GCC-4.0 (part 3)

Message

gralves · Post by **gralves** » Thu Sep 22, 2005 4:40 pm

edit by nixnut: continuation of GCC-4.0 (part 2)

I'm having some trouble compiling firefox. I think it is the same problem I have w/ blender - pointer to int conversions, on amd64, loose precision. There will probably be a lot of packages w/ this kind of problem.

Code: Select all

x86_64-pc-linux-gnu-g++ -o nsFreeType.o -c -DOSTYPE=\"Linux2.6\" -DOSARCH=\"Linux\" -I../.. -I../../../dist/include/xpcom -I../../../dist/include/string -I../../../dist/include/pref -I../../../dist/include/uconv -I../../../dist/include/unicharutil -I../../../dist/include/gfx -I../../../dist/include -I/var/tmp/portage/mozilla-firefox-1.0.6-r7/work/mozilla/dist/include/nspr     -I/usr/include/freetype2    -fPIC  -DGENTOO_NSPLUGINS_DIR=\"/usr/lib64/nsplugins\" -DGENTOO_NSBROWSER_PLUGINS_DIR=\"/usr/lib64/nsbrowser/plugins\"  -fno-rtti -fno-handle-exceptions -Wall -Wconversion -Wpointer-arith -Wcast-align -Woverloaded-virtual -Wsynth -Wno-ctor-dtor-privacy -Wno-non-virtual-dtor -Wno-long-long -march=athlon64 -pipe -fPIC -Wno-return-type -w
-fshort-wchar -pthread -pipe  -DNDEBUG -DTRIMMED -ffunction-sections -O1  -DGENTOO_NSPLUGINS_DIR=\"/usr/lib64/nsplugins\" -DGENTOO_NSBROWSER_PLUGINS_DIR=\"/usr/lib64/nsbrowser/plugins\"  -DMOZILLA_CLIENT -include ../../../mozilla-config.h -Wp,-MD,.deps/nsFreeType.pp nsFreeType.cpp
nsFreeType.cpp:101: error: cast from FT_Error (**)(FT_FaceRec_*) to int loses precision
nsFreeType.cpp:102: error: cast from FT_Error (**)(FT_LibraryRec_*) to int loses precision
nsFreeType.cpp:103: error: cast from FT_Error (**)(FT_GlyphRec_*) to int loses precision
nsFreeType.cpp:104: error: cast from FT_Error (**)(FT_FaceRec_*, FT_ULong) to int loses precision
nsFreeType.cpp:105: error: cast from FT_Error (**)(FT_GlyphSlotRec_*, FT_GlyphRec_**) to int loses precision
nsFreeType.cpp:106: error: cast from void* (**)(FT_FaceRec_*, FT_Sfnt_Tag) to int loses precision
nsFreeType.cpp:107: error: cast from FT_Error (**)(FT_GlyphRec_*, FT_UInt, FT_BBox*) to int loses precision
nsFreeType.cpp:108: error: cast from FT_Error (**)(FT_LibraryRec_**) to int loses precision
nsFreeType.cpp:109: error: cast from FT_Error (**)(FT_FaceRec_*, FT_UInt, FT_Int) to int loses precision
nsFreeType.cpp:110: error: cast from FT_Error (**)(FT_LibraryRec_*, const char*, FT_Long, FT_FaceRec_**) to int loses precision
nsFreeType.cpp:111: error: cast from FT_Error (**)(FT_Outline*, const FT_Outline_Funcs*, void*) to int loses precision
nsFreeType.cpp:112: error: cast from FT_Error (**)(FT_FaceRec_*, FT_UInt, FT_UInt) to int loses precision
nsFreeType.cpp:113: error: cast from FT_Error (**)(FT_FaceRec_*, FT_CharMapRec_*) to int loses precision
nsFreeType.cpp:115: error: cast from FT_Error (**)(FT_GlyphRec_*, FT_Matrix*, FT_Vector*) to int loses precision
nsFreeType.cpp:116: error: cast from FT_Error (**)(FT_FaceRec_*, FT_UInt, FT_UInt, FT_UInt, FT_Vector*) to int loses precision
nsFreeType.cpp:117: error: cast from FT_Error (**)(FT_GlyphRec_*, FT_GlyphRec_**) to int loses precision
nsFreeType.cpp:118: error: cast from FT_Error (**)(FT_GlyphRec_**, FT_Render_Mode, FT_Vector*, FT_Bool) to int loses precision
nsFreeType.cpp:120: error: cast from FT_ULong (**)(FT_FaceRec_*, FT_UInt*) to int loses precision
nsFreeType.cpp:121: error: cast from FT_ULong (**)(FT_FaceRec_*, FT_ULong, FT_UInt*) to int loses precision
gmake[4]: ** [nsFreeType.o] Erro 1
gmake[4]: Leaving directory `/var/tmp/portage/mozilla-firefox-1.0.6-r7/work/mozilla/gfx/src/freetype'
gmake[3]: ** [libs] Erro 2
gmake[3]: Leaving directory `/var/tmp/portage/mozilla-firefox-1.0.6-r7/work/mozilla/gfx/src'
gmake[2]: ** [libs] Erro 2
gmake[2]: Leaving directory `/var/tmp/portage/mozilla-firefox-1.0.6-r7/work/mozilla/gfx'
gmake[1]: ** [tier_9] Erro 2
gmake[1]: Leaving directory `/var/tmp/portage/mozilla-firefox-1.0.6-r7/work/mozilla'
make: ** [default] Erro 2

