View previous topic :: View next topic |
Author |
Message |
zxy Veteran
Joined: 06 Jan 2006 Posts: 1160 Location: in bed in front of the computer
|
Posted: Sat Jun 02, 2007 1:06 pm Post subject: |
|
|
well, --skipfirst isn't really my favourite
it's much better to mask things that break, or to solve them. Just skipping can and eventualy will lead to broken system. At least how the --skipfirst was implemented in emerge when i used it the last time. _________________ Nature does not hurry, yet everything is accomplished.
Lao Tzu |
|
Back to top |
|
|
devsk Advocate
Joined: 24 Oct 2003 Posts: 2995 Location: Bay Area, CA
|
|
Back to top |
|
|
Ishiki Tux's lil' helper
Joined: 31 Aug 2005 Posts: 86
|
|
Back to top |
|
|
creidiki Apprentice
Joined: 23 Mar 2007 Posts: 283 Location: Varese (Italy)
|
Posted: Sat Jun 02, 2007 5:52 pm Post subject: |
|
|
--skipfirst is a horrid idea that should never have been implemented.
btw, could anyone see if gmp will compile with gcc-4.2.0 and binutils-2.17.50.0.16 if you add -mfpmath=sse in you CXXFLAGS?
it currently causes ld (2.17.50.0.16, .15 is fine) to segf w. gcc-4.1.2, at least on x86 (i686-pc-linux-gnu) targets: https://bugs.gentoo.org/show_bug.cgi?id=180498
if it compiled fine for your on an x86_64 target then at least I'll know its related to x86 only, as mfpmath=sse is default for x86_64. _________________ '((eINIT) (soor overlay)) |
|
Back to top |
|
|
Uzytkownik Guru
Joined: 31 Oct 2004 Posts: 399 Location: Bay Area, US
|
Posted: Thu Jun 07, 2007 2:13 pm Post subject: |
|
|
I've tried to build gcc with patch (thid one with memory leak while compilong X11):
Code: | /var/tmp/portage/sys-devel/gcc-4.2.0/work/build/./gcc/gfortran -B/var/tmp/portage/sys-devel/gcc-4.2.0/work/build/./gcc/ -B/usr/i686-pc-linux-gnu/bin/ -B/usr/i686-pc-linux-gnu/lib/ -isystem /usr/i686-pc-linux-gnu/include -isystem /usr/i686-pc-linux-gnu/sys-include -I . -Wall -fno-repack-arrays -fno-underscoring -g -O2 -c /var/tmp/portage/sys-devel/gcc-4.2.0/work/gcc-4.2.0/libgfortran/intrinsics/selected_int_kind.f90 -fPIC -DPIC -o .libs/selected_int_kind.o
/var/tmp/portage/sys-devel/gcc-4.2.0/work/gcc-4.2.0/libgfortran/intrinsics/selected_int_kind.f90: In function 'selected_int_kind':
/var/tmp/portage/sys-devel/gcc-4.2.0/work/gcc-4.2.0/libgfortran/intrinsics/selected_int_kind.f90:22: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://bugs.gentoo.org/> for instructions.
