| Message ID | 20151126181234.GL5675@tucnak.redhat.com |
|---|---|
| State | New |
| Headers | show |
On November 26, 2015 7:12:34 PM GMT+01:00, Jakub Jelinek <jakub@redhat.com> wrote: >Hi! > >When upgrading a gcc tree < r230888 with objdir to r230888 or later >and just doing make in the gcc subdir, I'm getting errors - genmatch >complains that it doesn't know IFN_FMAX. >The problem seems to be a missing dependency, build/gencfn-macros >depends on internal-fn.def and thus cfn-operators.pd is updated, >but build/genmatch does not depend on it, even when it includes >internal-fn.def, so has not been rebuilt and fails to parse >the new cfn-operators.pd. >Fixed thusly, ok for trunk? OK. Richard. >2015-11-26 Jakub Jelinek <jakub@redhat.com> > > * Makefile.in (build/genmatch.o): Depend on internal-fn.def. > >--- gcc/Makefile.in.jj 2015-11-18 11:19:26.000000000 +0100 >+++ gcc/Makefile.in 2015-11-26 18:15:24.254279664 +0100 >@@ -2598,7 +2598,7 @@ build/genmddump.o : genmddump.c $(RTL_BA > coretypes.h $(GTM_H) errors.h $(READ_MD_H) gensupport.h > build/genmatch.o : genmatch.c $(BCONFIG_H) $(SYSTEM_H) \ > coretypes.h errors.h $(HASH_TABLE_H) hash-map.h $(GGC_H) is-a.h \ >- tree.def builtins.def >+ tree.def builtins.def internal-fn.def > build/gencfn-macros.o : gencfn-macros.c $(BCONFIG_H) $(SYSTEM_H) \ >coretypes.h errors.h $(HASH_TABLE_H) hash-set.h builtins.def >internal-fn.def > > > Jakub
--- gcc/Makefile.in.jj 2015-11-18 11:19:26.000000000 +0100 +++ gcc/Makefile.in 2015-11-26 18:15:24.254279664 +0100 @@ -2598,7 +2598,7 @@ build/genmddump.o : genmddump.c $(RTL_BA coretypes.h $(GTM_H) errors.h $(READ_MD_H) gensupport.h build/genmatch.o : genmatch.c $(BCONFIG_H) $(SYSTEM_H) \ coretypes.h errors.h $(HASH_TABLE_H) hash-map.h $(GGC_H) is-a.h \ - tree.def builtins.def + tree.def builtins.def internal-fn.def build/gencfn-macros.o : gencfn-macros.c $(BCONFIG_H) $(SYSTEM_H) \ coretypes.h errors.h $(HASH_TABLE_H) hash-set.h builtins.def internal-fn.def