!!! ERROR: www-client/mozilla-firefox-1.0.6-r7 failed.
!!! Function src_compile, Line 192, Exitcode 2
!!! (no error message)
!!! If you need support, post the topmost build error, NOT this status message.

Post by **Halcy0n** » Thu Sep 22, 2005 4:45 pm

gralves wrote:I'm having some trouble compiling firefox. I think it is the same problem I have w/ blender - pointer to int conversions, on amd64, loose precision. There will probably be a lot of packages w/ this kind of problem.

Code: Select all

x86_64-pc-linux-gnu-g++ -o nsFreeType.o -c -DOSTYPE="Linux2.6" -DOSARCH="Linux" -I../.. -I../../../dist/include/xpcom -I../../../dist/include/string -I../../../dist/include/pref -I../../../dist/include/uconv -I../../../dist/include/unicharutil -I../../../dist/include/gfx -I../../../dist/include -I/var/tmp/portage/mozilla-firefox-1.0.6-r7/work/mozilla/dist/include/nspr     -I/usr/include/freetype2    -fPIC  -DGENTOO_NSPLUGINS_DIR="/usr/lib64/nsplugins" -DGENTOO_NSBROWSER_PLUGINS_DIR="/usr/lib64/nsbrowser/plugins"  -fno-rtti -fno-handle-exceptions -Wall -Wconversion -Wpointer-arith -Wcast-align -Woverloaded-virtual -Wsynth -Wno-ctor-dtor-privacy -Wno-non-virtual-dtor -Wno-long-long -march=athlon64 -pipe -fPIC -Wno-return-type -w
-fshort-wchar -pthread -pipe  -DNDEBUG -DTRIMMED -ffunction-sections -O1  -DGENTOO_NSPLUGINS_DIR="/usr/lib64/nsplugins" -DGENTOO_NSBROWSER_PLUGINS_DIR="/usr/lib64/nsbrowser/plugins"  -DMOZILLA_CLIENT -include ../../../mozilla-config.h -Wp,-MD,.deps/nsFreeType.pp nsFreeType.cpp
nsFreeType.cpp:101: error: cast from FT_Error (**)(FT_FaceRec_*) to int loses precision
nsFreeType.cpp:102: error: cast from FT_Error (**)(FT_LibraryRec_*) to int loses precision
nsFreeType.cpp:103: error: cast from FT_Error (**)(FT_GlyphRec_*) to int loses precision
nsFreeType.cpp:104: error: cast from FT_Error (**)(FT_FaceRec_*, FT_ULong) to int loses precision
nsFreeType.cpp:105: error: cast from FT_Error (**)(FT_GlyphSlotRec_*, FT_GlyphRec_**) to int loses precision
nsFreeType.cpp:106: error: cast from void* (**)(FT_FaceRec_*, FT_Sfnt_Tag) to int loses precision
nsFreeType.cpp:107: error: cast from FT_Error (**)(FT_GlyphRec_*, FT_UInt, FT_BBox*) to int loses precision
nsFreeType.cpp:108: error: cast from FT_Error (**)(FT_LibraryRec_**) to int loses precision
nsFreeType.cpp:109: error: cast from FT_Error (**)(FT_FaceRec_*, FT_UInt, FT_Int) to int loses precision
nsFreeType.cpp:110: error: cast from FT_Error (**)(FT_LibraryRec_*, const char*, FT_Long, FT_FaceRec_**) to int loses precision
nsFreeType.cpp:111: error: cast from FT_Error (**)(FT_Outline*, const FT_Outline_Funcs*, void*) to int loses precision
nsFreeType.cpp:112: error: cast from FT_Error (**)(FT_FaceRec_*, FT_UInt, FT_UInt) to int loses precision
