Re: Bug gcj with date manipulations?

[Maurice DIAMANTINI <diam at ensta dot fr>]

> Me tooIn-Reply-To: <3A9EF540.11596FD9@waitaki.otago.ac.nz>; from bryce@waitaki.otago.ac.nz on ven, mar 02, 2001 at 02:20:01 +1300



> Is GCC 3.0 working on MacOS X yet?

I don't use MacOS-X for the moment (thank for the patch information
about gcc-3.0 on macOS-X)



> > I rise an error with date format in the following :
> > Hier is the code :
> >
> 
> Your test case works for me, using "gcc version 3.0 20010219 (prerelease)" 
> on linux x86. It gets the timezone wrong (which is probibly 
> a bug in our TimeZone implementation), but it does not crash:

Great news!

------------------------------------------------------------------------
> Please try updating your cvs tree to the gcc-3_0-branch. 
> If you still get the problem, a gdb stack trace would help.

OK, So I did a :

    cvs update -r gcc-3_0-branch

and I tried (without success) to build the new gcc-3.0 
(on linuxppc).
As the trace of my result is rather long, I begin by my
main final questions:


What should I do?
-----------------

As My interest is in using java + tcljava for starting on a
(free) plateform scilab/matlab-like for combinatorial optimization,
I think java is an alternative to c++ thanks to gcj (I have to 
convince some people of the efficacity of the java solution)

Why should I do:
- working with jikes/javax + java solution while waiting a better
  gcc-3.0 update?
- trying harder to install tools needed (rpm? binutils? ...) to
  build gcc-3.0?
- looking for a precompiled version of a full install gcc-3.0
  tar.bz2 (something to put into my /usr/local/pack/ directory
  in a seperate sub-directory)?
- reinstalling a new version of linux to get an uptodate developpement
  environnement, then retry to build gc-3.0?
- ???

thank you very much for reading me!
(and excuse my bad english!)

regards


------------------------------------------------------------------------

Here is the compilation trace (after about 1h30 of compilation 
by a "make bootstrap" on 
- Hard (home): Macintosh PM7500 upgraded by G3/275MHz (128M RAM)
- system: linuxppc (1999) whih many modifications and kernel  2.2.17

/usr/bin/ld -v
GNU ld version 2.9.5 (with BFD 2.9.5.0.14)

    make[6]: Entering directory \
       
`/mnt/rab/work/compil/gcc_build/powerpc-unknown-linux-gnu/nof/libffi'
    /bin/sh ./libtool --mode=link
    /mnt/rab/work/compil/gcc_build/gcc/xgcc
    -B/mnt/rab/work/compil/gcc_build/gcc/
   
-B/usr/local/pack/gcc-3.0-cvs/powerpc-linux/powerpc-unknown-linux-gnu/bin/
   
-B/usr/local/pack/gcc-3.0-cvs/powerpc-linux/powerpc-unknown-linux-gnu/lib/
    -isystem
   
/usr/local/pack/gcc-3.0-cvs/powerpc-linux/powerpc-unknown-linux-gnu/include
    -msoft-float -fPIC -mstrict-align -fexceptions -g -O2
    -msoft-float -fPIC -mstrict-align  -msoft-float -fPIC
    -mstrict-align -o ffitest  ffitest.o libffi.la 
    
    /mnt/rab/work/compil/gcc_build/gcc/xgcc \
        -B/mnt/rab/work/compil/gcc_build/gcc/ \
       
-B/usr/local/pack/gcc-3.0-cvs/powerpc-linux/powerpc-unknown-linux-gnu/bin/
\
       
-B/usr/local/pack/gcc-3.0-cvs/powerpc-linux/powerpc-unknown-linux-gnu/lib/
\
        -isystem
/usr/local/pack/gcc-3.0-cvs/powerpc-linux/powerpc-unknown-linux-gnu/include
\
        -msoft-float -fPIC -mstrict-align -fexceptions \
        -g -O2 -msoft-float -fPIC -mstrict-align \
        -msoft-float -fPIC -mstrict-align \
        -o .libs/ffitest ffitest.o  ./.libs/libffi.so \
        -Wl,--rpath \
        -Wl,/usr/local/pack/gcc-3.0-cvs/powerpc-linux/lib/nof
        
    /usr/bin/ld: warning: libgcc_s_nof.so.0, needed by
./.libs/libffi.so, not found (try using --rpath)
    ./.libs/libffi.so: undefined reference to `__extendsfdf2@@GCC_3.0'
    collect2: ld returned 1 exit status
    make[6]: *** [ffitest] Error 1
    make[6]: Leaving directory
`/mnt/rab/work/compil/gcc_build/powerpc-unknown-linux-gnu/nof/libffi'

------------------------------------------------------------------------

Then I did:
cd /mnt/rab/work/compil/gcc_build/powerpc-unknown-linux-gnu/nof/libffi
nm ./.libs/libffi.so | grep __extendsfdf2
         U __extendsfdf2@@GCC_3.0
ll /mnt/rab/work/compil/gcc_build/gcc/libgcc_s_nof.so.0

lrwxrwxrwx 1 diam  uma  15 Mar  4 10:40 \
    /mnt/rab/work/compil/gcc_build/gcc/libgcc_s_nof.so.0 ->
libgcc_s_nof.so

So the file exists !

------------------------------------------------------------------------
Then I tryed to manually modyfing the last command modifing the 
"--rpath" like this:

/mnt/rab/work/compil/gcc_build/gcc/xgcc \
    -B/mnt/rab/work/compil/gcc_build/gcc/ \
   
-B/usr/local/pack/gcc-3.0-cvs/powerpc-linux/powerpc-unknown-linux-gnu/bin/
\
   
-B/usr/local/pack/gcc-3.0-cvs/powerpc-linux/powerpc-unknown-linux-gnu/lib/
\
    -isystem
/usr/local/pack/gcc-3.0-cvs/powerpc-linux/powerpc-unknown-linux-gnu/include
\
    -msoft-float -fPIC -mstrict-align -fexceptions \
    -g -O2 -msoft-float -fPIC -mstrict-align \
    -msoft-float -fPIC -mstrict-align \
    -o .libs/ffitest ffitest.o  ./.libs/libffi.so \
    -Wl,--rpath,/mnt/rab/work/compil/gcc_build/gcc/
    
Résultat : "/-/" is for  "/mnt/rab/work/compil/"

/-/gcc_build/gcc//libgcc_s_nof.so.0: undefined reference to
`__SBSS2_END__'
/-/gcc_build/gcc//libgcc_s_nof.so.0: undefined reference to `__init'
/-/gcc_build/gcc//libgcc_s_nof.so.0: undefined reference to
`__SBSS_END__'
/-/gcc_build/gcc//libgcc_s_nof.so.0: undefined reference to
`__SDATA2_START__'
/-/gcc_build/gcc//libgcc_s_nof.so.0: undefined reference to
`__SDATA_START__'
collect2: ld returned 1 exit status

So I give up this solution.
------------------------------------------------------------------------

I try to build gcc with a new  binutils but it doesn't work 
either :-(

I try to install a new rpm of binutils but my rpm utility is too 
old and does'nt support the new rpm packages :-((

I was unable to build the new rpm from the tar.gz :-((

Perhaps my linux distribution is to old? (linuxppc-99 modyfied) 

------------------------------------------------------------------------

-- 
Maurice.Diamantini@ensta.fr       -       ENSTA/LMA
École Nationale Supérieure  de  Techniques Avancées
 Laboratoire    de     Mathématiques    Appliquées 
     http://www.ensta.fr/~diam