make[3]: *** [selected_int_kind.lo] Error 1
make[3]: Leaving directory `/var/tmp/portage/sys-devel/gcc-4.2.0/work/build/i686-pc-linux-gnu/libgfortran'
make[2]: *** [all] Error 2
make[2]: Leaving directory `/var/tmp/portage/sys-devel/gcc-4.2.0/work/build/i686-pc-linux-gnu/libgfortran'
make[1]: *** [all-target-libgfortran] Error 2
make[1]: Leaving directory `/var/tmp/portage/sys-devel/gcc-4.2.0/work/build'
make: *** [profiledbootstrap] Error 2
|
_________________ I've probably left my head... somwhere. Please wait untill I find it. |
|
Back to top |
|
|
mbar Veteran
Joined: 19 Jan 2005 Posts: 1990 Location: Poland
|
Posted: Thu Jun 07, 2007 2:20 pm Post subject: |
|
|
disable fortran. |
|
Back to top |
|
|
Genewb Apprentice
Joined: 09 Jan 2007 Posts: 165
|
Posted: Thu Jun 07, 2007 3:45 pm Post subject: |
|
|
mbar wrote: | disable fortran. |
That's bugged me for some time actually. Do generic desktop users really need Fortran at all? And perl for that matter ... |
|
Back to top |
|
|
Vlad.Sharp Guru
Joined: 08 Dec 2004 Posts: 337 Location: Cambridgeshire, UK
|
Posted: Thu Jun 07, 2007 3:48 pm Post subject: |
|
|
Genewb wrote: | mbar wrote: | disable fortran. |
That's bugged me for some time actually. Do generic desktop users really need Fortran at all? And perl for that matter ... |
Certainly not fortran! Perl can be useful, however - genlop for example comes to mind, or makepasswd, or any number of scripts... |
|
Back to top |
|
|
jcb118 n00b
Joined: 12 Oct 2006 Posts: 2
|
Posted: Fri Jun 08, 2007 2:55 am Post subject: openmp |
|
|
gcc 4.2 added OpenMP which allows you to add a handful of directives to existing source code to turn a standard program (fortran, c, c++?) into one that's executes in parallel on multiprocessor system. The best part is the directives read like comments to a non-OpenMP compiler, so your code can be compiled as usual or in parallel with the switch of a compile flag.
I'll be a happy guy when this gets closer to stable. |
|
Back to top |
|
|
Uzytkownik Guru
Joined: 31 Oct 2004 Posts: 399 Location: Bay Area, US
|
Posted: Fri Jun 08, 2007 8:26 am Post subject: |
|
|
boost-build cannot be build with gcc 4.2 (it builds ok with 4.1).
Code: | ###
### Using 'cc' toolset.
###
rm -rf bootstrap
mkdir bootstrap
i686-pc-linux-gnu-gcc -o bootstrap/jam0 -O2 -march=native -pipe -momit-leaf-frame-pointer -ggdb command.c compile.c debug.c execunix.c expand.c fileunix.c glob.c hash.c hdrmacro.c headers.c jam.c jambase.c jamgram.c lists.c make.c make1.c newstr.c option.c parse.c pathunix.c pathvms.c regexp.c rules.c scan.c search.c subst.c timestamp.c variable.c modules.c strings.c filesys.c builtins.c pwd.c class.c native.c w32_getreg.c modules/set.c modules/path.c modules/regex.c modules/property-set.c modules/sequence.c modules/order.c execnt.c filent.c
./bootstrap/jam0 -f build.jam --toolset=cc --toolset-root= clean
./build.sh: line 13: 32516 Segmentation fault $@
|
With -mfpmath=sse it warns that it use 387 and disable sse.
Boost itself seems to build ok.
Is it my bug, known bug of gcc/boost or unknown bug? _________________ I've probably left my head... somwhere. Please wait untill I find it. |
|
Back to top |
|
|
kernelOfTruth Watchman
Joined: 20 Dec 2005 Posts: 6111 Location: Vienna, Austria; Germany; hello world :)
|
Posted: Fri Jun 08, 2007 9:03 am Post subject: |
|
|
Uzytkownik wrote: | boost-build cannot be build with gcc 4.2 (it builds ok with 4.1).
Code: | ###
### Using 'cc' toolset.
###
rm -rf bootstrap
mkdir bootstrap
i686-pc-linux-gnu-gcc -o bootstrap/jam0 -O2 -march=native -pipe -momit-leaf-frame-pointer -ggdb command.c compile.c debug.c execunix.c expand.c fileunix.c glob.c hash.c hdrmacro.c headers.c jam.c jambase.c jamgram.c lists.c make.c make1.c newstr.c option.c parse.c pathunix.c pathvms.c regexp.c rules.c scan.c search.c subst.c timestamp.c variable.c modules.c strings.c filesys.c builtins.c pwd.c class.c native.c w32_getreg.c modules/set.c modules/path.c modules/regex.c modules/property-set.c modules/sequence.c modules/order.c execnt.c filent.c
./bootstrap/jam0 -f build.jam --toolset=cc --toolset-root= clean
./build.sh: line 13: 32516 Segmentation fault $@
|
With -mfpmath=sse it warns that it use 387 and disable sse.
