Subject: Re: aix gcc34 problem
To: Hubert Feyrer <>
From: Florian Heigl <>
List: tech-pkg
Date: 07/18/2005 03:51:05
2005/7/14, Hubert Feyrer <>:
> On Thu, 14 Jul 2005, Florian Heigl wrote:
> >> Maybe look at the named Makefile at line 777, and
> >> feel free to post here (with some context?).
> ...
> >  +775  # ??? The choices here will need some experimenting with.
> >  +776
> >  +777  export AR_FOR_TARGET
> I'd say this looks a lot like Shell commands that aren't indented, and
> esp. the blink line 776 probably breaks things altogether. Next step woul=
> be to see why this is... I guess that that Makefile is created from a
> by the configure script. To debug, you can re-run the
> "config.status" script, and see if it does the thing consistently (wrong
> ;).

1. It appears that the aix-supplied autoconf/automake combinations had
some problems.
I spent some time testing the results of building gcc outside the tree
with both gmake and (n)bmake, and at some point decided to rebuild the
autotools, and after that libtool which brought a large step ahead.

2. Also I had to find that rs6000-ibm-aix5.2.0.0 will define a powerpc
cpu instead of power, which is debatable as probably 80% of all rs6000
are based on power cpus. As I supplied -mcpu=3Dpower, the assembler
choked on the ambigous system definition it got from my mk.conf :)
i switched over to using power-ibm-aix5.2.0.0, which now appears to
define ppc64. which is in fact wrong (the cpu is a powerIII, not
really similar to a regular ppc64) but gets me working code.

3. as a side note: the binutils have another problem (file header of
/opt/pkg/bin/as has an alignment problem), but this is another topic.

4. I commented out c++/g77/objc/java in the top-level Makefile and now
really got a working gcc binary installed.

The next step was to rebuild gcc with the newly generated gcc, which
will break within libstc++, but I think this is a problem with gcc,
not pkgsrc (verified building outside of pkgsrc generates the same

echo regex.o cplus-dem.o cp-demangle.o cp-demint.o md5.o alloca.o
argv.o choose-temp.o concat.o dyn-string.o fdmatch.o fibheap.o
floatformat.o fnmatch.o getopt.o getopt1.o getpwd.o getruntime.o
hashtab.o hex.o lbasename.o lrealpath.o make-relative-prefix.o
make-temp-file.o objalloc.o obstack.o partition.o physmem.o pex-unix.o
safe-ctype.o sort.o spaces.o splay-tree.o strerror.o strsignal.o
ternary.o xatexit.o xexit.o xmalloc.o xmemdup.o xstrdup.o xstrerror.o
> required-list
gmake[5]: Entering directory
gmake[5]: Nothing to be done for `all'.
gmake[5]: Leaving directory
gmake[4]: Leaving directory
gmake[3]: Leaving directory
gmake[2]: Leaving directory
gmake[2]: Entering directory
gmake[2]: *** No rule to make target `all'.  Stop.
gmake[2]: Leaving directory
gmake[1]: *** [all-target-libstdc++-v3] Error 2
gmake[1]: Leaving directory `/ramobj/pkg/lang/gcc34/work/obj'
gmake: *** [bootstrap] Error 2
*** Error code 2

bmake: stopped in /usr/pkgsrc/lang/gcc34
=3D=3D=3D> There was an error during the ``build'' phase.
=3D=3D=3D> Please investigate the following for more information:
=3D=3D=3D>      * log of the build
=3D=3D=3D>      * /ramobj/pkg/lang/gcc34/work/.work.log
*** Error code 1

bmake: stopped in /usr/pkgsrc/lang/gcc34

For reference - some more people appear to see this problem:

I think I'll file a bug report for gcc... and ask You to see this for
resolved by rebuilding auto* and libtool.


'Sie brauchen sich um Ihre Zukunft keine Gedanken zu machen'