Bug#276291: wmkbd: FTBFS: g++-3.3 internal compiler error

2004-10-13 Thread Roland Stigge
Package: wmkbd, gcc-3.3
Version: 0.4.1-2
Severity: serious

Hi,

building the package wmkbd in a clean build environment
(with pbuilder) on i386 results in:

=
[...]
# Add here commands to compile the package.
/usr/bin/make
make[1]: Entering directory `/tmp/buildd/wmkbd-0.4.1'
g++ -g -O -I/usr/openwin/include -I/usr/local/include `glib-config --cflags` 
`gnome-config gnomeui --cflags` `gtk-config --cflags` -DXLIB_ILLEGAL_ACCESS 
app.c -c -o app.o
In file included from app.c:76:
keysyms.c:921: internal compiler error: in tree_low_cst, at tree.c:3255
Please submit a full bug report,
with preprocessed source if appropriate.
See http://gcc.gnu.org/bugs.html> for instructions.
For Debian GNU/Linux specific bug reporting instructions, see
.
make[1]: *** [app.o] Error 1
make[1]: Leaving directory `/tmp/buildd/wmkbd-0.4.1'
make: *** [build-stamp] Error 2
=

With gcc 3.2 an 3.4, it works.

Thanks for considering.

bye,
  Roland




size gains

2004-10-13 Thread reyes ray
Within a few days you should notice immediate erection size increases

90% of males were interested in improving their sexual stamina,
performance, and the size of their manhood. Are you one of the 90%?

I want to let u guys know that I have seen over 1 inch in length increase
since I started taking ur system. The exercises are easy too. I use them
both and this is awesome. Clancy, Spokane

check out the only Male Enhancement formula with a free DVD

http://www.supervalueproduct.com/lg/



i am busy, no thank you, go above




The effect of the electric shock conveyed by the tube was beginning to wear
away, and now the buccaneer sat up, rubbed his head in a bewildered fashion
and looked around him. When he saw Rob he gave a shout of rage and drew his
knife, but one motion of the electric tube made him cringe and slip away to
the cabin, where he remained out of danger
And now the other four sat up, groaning and muttering in their outlandish
speech; But they had no notion of facing Rob's tube a second time, so one by
one they joined their leader in the cabin, leaving the boy undisturbed




Heighten sensation

2004-10-13 Thread johnathan hart
Male enhancement is achieving your goals of becoming a better man

90% of males were interested in improving their sexual stamina,
performance, and the size of their manhood. Are you one of the 90%?

I cannot afford the ridiculous prescription costs since I have no health
insurance and am unemployed. Your discount rates on the Internet really
helped me, and your staff was very knowledgeable and helpful. I saw
immediate results, and now I have regained the confidence I had when I was
younger, in the bedroom. -Kenneth, Salt Lake City

check out the only Male Enhancement formula with a free DVD

http://www.superbitems.com/lg/



no, then link above




He followed the coast line, keeping but a short distance above the earth,
and after an hour's swift flight reached the city of San Francisco. His
shoulders were sore and stiff from the heavy strain upon them of the
previous day, and he wished more than once that he had some of his mother's
household liniment to rub them with
Yet so great was his delight at reaching once more his native land that all
discomforts were speedily forgotten




Re: cross-gcc patches

2004-10-13 Thread Nikita V. Youshchenko
Hello

> > 3. I'm having bad time with gcc-3.4 sparc multilib target. Current
> > debian/rules.conf contains code that changes TARGET_ALIAS from
> > sparc-linux to sparc64-linux. But this causes unwanted side-effects
> > (such as attempts to use 'sparc64-linux-as' instead of
> > 'sparc-linux-as' as assembler; that fails because there is no
> > 'sparc64-linux-as', so it falls back to 'as'; that also fails because
> > 'as' is native and can't process sparc assembler. Also, files go to
> > /usr/sparc64-linux/... instead of /usr/sparc-linux/...) If I comment
> > out TARGET_ALIAS change, buld fails with
> > In file included from ../../src/gcc/libgcc2.c:56:
> > ../../src/gcc/libgcc2.h:81: error: no data type for mode `TI'
> > ../../src/gcc/libgcc2.h:82: error: no data type for mode `TI'
> > make[4]: *** [libgcc/64/_muldi3.o] Error 1
> > Before I go deep into upstream code, I'd love to get some advice from
> > somebody who has better understanding of what that means and how
> > things work...
>
> I remember having stopped at the very same point and then found out
> about the configury done as above. CC'ed Clint and Ben.

