Building on powerpc-linux-gnu

Ian Piumarta ian.piumarta@inria.fr
Tue Oct 15 21:12:00 GMT 2002


Folks,

The current CVS tree is broken on the above platform.  Attached are cvs
diffs that make it work for me.

The example in the README compiles and runs fine.  I can't run the
testsuite since configure doesn't seem to want to build it and I'm too
lazy to fight over it.  (Forgive me if this issue has already been
resolved, since I'm also too lazy to trawl through the list archives.)

In case it matters (the last line almost certainly does):

  emilia$ gcc -v
  Reading specs from /usr/lib/gcc-lib/ppc-yellowdog-linux/2.95.4/specs
  gcc version 2.95.4 20010319 (prerelease/franzo/20011204)
  emilia$ uname -a
  Linux emilia.inria.fr 2.4.18-0.8a #2 Sun May 26 06:00:20 CEST 2002 ppc
  emilia$ autoconf -V
  autoconf (GNU Autoconf) 2.53

Regards,
Ian


Index: include/ffi.h
===================================================================
RCS file: /cvs/libffi/libffi/include/ffi.h,v
retrieving revision 1.3
diff -r1.3 ffi.h
304c304
< #elif defined(POWERPC)
---
> #elif defined(__powerpc__)
Index: configure.in
===================================================================
RCS file: /cvs/libffi/libffi/configure.in,v
retrieving revision 1.11
diff -r1.11 configure.in
86c86,87
< 	$LN_S $srcdir/powerpc/sysv.S t-asm.S;;
---
> 	$LN_S $srcdir/powerpc/sysv.S t-asm.S;
> 	$LN_S $srcdir/powerpc/ppc_closure.S ppc_closure.S;;
145c146,147
< AC_OUTPUT(include/Makefile testsuite/Makefile Makefile,
---
> AC_CONFIG_FILES(include/Makefile testsuite/Makefile Makefile)
> AC_CONFIG_COMMANDS([default],
147,150d148
< if test -n "$CONFIG_FILES"; then
<    ac_file=Makefile . ${libffi_basedir}/config-ml.in
< fi
< ],
167a166,170
> ],[
> if test -n "$CONFIG_FILES"; then
>    ac_file=Makefile . ${libffi_basedir}/config-ml.in
> fi
> ]
168a172
> AC_OUTPUT
Index: powerpc/sysv.S
===================================================================
RCS file: /cvs/libffi/libffi/powerpc/sysv.S,v
retrieving revision 1.1
diff -r1.1 sysv.S
31a32,33
> #include "ppc_closure.S"
> 



More information about the Libffi-discuss mailing list