nsFreeType.cpp:113: error: cast from FT_Error (**)(FT_FaceRec_*, FT_CharMapRec_*) to int loses precision
nsFreeType.cpp:115: error: cast from FT_Error (**)(FT_GlyphRec_*, FT_Matrix*, FT_Vector*) to int loses precision
nsFreeType.cpp:116: error: cast from FT_Error (**)(FT_FaceRec_*, FT_UInt, FT_UInt, FT_UInt, FT_Vector*) to int loses precision
nsFreeType.cpp:117: error: cast from FT_Error (**)(FT_GlyphRec_*, FT_GlyphRec_**) to int loses precision
nsFreeType.cpp:118: error: cast from FT_Error (**)(FT_GlyphRec_**, FT_Render_Mode, FT_Vector*, FT_Bool) to int loses precision
nsFreeType.cpp:120: error: cast from FT_ULong (**)(FT_FaceRec_*, FT_UInt*) to int loses precision
nsFreeType.cpp:121: error: cast from FT_ULong (**)(FT_FaceRec_*, FT_ULong, FT_UInt*) to int loses precision
gmake[4]: ** [nsFreeType.o] Erro 1
gmake[4]: Leaving directory `/var/tmp/portage/mozilla-firefox-1.0.6-r7/work/mozilla/gfx/src/freetype'
gmake[3]: ** [libs] Erro 2
gmake[3]: Leaving directory `/var/tmp/portage/mozilla-firefox-1.0.6-r7/work/mozilla/gfx/src'
gmake[2]: ** [libs] Erro 2
gmake[2]: Leaving directory `/var/tmp/portage/mozilla-firefox-1.0.6-r7/work/mozilla/gfx'
gmake[1]: ** [tier_9] Erro 2
gmake[1]: Leaving directory `/var/tmp/portage/mozilla-firefox-1.0.6-r7/work/mozilla'
make: ** [default] Erro 2

!!! ERROR: www-client/mozilla-firefox-1.0.6-r7 failed.
!!! Function src_compile, Line 192, Exitcode 2
!!! (no error message)
!!! If you need support, post the topmost build error, NOT this status message.

Grr, I thought I checked for and fixed these already. I'll look into it later at work.

gralves · Post by **gralves** » Thu Sep 22, 2005 5:26 pm

Halcy0n wrote:
gralves wrote:I'm having some trouble compiling firefox. I think it is the same problem I have w/ blender - pointer to int conversions, on amd64, loose precision. There will probably be a lot of packages w/ this kind of problem.
Code: Select all
!!! ERROR: www-client/mozilla-firefox-1.0.6-r7 failed.
!!! Function src_compile, Line 192, Exitcode 2
!!! (no error message)
!!! If you need support, post the topmost build error, NOT this status message.
Grr, I thought I checked for and fixed these already. I'll look into it later at work.

By the way, what is the correct fix for these cases? I think that just changing the int for long would impose performance hits on non-amd64 machines.

chtephan · Post by **chtephan** » Thu Sep 22, 2005 8:31 pm

long is 32 bit on x86 systems.

While it's not written anywhere a lot of programs are fine with the assumption sizeof(int) <= sizeof(void*) <= sizeof(long), including the kernel. On most(all?) current systems int is 32 bit, long is the actual register size, and long long is usually 64 bit.

Post by **Halcy0n** » Thu Sep 22, 2005 9:42 pm

chtephan wrote:long is 32 bit on x86 systems.

