|
darkphader
Veteran
Joined: 09 May 2002
Posts: 1217
Location: Motown
|
| Posted: Sat Sep 11, 2004 8:56 pm Post subject: |
|
|
| thubble wrote: | | The problem is with "-msse2". Compile xorg with "-march=pentium4 -mno-sse2" and it should work fine, at least it does for me. |
Does gcc-3.3.x work differently than gcc-3.4.x in this regard?
Does one enable sse2 support by default with -arch=pentium4 and the other not? Or is just the way sse2 code is handled by gcc-3.4.x that causes the problem?
If so, is it better, with a P4, to compile xorg-x11 with gcc-3.3.x and -arch=pentium4 rather than using one of the gcc-3.4.x workarounds (-mno-sse2 or -arch=pentium3)?
Thanks.
Chris
_________________
WYSIWYG - What You See Is What You Grep |
|
| Back to top |
|
|
morfic
Retired Dev
Joined: 28 Jul 2004
Posts: 31
|
| Posted: Sat Sep 11, 2004 11:35 pm Post subject: |
|
|
gcc 3.4 is no longer profile masked in 2004.2 default profiles
it is now ~x86 in 2004.2 profiles w/o a need for a special profile
hope anyone waiting for this will be glad, not this wont get you 3.4.2
but 3.4.2-r1 is good till 3.4.2 is all finished up
the great lv will bring this to you shortly |
|
| Back to top |
|
|
StringCheesian
l33t
Joined: 21 Oct 2003
Posts: 887
|
| Posted: Sun Sep 12, 2004 12:12 am Post subject: |
|
|
| teutzz: Try taking stuff out of your CFLAGS. That error is often caused by misspelled or unsupported flags in CFLAGS. Maybe your 3.4.2 doesn't have the visibility patch? |
|
| Back to top |
|
|
morfic
Retired Dev
Joined: 28 Jul 2004
Posts: 31
|
| Posted: Sun Sep 12, 2004 12:15 am Post subject: |
|
|
| nice typo s/3.4.2-r1/3.4.1-r2/ |
|
| Back to top |
|
|
|
Portage & Programming
