| Message ID | 201209071611.q87GB48n013212@glazunov.sibelius.xs4all.nl |
|---|---|
| State | New |
| Headers | show |
OK, thanks. Jason
> Date: Fri, 07 Sep 2012 14:12:29 -0400 > From: Jason Merrill <jason@redhat.com> > > OK, thanks. I don't have commit access (or at least I'm not on the write after approval list). Would you (or somebody else) be so kind to commit this fix for me? Thanks, Mark
On Sat, 8 Sep 2012, Mark Kettenis wrote: > I don't have commit access (or at least I'm not on the write after > approval list). Would you (or somebody else) be so kind to commit > this fix for me? I'll take care of it. Gerald
Index: decl.c =================================================================== --- decl.c (revision 191075) +++ decl.c (working copy) @@ -5131,7 +5131,7 @@ if (field_init == error_mark_node) return error_mark_node; - if (d->cur->index && d->cur == old_cur) + if (d->cur == old_cur && d->cur->index) { /* This can happen with an invalid initializer for a flexible array member (c++/54441). */
The following change (r190962): 2012-09-04 Jason Merrill <jason@redhat.com> PR c++/54441 * decl.c (reshape_init_class): Handle invalid initializer for 0-length array member. introduces a poinential dereference of d->end. I hit this when bootstrapping on OpenBSD/i386. Looks Like this change didn't take into account that reshape_init_r() might move d->cur past the end of the list of fields. Potential fix below. Simply reordering the checks in the if-statement avoids the problem. gcc/cp: 2012-09-07 Mark Kettenis <kettenis@openbsd.org> * decl.c (reshape_init_class): Avoid dereferencing a past-the-end pointer.