Boost itself seems to build ok.
Is it my bug, known bug of gcc/boost or unknown bug? |
more a known bug (known to upstream?) this also happened with snapshots from january / december last year ... _________________ https://github.com/kernelOfTruth/ZFS-for-SystemRescueCD/tree/ZFS-for-SysRescCD-4.9.0
https://github.com/kernelOfTruth/pulseaudio-equalizer-ladspa
Hardcore Gentoo Linux user since 2004 |
|
Back to top |
|
|
cneil n00b
Joined: 28 May 2007 Posts: 39 Location: Stillwater, OK
|
Posted: Sat Jun 09, 2007 12:30 am Post subject: |
|
|
Uzytkownik wrote: | boost-build cannot be build with gcc 4.2 (it builds ok with 4.1). |
See Bug #181139. |
|
Back to top |
|
|
neuron Advocate
Joined: 28 May 2002 Posts: 2371
|
Posted: Sun Jun 10, 2007 3:56 pm Post subject: |
|
|
Anyone tried openmp? I'm getting very mixed results, see https://forums.gentoo.org/viewtopic-t-564312-highlight-.html
Also, anyone seen the gcc 4.3 ebuilds? Thought I saw em around on the forums but I can't find that thread anymore. Figured I'd check if it's been fixed upstream or not. |
|
Back to top |
|
|
Tom_ Guru
Joined: 20 May 2004 Posts: 444 Location: France
|
Posted: Mon Jun 11, 2007 2:57 pm Post subject: |
|
|
Hi,
What about -ftree-vectorize? Is it stable enough ? Does GCC-4.2 really provides some improvements about this flag?
C++ visibility handling has been improved in GCC 4.2 but it is not useful to enable it in the make.conf, is it ? I've heard it is already enabled in the software that can really use it. Is it correct ?
Thank you in advance.
Tom. |
|
Back to top |
|
|
kernelOfTruth Watchman
Joined: 20 Dec 2005 Posts: 6111 Location: Vienna, Austria; Germany; hello world :)
|
Posted: Mon Jun 11, 2007 3:09 pm Post subject: |
|
|
anyone tried to install gcc-4.2.0 from kevquinn's overlay on hardened?
install (=make install) part always fails with:
Quote: | * Installing GCC...
make[1]: Entering directory `/var/tmp/paludis/sys-devel/gcc-4.2.0/work/build'
/bin/sh /var/tmp/paludis/sys-devel/gcc-4.2.0/work/gcc-4.2.0/mkinstalldirs /var/tmp/paludis/sys-devel/gcc-4.2.0/image//usr /var/tmp/paludis/sys-devel/gcc-4.2.0/image//usr
mkdir -p -- /var/tmp/paludis/sys-devel/gcc-4.2.0/image//usr /var/tmp/paludis/sys-devel/gcc-4.2.0/image//usr
make[2]: Entering directory `/var/tmp/paludis/sys-devel/gcc-4.2.0/work/build/fixincludes'
rm -rf /var/tmp/paludis/sys-devel/gcc-4.2.0/image//usr/libexec/gcc/i686-pc-linux-gnu/4.2.0/install-tools
/bin/sh /var/tmp/paludis/sys-devel/gcc-4.2.0/work/gcc-4.2.0/fixincludes/../mkinstalldirs /var/tmp/paludis/sys-devel/gcc-4.2.0/image//usr/libexec/gcc/i686-pc-linux-gnu/4.2.0/install-tools
mkdir -p -- /var/tmp/paludis/sys-devel/gcc-4.2.0/image//usr/libexec/gcc/i686-pc-linux-gnu/4.2.0/install-tools
/bin/sh /var/tmp/paludis/sys-devel/gcc-4.2.0/work/gcc-4.2.0/fixincludes/../mkinstalldirs /var/tmp/paludis/sys-devel/gcc-4.2.0/image//usr/lib/gcc/i686-pc-linux-gnu/4.2.0/install-tools/include