I think I've found a fix to this.
See the attachement.
Adding a single line to gcc/config.gcc makes sparc target to build 
correctly, without having to change it to sparc64.

[EMAIL PROTECTED]:~/debian/gcc/3.4/test> dpkg-architecture -qDEB_HOST_ARCH
i386
[EMAIL PROTECTED]:~/debian/gcc/3.4/test> sparc-linux-gcc-3.4 -o h hello.c
[EMAIL PROTECTED]:~/debian/gcc/3.4/test> file h
h: ELF 32-bit MSB executable, SPARC, version 1 (SYSV), for GNU/Linux 2.2.0, 
dynamically linked (uses shared libs), not stripped
[EMAIL PROTECTED]:~/debian/gcc/3.4/test> sparc-linux-gcc-3.4 -m64 -o h64 
hello.c
[EMAIL PROTECTED]:~/debian/gcc/3.4/test> file h64
h64: ELF 64-bit MSB executable, SPARC V9, version 1 (SYSV), for GNU/Linux 
2.4.18, dynamically linked (uses shared libs), not stripped

I think the attached change (that is not cross-specific) should go in 
gcc-3.4 package.

With this change and the cross patch I've sent some days ago, gcc-3.4 
becomes in sync with gcc-3.3 (that means, builds for all debian linux 
targets, including biarch s390 and sparc).

Nikita
--- gcc-3.4/debian/patches/sparc64-build.dpatch	2004-10-04 10:54:41.0 +0400
+++ test/gcc-3.4/debian/patches/sparc64-build.dpatch	2004-10-13 18:35:05.0 +0400
@@ -37,7 +37,7 @@
 diff -urNad sparc64-build.gcc-3.3.tmp/src/gcc/config.gcc sparc64-build.gcc-3.3/src/gcc/config.gcc
 --- sparc64-build.gcc-3.3.tmp/src/gcc/config.gcc	2003-03-13 08:40:33.0 +
 +++ sparc64-build.gcc-3.3/src/gcc/config.gcc	2003-03-13 08:48:27.0 +
-@@ -2383,8 +2383,17 @@
+@@ -2383,8 +2383,18 @@
  	gnu_ld=yes
  	;;
  sparc-*-linux*)		# SPARC's running GNU/Linux, libc6
@@ -49,6 +49,7 @@
 +	tmake_file="t-slibgcc-elf-ver t-linux sparc/t-linux64 sparc/t-crtfm"
 +	tm_file="sparc/biarch64.h ${tm_file} dbxelf.h elfos.h svr4.h sparc/sysv4.h sparc/linux64.h"
 +	float_format=sparc
++	need_64bit_hwint=yes
 +	fi
 +	#tm_file="${tm_file} dbxelf.h elfos.h svr4.h sparc/sysv4.h sparc/linux.h"
 +	#tmake_file="t-slibgcc-elf-ver t-linux sparc/t-crtfm"
--- gcc-3.4/debian/rules.defs	2004-10-06 01:00:12.0 +0400
+++ test/gcc-3.4/debian/rules.defs	2004-10-13 23:30:12.0 +0400
@@ -559,9 +559,6 @@
   biarch := yes
   with_lib64gcc	:= yes
   with_lib64cxx	:= yes
-  ifeq ($(TARGET_ALIAS),sparc-linux)
-TARGET_ALIAS := sparc64-linux
-  endif
 endif
 
 # s390x build 


Re: cross-gcc patches

2004-10-13 Thread Matthias Klose
Nikita V. Youshchenko writes:
> Hello
> 
> > > 3. I'm having bad time with gcc-3.4 sparc multilib target. Current
> > > debian/rules.conf contains code that changes TARGET_ALIAS from
> > > sparc-linux to sparc64-linux. But this causes unwanted side-effects
> > > (such as attempts to use 'sparc64-linux-as' instead of
> > > 'sparc-linux-as' as assembler; that fails because there is no
> > > 'sparc64-linux-as', so it falls back to 'as'; that also fails because
> > > 'as' is native and can't process sparc assembler. Also, files go to
> > > /usr/sparc64-linux/... instead of /usr/sparc-linux/...) If I comment
> > > out TARGET_ALIAS change, buld fails with
> > > In file included from ../../src/gcc/libgcc2.c:56:
> > > ../../src/gcc/libgcc2.h:81: error: no data type for mode `TI'
> > > ../../src/gcc/libgcc2.h:82: error: no data type for mode `TI'
> > > make[4]: *** [libgcc/64/_muldi3.o] Error 1
> > > Before I go deep into upstream code, I'd love to get some advice from
> > > somebody who has better understanding of what that means and how
> > > things work...
> >
> > I remember having stopped at the very same point and then found out
> > about the configury done as above. CC'ed Clint and Ben.
> 
> I think I've found a fix to this.
> See the attachement.
> Adding a single line to gcc/config.gcc makes sparc target to build 
> correctly, without having to change it to sparc64.

