[CVS 980420] Fixproto and stdlib.h - final(?-) PATCH included
manfred@s-direktnet.de
manfred@s-direktnet.de
Thu Apr 23 01:07:00 GMT 1998
On Tue, 21 April 1998, 13:06:41, ghazi@caip.rutgers.edu wrote:
> I still don't think this is right, you are only saying that
> libgcc.a cannot be created via "ar" until fixproto is run. Remember,
> a parallel make can build LIBGCC2 at the same time as STMP_FIXPROTO in
> the above dependency list.
>
> What you really need is to make sure the individual .o files
> making up the archive are not _compiled_ until fixproto is run. Note,
> if you set up the dependency, then your patch wouldn't include all the
> changes you had to all.cross, etc, as long as they already depend on
> LIBGCC, all should work.
>
> Maybe having libgcc2.ready depend on STMP_FIXPROTO is along the
> correct lines. But an expert should verify if this is sufficient. (I'm
> not sure it is complete.)
Yep, looks like libgcc2.ready is the right one. Since I don't have a
MP system available, I can't check, if it's actually sufficient;
but, IMHO this should work:
Thu Apr 23 09:41:47 1998 Manfred Hollstein <manfred@s-direktnet.de>
* Makefile.in (libgcc2.ready): Add explicit dependancy from
$(STMP_FIXPROTO) to ensure all necessary include files have
been created and to guarantee proper parallel builds.
diff -rup -x CVS -x RCS -x *.o -x *.info* -x *.html* -x *.elc -x *.dvi -x *.orig -x *~ -x version.el egcs-980422.orig/gcc/Makefile.in egcs-980422/gcc/Makefile.in
--- egcs-980422.orig/gcc/Makefile.in Sun Apr 19 15:38:34 1998
+++ egcs-980422/gcc/Makefile.in Wed Apr 22 18:44:40 1998
@@ -981,7 +981,7 @@ libgcc1.S: libgcc1.c $(CONFIG_H) config.
# Compiling libgcc2.a requires making sure that cc1, etc. have been compiled.
# But recompiling cc1 should not force recompilation of libgcc2.a.
# If you want to force recompilation, delete libgcc2.a.
-libgcc2.ready: $(GCC_PASSES) $(LIBGCC2_DEPS) stmp-int-hdrs
+libgcc2.ready: $(GCC_PASSES) $(LIBGCC2_DEPS) stmp-int-hdrs $(STMP_FIXPROTO)
-if [ -f libgcc2.ready ] ; then \
true; \
else \
More information about the Gcc-bugs
mailing list