mkdir -p -- /var/tmp/paludis/sys-devel/gcc-4.2.0/image//usr/lib/gcc/i686-pc-linux-gnu/4.2.0/install-tools/include
/usr/bin/install -c -m 644 /var/tmp/paludis/sys-devel/gcc-4.2.0/work/gcc-4.2.0/fixincludes/README-fixinc \
/var/tmp/paludis/sys-devel/gcc-4.2.0/image//usr/lib/gcc/i686-pc-linux-gnu/4.2.0/install-tools/include/README
/usr/bin/install -c fixinc.sh /var/tmp/paludis/sys-devel/gcc-4.2.0/image//usr/libexec/gcc/i686-pc-linux-gnu/4.2.0/install-tools/fixinc.sh
/usr/bin/install -c fixincl \
/var/tmp/paludis/sys-devel/gcc-4.2.0/image//usr/libexec/gcc/i686-pc-linux-gnu/4.2.0/install-tools/fixincl
/usr/bin/install -c mkheaders /var/tmp/paludis/sys-devel/gcc-4.2.0/image//usr/libexec/gcc/i686-pc-linux-gnu/4.2.0/install-tools/mkheaders
make[2]: Leaving directory `/var/tmp/paludis/sys-devel/gcc-4.2.0/work/build/fixincludes'
make[2]: Entering directory `/var/tmp/paludis/sys-devel/gcc-4.2.0/work/build/gcc'
gcc -DEFAULT_PIE -DEFAULT_RELRO -DEFAULT_BIND_NOW -march=pentium-m -pipe -O2 -DIN_GCC -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long -Wno-variadic-macros -Wno-overlength-strings -Wold-style-definition -Wmissing-format-attribute -DHAVE_CONFIG_H -I. -I. -I/var/tmp/paludis/sys-devel/gcc-4.2.0/work/gcc-4.2.0/gcc -I/var/tmp/paludis/sys-devel/gcc-4.2.0/work/gcc-4.2.0/gcc/. -I/var/tmp/paludis/sys-devel/gcc-4.2.0/work/gcc-4.2.0/gcc/../include -I/var/tmp/paludis/sys-devel/gcc-4.2.0/work/gcc-4.2.0/gcc/../libcpp/include -I/var/tmp/paludis/sys-devel/gcc-4.2.0/work/gcc-4.2.0/gcc/../libdecnumber -I../libdecnumber \
-DGCC_INCLUDE_DIR=\"/usr/lib/gcc/i686-pc-linux-gnu/4.2.0/include\" -DGPLUSPLUS_INCLUDE_DIR=\"/usr/lib/gcc/i686-pc-linux-gnu/4.2.0/include/g++-v4\" -DGPLUSPLUS_TOOL_INCLUDE_DIR=\"/usr/lib/gcc/i686-pc-linux-gnu/4.2.0/include/g++-v4/i686-pc-linux-gnu\" -DGPLUSPLUS_BACKWARD_INCLUDE_DIR=\"/usr/lib/gcc/i686-pc-linux-gnu/4.2.0/include/g++-v4/backward\" -DLOCAL_INCLUDE_DIR=\"/usr/local/include\" -DCROSS_INCLUDE_DIR=\"/usr/lib/gcc/i686-pc-linux-gnu/4.2.0/../../../../i686-pc-linux-gnu/sys-include\" -DTOOL_INCLUDE_DIR=\"/usr/lib/gcc/i686-pc-linux-gnu/4.2.0/../../../../i686-pc-linux-gnu/include\" \
-c /var/tmp/paludis/sys-devel/gcc-4.2.0/work/gcc-4.2.0/gcc/cppdefault.c -o cppdefault.o
cc1: error: unrecognized command line option "-Wno-overlength-strings"
make[2]: *** [cppdefault.o] Error 1
make[2]: Leaving directory `/var/tmp/paludis/sys-devel/gcc-4.2.0/work/build/gcc'
make[1]: *** [install-gcc] Error 2
make[1]: Leaving directory `/var/tmp/paludis/sys-devel/gcc-4.2.0/work/build'
make: *** [install] Error 2
!!! ERROR in sys-devel/gcc-4.2.0:
!!! In gcc-compiler_src_install at line 3312
!!! (no message provided)
!!! Call stack:
!!! * gcc-compiler_src_install (/var/tmp/paludis/sys-devel/gcc-4.2.0/temp/loadsaveenv:3312)
!!! * toolchain_src_install (/var/tmp/paludis/sys-devel/gcc-4.2.0/temp/loadsaveenv:6022)
!!! * src_install (/var/tmp/paludis/sys-devel/gcc-4.2.0/temp/loadsaveenv:5621)