thanks, please could you check if the testsuite is run for 32 and 64?

Matthias




Bug#276291: wmkbd: FTBFS: g++-3.3 internal compiler error

2004-10-13 Thread Matthias Klose
clone 276291 -1
reassign 276291 wmkbd
reassign -1 g++-3.3
severity -1 important
tags -1 + upstream
tags -1 + fixed-upstream
retitle -1 [fixed in 3.2/3.4] ICE building wmkbd
thanks

there's a workaround, so it's not release critical in the gcc-3.3
package.

Roland Stigge writes:
> Package: wmkbd, gcc-3.3
> Version: 0.4.1-2
> Severity: serious
> 
> Hi,
> 
> building the package wmkbd in a clean build environment
> (with pbuilder) on i386 results in:
> 
> =
> [...]
> # Add here commands to compile the package.
> /usr/bin/make
> make[1]: Entering directory `/tmp/buildd/wmkbd-0.4.1'
> g++ -g -O -I/usr/openwin/include -I/usr/local/include `glib-config --cflags` 
> `gnome-config gnomeui --cflags` `gtk-config --cflags` -DXLIB_ILLEGAL_ACCESS 
> app.c -c -o app.o
> In file included from app.c:76:
> keysyms.c:921: internal compiler error: in tree_low_cst, at tree.c:3255
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See http://gcc.gnu.org/bugs.html> for instructions.
> For Debian GNU/Linux specific bug reporting instructions, see
> .
> make[1]: *** [app.o] Error 1
> make[1]: Leaving directory `/tmp/buildd/wmkbd-0.4.1'
> make: *** [build-stamp] Error 2
> =
> 
> With gcc 3.2 an 3.4, it works.
> 
> Thanks for considering.
> 
> bye,
>   Roland
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]




linking with multiple versions of gcc installed

2004-10-13 Thread Brendan Simon
Hi,
I have an older powerpc machine that is basically running Woody (stable).
   gcc-2.95 installed as the default compiler.
   glibc-2.2 installed as the C libriary.
I have a newer powerpc machine that has Sarge (testing) installed on it.
   gcc-3.3 installed as the default compiler.
   glibc-2.3 installed as the C libriary.
   gcc-2.95 also installed.
The newer machine has mounted the /home directory of the older machine 
via NFS.
My project compiles and links fine on the old macihne using 'gcc' or 
'gcc-2.95'.
The new machine fails linking when using 'gcc-2.95' as the invoked command.
   eg. undefined references to '[EMAIL PROTECTED]'

BTW, my project is using glibc-2.2.5 which is stored under configuration 
management.  ie. None of the standard libraries or headers _SHOULD_ be 
used when linking or compiling etc.  I am using the -nostdlib switches, etc.

I think the problem is that the file 
/repository/glibc-2.2.5/usr/lib/libc.so has the following.
   GROUP ( /lib/libc.so.6 /usr/lib/libc_nonshared.a )
I'm assuming that this causes the local system files to be referenced 
which is not what I want.
I don't want to edit the libc.so as this file is copied to the end 
product filesystem.
Is the correct solution to use -rpath or more likely -rpath-link in my 
linking commands ???
If so, could some one give me some idea of the best way to use them.

Many thanks,
Brendan Simon.




Bug#275655: Bug in earlier versions too; not in 3.4

2004-10-13 Thread Chris Howie
I forgot to mention that earlier versions of gcc
(specifically, 3.0) caused the same problem.  I just
installed 3.4 and everything works perfectly, so I
guess this is a "<= 3.3" problem.