While it's not written anywhere a lot of programs are fine with the assumption sizeof(int) <= sizeof(void*) <= sizeof(long), including the kernel. On most(all?) current systems int is 32 bit, long is the actual register size, and long long is usually 64 bit.

This is normally the case, but is not necessarily guaranteed by the actual definitions. For most cases, this assumption should be correct though.

luca82 · Post by **luca82** » Fri Sep 23, 2005 4:21 pm

After upgrade from gcc-4.0.1 to gcc.4.0.2_pre20050917 I have noted that there is only gcj but no gij and gcj-dbtool... I have compiled with +gcj flag. What's changed?

luca82 · Post by **luca82** » Fri Sep 23, 2005 6:02 pm

luca82 wrote:After upgrade from gcc-4.0.1 to gcc.4.0.2_pre20050917 I have noted that there is only gcj but no gij and gcj-dbtool... I have compiled with +gcj flag. What's changed?

Solved. I have recompiled gcc-4.0.2 with himself and now I have gij and gcj-dbtool.

chrroessner · Post by **chrroessner** » Sat Sep 24, 2005 6:56 am

Short question:

I compiled gcc, emerged the system, revdep-rebuilt, kernel + external stuff, prelinked. Do I have to do more than this or is the main part finished with this?

Thanks

Christian

N.B.: I had big problems with gcc, because I had the use flag vanilla set in gcc, so PIE was not built and gcc did not know -fno-stack-protector. This made problems with perl and glibc. After recompiling gcc, glibc finished successfully (perl not, yet). Maybe someone could check flag-o-matic.eclass or whatever to have this in mind

Xake · Post by **Xake** » Sat Sep 24, 2005 11:09 am

Can't compile perl, but that has something to do with gdbm.h It compilains about not finding it and then gives me a lot of errors.

I have also lately had some other problems, but have not had the time to check up if they're known or not.
Just got a ner apartment and a new connection..
Had some abstinence after being offline for about three weeks....

chrroessner · Post by **chrroessner** » Sat Sep 24, 2005 1:05 pm

Fix for perl-5.8.7-r1:

After emerge is breaking:
vi /var/tmp/portage/perl-5.8.7-r1/work/perl-5.8.7/ext/GDBM_File/GDBM_File.xs +5
You will see a line #include <gdbm.h>
Change it to #include <gdbm/gdbm.h>
After that:
FEATURES="keeptemp keepwork" emerge perl

Fixed it for me.

Xake · Post by **Xake** » Sat Sep 24, 2005 2:31 pm

Revdump for gdbm to fix it. Seems like the ebuild for gdbm installed it to the wrong place.

Bonkie · Post by **Bonkie** » Mon Sep 26, 2005 12:31 pm

mjpegtools-1.6.2-r4 fails with gcc version 4.0.2-pre20050917 (Gentoo 4.0.2_pre20050917, pie-8.7.

Code: Select all

make[2]: Entering directory `/var/tmp/portage/mjpegtools-1.6.2-r4/work/mjpegtools-1.6.2/mpeg2enc'
if /bin/sh ../libtool --mode=compile i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I.. -I.. -I../utils    -mcpu=i686 -march=i686 -march=athlon-xp -O2 -pipe -Wall -Wunused -MT transfrm_x86.lo -MD -MP -MF ".deps/transfrm_x86.Tpo" \
  -c -o transfrm_x86.lo `test -f 'transfrm_x86.c' || echo './'`transfrm_x86.c; \
then mv -f ".deps/transfrm_x86.Tpo" ".deps/transfrm_x86.Plo"; \
else rm -f ".deps/transfrm_x86.Tpo"; exit 1; \
fi
 i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I.. -I.. -I../utils -mcpu=i686 -march=i686 -march=athlon-xp -O2 -pipe -Wall -Wunused -MT transfrm_x86.lo -MD -MP -MF .deps/transfrm_x86.Tpo -c transfrm_x86.c  -fPIC -DPIC -o .libs/transfrm_x86.o
`-mcpu=' is deprecated. Use `-mtune=' or '-march=' instead.
{standard input}: Assembler messages:
{standard input}:57: Error: suffix or operands invalid for `movq'
{standard input}:58: Error: suffix or operands invalid for `movq'
{standard input}:61: Error: suffix or operands invalid for `movq'
{standard input}:89: Error: suffix or operands invalid for `movq'
{standard input}:94: Error: suffix or operands invalid for `movq'
{standard input}:114: Error: suffix or operands invalid for `movq'
{standard input}:119: Error: suffix or operands invalid for `movq'
{standard input}:126: Error: suffix or operands invalid for `movq'
{standard input}:131: Error: suffix or operands invalid for `movq'
{standard input}:144: Error: suffix or operands invalid for `movq'
{standard input}:149: Error: suffix or operands invalid for `movq'
{standard input}:166: Error: suffix or operands invalid for `movq'
{standard input}:171: Error: suffix or operands invalid for `movq'
{standard input}:178: Error: suffix or operands invalid for `movq'
{standard input}:183: Error: suffix or operands invalid for `movq'
{standard input}:194: Error: suffix or operands invalid for `movq'
{standard input}:205: Error: suffix or operands invalid for `movq'
make[2]: *** [transfrm_x86.lo] Error 1
make[2]: Leaving directory `/var/tmp/portage/mjpegtools-1.6.2-r4/work/mjpegtools-1.6.2/mpeg2enc'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/var/tmp/portage/mjpegtools-1.6.2-r4/work/mjpegtools-1.6.2'
make: *** [all] Error 2