!!! * ebuild_f_install (/usr/libexec/paludis/src_install.bash:4
!!! * ebuild_main (/usr/libexec/paludis/ebuild.bash:363)
!!! * main (/usr/libexec/paludis/ebuild.bash:381)
die trap: exiting with error.
Install error:
* In program paludis -i =sys-devel/gcc-4.2.0 -1:
* When performing install action from command line:
* When executing install task:
* When installing 'sys-devel/gcc-4.2.0':
* Install error: Install failed for 'sys-devel/gcc-4.2.0::portage'
|
this seems to be a
known error
many thanks in advance _________________ https://github.com/kernelOfTruth/ZFS-for-SystemRescueCD/tree/ZFS-for-SysRescCD-4.9.0
https://github.com/kernelOfTruth/pulseaudio-equalizer-ladspa
Hardcore Gentoo Linux user since 2004 |
|
Back to top |
|
|
Genewb Apprentice
Joined: 09 Jan 2007 Posts: 165
|
Posted: Tue Jun 12, 2007 12:00 am Post subject: |
|
|
Tom_ wrote: | Hi,
What about -ftree-vectorize? Is it stable enough ? Does GCC-4.2 really provides some improvements about this flag?
C++ visibility handling has been improved in GCC 4.2 but it is not useful to enable it in the make.conf, is it ? I've heard it is already enabled in the software that can really use it. Is it correct ?
Thank you in advance.
Tom. |
-ftree-vectorize seems perfectly stable on ~amd64 for most packages. I haven't tried it with known problem packages, aside from librsvg. That being said, the problem may be at least in part Pentium 4 specific. |
|
Back to top |
|
|
devsk Advocate
Joined: 24 Oct 2003 Posts: 2995 Location: Bay Area, CA
|
Posted: Tue Jun 12, 2007 12:43 am Post subject: |
|
|
Genewb wrote: | Tom_ wrote: | Hi,
What about -ftree-vectorize? Is it stable enough ? Does GCC-4.2 really provides some improvements about this flag?
C++ visibility handling has been improved in GCC 4.2 but it is not useful to enable it in the make.conf, is it ? I've heard it is already enabled in the software that can really use it. Is it correct ?
Thank you in advance.
Tom. |
-ftree-vectorize seems perfectly stable on ~amd64 for most packages. I haven't tried it with known problem packages, aside from librsvg. That being said, the problem may be at least in part Pentium 4 specific. | Has anybody quantized the gains with -ftree-vectorize on a specific package? like some sort of comparison with and without. anyone? |
|
Back to top |
|
|
Tom_ Guru
Joined: 20 May 2004 Posts: 444 Location: France
|
Posted: Tue Jun 12, 2007 10:11 am Post subject: |
|
|
Genewb wrote: |
-ftree-vectorize seems perfectly stable on ~amd64 for most packages. I haven't tried it with known problem packages, aside from librsvg. That being said, the problem may be at least in part Pentium 4 specific. |
Ok it is cool since I also run an ~amd64 system!
devsk wrote: |
Has anybody quantized the gains with -ftree-vectorize on a specific package? like some sort of comparison with and without. anyone? |
Yeah it would be great to have this kind of comparison.