Also, it's no longer critical that this get resolved
soon since 3.4 works.  =)




___
Do you Yahoo!?
Declare Yourself - Register online to vote today!
http://vote.yahoo.com




Testing sparc target (was: Re: cross-gcc patches)

2004-10-13 Thread Nikita V. Youshchenko
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1


> Nikita V. Youshchenko writes:
> > Hello
> >
> > > > 3. I'm having bad time with gcc-3.4 sparc multilib target. Current
> > > > debian/rules.conf contains code that changes TARGET_ALIAS from
> > > > sparc-linux to sparc64-linux. But this causes unwanted
> > > > side-effects (such as attempts to use 'sparc64-linux-as' instead
> > > > of
> > > > 'sparc-linux-as' as assembler; that fails because there is no
> > > > 'sparc64-linux-as', so it falls back to 'as'; that also fails
> > > > because 'as' is native and can't process sparc assembler. Also,
> > > > files go to /usr/sparc64-linux/... instead of
> > > > /usr/sparc-linux/...) If I comment out TARGET_ALIAS change, buld
> > > > fails with
> > > > In file included from ../../src/gcc/libgcc2.c:56:
> > > > ../../src/gcc/libgcc2.h:81: error: no data type for mode `TI'
> > > > ../../src/gcc/libgcc2.h:82: error: no data type for mode `TI'
> > > > make[4]: *** [libgcc/64/_muldi3.o] Error 1
> > > > Before I go deep into upstream code, I'd love to get some advice
> > > > from somebody who has better understanding of what that means and
> > > > how things work...
> > >
> > > I remember having stopped at the very same point and then found out
> > > about the configury done as above. CC'ed Clint and Ben.
> >
> > I think I've found a fix to this.
> > See the attachement.
> > Adding a single line to gcc/config.gcc makes sparc target to build
> > correctly, without having to change it to sparc64.
>
> thanks, please could you check if the testsuite is run for 32 and 64?

Unfortunately I can't, because I don't have access to any sparc debian 
systems, and cros-running testsuite doesn't seem to be possible.

Probably somebody on the list may help? I'm reattaching the fix to make it 
available on list. The fix is not cross-specific, show thing to test is 
that it doesn't break anything for native sparc compilers.

> one more question, if you do have access to a sparc64, please could
> you build the current gcc-3.3 for unstable?

Again, I can't help, I don't have hardware access.

Nikita
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFBbgJzv3x5OskTLdsRAt0VAJ9ULMm+3IYBXZE4Y/3nh3YQ2TVS1ACgouHI
hmPH11zfalmRyehFDy4iteY=
=0hmo
-END PGP SIGNATURE-
--- gcc-3.4/debian/patches/sparc64-build.dpatch	2004-10-04 10:54:41.0 +0400
+++ test/gcc-3.4/debian/patches/sparc64-build.dpatch	2004-10-13 18:35:05.0 +0400
@@ -37,7 +37,7 @@
 diff -urNad sparc64-build.gcc-3.3.tmp/src/gcc/config.gcc sparc64-build.gcc-3.3/src/gcc/config.gcc
 --- sparc64-build.gcc-3.3.tmp/src/gcc/config.gcc	2003-03-13 08:40:33.0 +
 +++ sparc64-build.gcc-3.3/src/gcc/config.gcc	2003-03-13 08:48:27.0 +
-@@ -2383,8 +2383,17 @@
+@@ -2383,8 +2383,18 @@
  	gnu_ld=yes
  	;;
  sparc-*-linux*)		# SPARC's running GNU/Linux, libc6
@@ -49,6 +49,7 @@
 +	tmake_file="t-slibgcc-elf-ver t-linux sparc/t-linux64 sparc/t-crtfm"
 +	tm_file="sparc/biarch64.h ${tm_file} dbxelf.h elfos.h svr4.h sparc/sysv4.h sparc/linux64.h"
 +	float_format=sparc
++	need_64bit_hwint=yes
 +	fi
 +	#tm_file="${tm_file} dbxelf.h elfos.h svr4.h sparc/sysv4.h sparc/linux.h"
 +	#tmake_file="t-slibgcc-elf-ver t-linux sparc/t-crtfm"
--- gcc-3.4/debian/rules.defs	2004-10-06 01:00:12.0 +0400
+++ test/gcc-3.4/debian/rules.defs	2004-10-13 23:30:12.0 +0400
@@ -559,9 +559,6 @@
   biarch := yes
   with_lib64gcc	:= yes
   with_lib64cxx	:= yes
-  ifeq ($(TARGET_ALIAS),sparc-linux)
-TARGET_ALIAS := sparc64-linux
-  endif
 endif
 
 # s390x build