Cant find patch, fix or info, all I found was this post in the mailing list, although cvs should compile with gcc4 (as ive read).
Sorry if it is a known issue

Exner · Post by **Exner** » Mon Sep 26, 2005 12:52 pm

dev-ml/lablgtk-2.4.0 fails with sys-devel/gcc-4.0.2_pre20050917

chrroessner · Post by **chrroessner** » Mon Sep 26, 2005 1:05 pm

I did a revdep-rebuild --library libstdc++.so.6 and recompiled about 200 apps. The following ones died:

app-accessibility/festival-1.4.3-r4
app-accessibility/speech-tools-1.2.3-r3
app-benchmarks/acovea-4.0.0
app-mobilephone/gsmlib-1.11_pre041028
dev-cpp/gconfmm-2.0.2
dev-php/php-4.4.0-r1
dev-python/PyQt-3.14.1
media-libs/lcms-1.14-r1
media-plugins/gst-plugins-ffmpeg-0.8.3
media-sound/alsa-tools-1.0.9
media-sound/kamix-0.6.1
media-video/avifile-0.7.43.20050224-r2
media-video/mjpegtools-1.6.2-r4
media-video/transcode-0.6.14-r2
net-im/gnomemeeting-1.2.2
net-libs/gecko-sdk-1.7.8
net-libs/openh323-1.15.6
net-mail/courier-imap-4.0.4

gcc (GCC) 4.0.2-pre20050917 (Gentoo 4.0.2_pre20050917, pie-8.7.

Regards

Christian

chtephan · Post by **chtephan** » Mon Sep 26, 2005 2:02 pm

> app-mobilephone/gsmlib-1.11_pre041028
../gsmlib/gsm_map_key.h:82: error: explicit qualification in declaration of `bool gsmlib::operator<(const gsmlib::MapKey<SortedStore>&, const gsmlib::MapKey<SortedStore>&)'

> dev-cpp/gconfmm-2.0.2
Works for me.

> dev-php/php-4.4.0-r1
Used to work for me, currently using dev-lang/php-4.4.0-r1

> media-libs/lcms-1.14-r1
Works for me.

> media-plugins/gst-plugins-ffmpeg-0.8.3
Works for me.

> media-sound/alsa-tools-1.0.9
Confirmed. Invalid LValue.

> media-video/avifile-0.7.43.20050224-r2
FFReadHandler.cpp:69: error: cannot convert 'AVCodecContext**' to 'AVCodecContext*' in initialization

> media-video/mjpegtools-1.6.2-r4
{standard input}:57: Error: suffix or operands invalid for `movq'

> media-video/transcode-0.6.14-r2
Invalid LValue again.

> net-libs/gecko-sdk-1.7.8
Works for me.

> net-mail/courier-imap-4.0.4
Works for me.

liboil has a fix pending, don't know about libsdl.

Post by **Halcy0n** » Mon Sep 26, 2005 2:13 pm

chtephan wrote:liboil has a fix pending, don't know about libsdl.