Thank you a lot! |
|
Back to top |
|
|
irondog l33t
Joined: 07 Jul 2003 Posts: 715 Location: Voor mijn TV. Achter mijn pc.
|
Posted: Tue Jun 12, 2007 7:55 pm Post subject: |
|
|
Tom_ wrote: | C++ visibility handling has been improved in GCC 4.2 but it is not useful to enable it in the make.conf, is it ? | It will never be good for make.conf. You might be lucky not running into problems when enabling it, but it's simply incorrect to put it in make.conf.
Some things will never work with hidden visibility and that's why it's up to a developer to enable it or not for specific packages or even specific objecs within a package. _________________ Alle dingen moeten onzin zijn. |
|
Back to top |
|
|
creidiki Apprentice
Joined: 23 Mar 2007 Posts: 283 Location: Varese (Italy)
|
Posted: Wed Jun 13, 2007 9:41 am Post subject: |
|
|
-fvisiblity-inlines-hidden you can add to your core CXXFLAGS with no issues, -fvisibility=hidden is best left alone. _________________ '((eINIT) (soor overlay)) |
|
Back to top |
|
|
loki_val Retired Dev
Joined: 13 Nov 2006 Posts: 418 Location: Denmark
|
Posted: Mon Jun 18, 2007 8:30 pm Post subject: |
|
|
For people having problems with xorg-server taking *quite* a bit of time to compile, see this happy news from #gentoo-commits: Code: | 03vapier * 10gentoo/src/patchsets/gcc/4.2.0/gentoo/ (40_all_gcc42-PR30052.patch README.history): fix from upstream for memory hog #180304 | Remember that since GCC 4.2 is still package.masked, there will be no revision bumps. Emerge --sync and reemerge GCC to get this patch. |
|
Back to top |
|
|
unK l33t
Joined: 06 Feb 2007 Posts: 769
|
Posted: Tue Jun 19, 2007 7:29 am Post subject: |
|
|
unfortunately, emerge gcc fails with this message:
Code: | !!! Couldn't download 'gcc-4.2.0-patches-1.2.tar.bz2'. Aborting. |
has anyone got this file already?
edit: nevermind, I've got it. |
|
Back to top |
|
|
irondog l33t
Joined: 07 Jul 2003 Posts: 715 Location: Voor mijn TV. Achter mijn pc.
|
Posted: Tue Jun 19, 2007 7:25 pm Post subject: |
|
|
Patchset 1.2 sucks
I did not have any unexpected problems with the gcc-4.2.0 package and the previous patchset. I neither had xorg-x11 build problemes nor runtime errors, both 1.2 and 1.3 work.
With the current patchset some objects take ages to build
This process is taking 15% of all memory and is taking 100% of one core.:
Code: | 10000 gerte 25 0 329m 314m 4700 R 100 15.6 6:25.91 cc1plus |
It has been compiling a simple C++ file for over 6 minutes. Not a proper fix in my eyes..... _________________ Alle dingen moeten onzin zijn. |
|
Back to top |
|
|
loki_val Retired Dev
Joined: 13 Nov 2006 Posts: 418 Location: Denmark
|
Posted: Tue Jun 19, 2007 8:01 pm Post subject: |
|
|
Irondog: bugs.gentoo.org, please. That sounds like a serious regression. What was the compile time with patchset 1.1?
Last edited by loki_val on Tue Jun 19, 2007 8:26 pm; edited 1 time in total |
|
Back to top |
|
|
Vlad.Sharp Guru
Joined: 08 Dec 2004 Posts: 337 Location: Cambridgeshire, UK
|
Posted: Tue Jun 19, 2007 8:02 pm Post subject: |
|
|
Oooh could you post the source of this c++ file? |
|
Back to top |
|
|
|