------- Comment #4 from ktietz at gcc dot gnu dot org 2010-05-26 20:00 -------
(In reply to comment #3)
> (In reply to comment #1)
> > Hmm, is here rtti in use? Could you please test following patch, if it
> > solves
> > your bootstrap issue?
> >
> > Kai
> >
> > Index: gcc/gcc/cp/rtti.c
> > ===================================================================
> > --- gcc.orig/gcc/cp/rtti.c 2010-05-26 17:54:18.000000000 +0200
> > +++ gcc/gcc/cp/rtti.c 2010-05-26 21:46:53.066961700 +0200
> > @@ -1502,6 +1502,8 @@ emit_support_tinfos (void)
> > tree types[3];
> > int i;
> >
> > + if (bltn == NULL_TREE)
> > + continue;
> > types[0] = bltn;
> > types[1] = build_pointer_type (bltn);
> > types[2] = build_pointer_type (cp_build_qualified_type (bltn,
> >
>
> I don't think so since it also failed in ibgfortran.
>
hmm, this should be the real reason here.
Index: gcc/gcc/c-lex.c
===================================================================
--- gcc.orig/gcc/c-lex.c 2010-05-21 20:16:07.000000000 +0200
+++ gcc/gcc/c-lex.c 2010-05-26 21:58:52.839130300 +0200
@@ -480,7 +480,11 @@ narrowest_unsigned_type (unsigned HOST_W
for (; itk < itk_none; itk += 2 /* skip unsigned types */)
{
- tree upper = TYPE_MAX_VALUE (integer_types[itk]);
+ tree upper;
+
+ if (integer_types[itk] == NULL_TREE)
+ continue;
+ upper = TYPE_MAX_VALUE (integer_types[itk]);
if ((unsigned HOST_WIDE_INT) TREE_INT_CST_HIGH (upper) > high
|| ((unsigned HOST_WIDE_INT) TREE_INT_CST_HIGH (upper) == high
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44287