We have a patch up on the libsdl bug, http://bugs.gentoo.org/show_bug.cgi?id=104437 that you all should give a shot and let us know if it breaks anything.

chrroessner · Post by **chrroessner** » Mon Sep 26, 2005 10:10 pm

Hi,

I tested again, after updating to gcc-4.0.2rc2 and updates from today (27.09.2005). Here are my results for the ones that are still broken:

Code: Select all

dev-cpp/gconfmm-2.0.2:

/var/tmp/portage/gconfmm-2.0.2/work/gconfmm-2.0.2/gconf/gconfmm/

client.cc:185: error: cast from 'void*' to 'int' loses precision
client.cc:188: error: cast from 'void*' to 'gboolean' loses precision

I removed it, because it seems, no other app is using it (2.12.0 is installed, too)

Code: Select all

media-plugins/gst-plugins-ffmpeg-0.8.3:

/var/tmp/portage/gst-plugins-ffmpeg-0.8.3/work/gst-ffmpeg-0.8.3/gst-libs/ext/ffmpeg/libavcodec

 x86_64-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I.. -DHAVE_AV_CONFIG_H=1 -Wall -O2 -march=athlon64 -pipe -fomit-frame-pointer -fno-pie -fno-pie -MT common.lo -MD -MP -MF .deps/common.Tpo -c common.c  -fPIC -DPIC -o .libs/common.o
In file included from avcodec.h:15,
                 from common.c:29:
common.h:69: error: array type has incomplete element type
common.h:71: error: array type has incomplete element type

Regards
Christian

chtephan · Post by **chtephan** » Tue Sep 27, 2005 7:38 am

The libavcodec in that gst-plugins-ffmpeg version isn't read for gcc4, the version in media-plugins/gst-plugins-ffmpeg-0.8.6 is.

chrroessner · Post by **chrroessner** » Tue Sep 27, 2005 7:49 am

There you can see that I have problems understanding portage´s emerge command: Look at the following

Code: Select all

croessner@amd64 ~ $ eix gst-plugins-ffmpeg
* media-plugins/gst-plugins-ffmpeg
     Available versions:  0.8.3 0.8.4 0.8.4-r1 0.8.5 0.8.6
     Installed:           0.8.3
     Homepage:            http://gstreamer.freedesktop.org/modules/gst-ffmpeg.html
     Description:         FFmpeg based gstreamer plugin


Found 1 matches
croessner@amd64 ~ $ emerge -uDp world

These are the packages that I would merge, in order:

Calculating world dependencies ...done!
[ebuild  N    ] www-client/mozilla-1.7.12-r1
[ebuild     U ] net-print/libgnomecups-0.2.2 [0.2.0-r2]
[ebuild     U ] gnome-base/libgnomeprint-2.12.1 [2.12.0]
[ebuild     U ] dev-python/PyQt-3.15 [3.14.1]
[ebuild     U ] app-arch/p7zip-4.27 [4.20-r1]
[ebuild  N    ] media-libs/urt-3.1b
[ebuild     U ] media-libs/netpbm-10.29-r1 [10.29]
[ebuild     U ] sys-apps/initng-0.2.5 [0.2.3]

First I did an eix and found out that I do not have the latest version of gst-plugins-ffmpeg. Right after that I did an -uDp world. So why does emerge not show an update for this?

Thanks

Christian

Bonkie · Post by **Bonkie** » Tue Sep 27, 2005 8:53 am

why dont you try

Code: Select all

emerge =media-plugins/gst-plugins-ffmpeg-0.8.6 -vp

and let portage tell you why it wont install this version

chrroessner · Post by **chrroessner** » Tue Sep 27, 2005 9:03 am

I have explicitly tried your suggestion:

Code: Select all

croessner@amd64 ~ $ emerge =media-plugins/gst-plugins-ffmpeg-0.8.6 -vp


These are the packages that I would merge, in order:

Calculating dependencies ...done!
[ebuild     U ] media-plugins/gst-plugins-ffmpeg-0.8.6 [0.8.3] 1,992 kB

As you can see, I could install the app without problems. My question was: Why does emerge -uDp world not provide an update for the app, while emerge =class/app-rev does? I know that I am not the only one who has discovered this problem. But I guess this is an other story and does not belong to the gcc-4 thread, does it?

Thanks anyway

Christian

chrroessner · Post by **chrroessner** » Tue Sep 27, 2005 2:18 pm

Fixes for

Code: Select all

dev-cpp/gconfmm-2.0.2:

/var/tmp/portage/gconfmm-2.0.2/work/gconfmm-2.0.2/gconf/gconfmm/

client.cc:185: error: cast from 'void*' to 'int' loses precision
client.cc:188: error: cast from 'void*' to 'gboolean' loses precision

Code: Select all

185         gconf_value_set_int(v,*(int*)i->data);
188         gconf_value_set_bool(v,*(gboolean*)i->data);

Regards

Christian

zomps · Post by **zomps** » Tue Sep 27, 2005 6:19 pm

Bonkie wrote:mjpegtools-1.6.2-r4 fails with gcc version 4.0.2-pre20050917 (Gentoo 4.0.2_pre20050917, pie-8.7.

Code: Select all

make[2]: Entering directory `/var/tmp/portage/mjpegtools-1.6.2-r4/work/mjpegtools-1.6.2/mpeg2enc'
if /bin/sh ../libtool --mode=compile i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I.. -I.. -I../utils    -mcpu=i686 -march=i686 -march=athlon-xp -O2 -pipe -Wall -Wunused -MT transfrm_x86.lo -MD -MP -MF ".deps/transfrm_x86.Tpo" \
  -c -o transfrm_x86.lo `test -f 'transfrm_x86.c' || echo './'`transfrm_x86.c; \
then mv -f ".deps/transfrm_x86.Tpo" ".deps/transfrm_x86.Plo"; \
else rm -f ".deps/transfrm_x86.Tpo"; exit 1; \
fi
 i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I.. -I.. -I../utils -mcpu=i686 -march=i686 -march=athlon-xp -O2 -pipe -Wall -Wunused -MT transfrm_x86.lo -MD -MP -MF .deps/transfrm_x86.Tpo -c transfrm_x86.c  -fPIC -DPIC -o .libs/transfrm_x86.o
`-mcpu=' is deprecated. Use `-mtune=' or '-march=' instead.
{standard input}: Assembler messages:
{standard input}:57: Error: suffix or operands invalid for `movq'
{standard input}:58: Error: suffix or operands invalid for `movq'
{standard input}:61: Error: suffix or operands invalid for `movq'
{standard input}:89: Error: suffix or operands invalid for `movq'
{standard input}:94: Error: suffix or operands invalid for `movq'
{standard input}:114: Error: suffix or operands invalid for `movq'
{standard input}:119: Error: suffix or operands invalid for `movq'
{standard input}:126: Error: suffix or operands invalid for `movq'
{standard input}:131: Error: suffix or operands invalid for `movq'
{standard input}:144: Error: suffix or operands invalid for `movq'
{standard input}:149: Error: suffix or operands invalid for `movq'
{standard input}:166: Error: suffix or operands invalid for `movq'
{standard input}:171: Error: suffix or operands invalid for `movq'
{standard input}:178: Error: suffix or operands invalid for `movq'
{standard input}:183: Error: suffix or operands invalid for `movq'
{standard input}:194: Error: suffix or operands invalid for `movq'
{standard input}:205: Error: suffix or operands invalid for `movq'
make[2]: *** [transfrm_x86.lo] Error 1
make[2]: Leaving directory `/var/tmp/portage/mjpegtools-1.6.2-r4/work/mjpegtools-1.6.2/mpeg2enc'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/var/tmp/portage/mjpegtools-1.6.2-r4/work/mjpegtools-1.6.2'
make: *** [all] Error 2

Cant find patch, fix or info, all I found was this post in the mailing list, although cvs should compile with gcc4 (as ive read).
Sorry if it is a known issue

there is new version out now 1.8.0 which seems to fix it, least for me

Bonkie · Post by **Bonkie** » Tue Sep 27, 2005 6:51 pm

zomps wrote:there is new version out now 1.8.0 which seems to fix it, least for me

Yes, mjpegtools-1.8.0 successfully compiles for me as well ! Thanx for sharing

SteveYin · Post by **SteveYin** » Tue Sep 27, 2005 6:56 pm

gcc mail list says gcc 4.0.2 will be released today or tommorow..