------- Comment #9 from rguenth at gcc dot gnu dot org 2010-07-20 14:21 -------
The issue is with the cache for outcomes of gimple_types_compatible_p. When
hitting cycles in the type graph we assume equality but also record final
decisions based on it.
So we end up merging the two struct link instances, but only at the 2nd
time visiting it. IIRC this is a similar reason as to why we introduced
SCC visiting for the hashing part.
I have a more simpler workaround.
--
rguenth at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
AssignedTo|unassigned at gcc dot gnu |rguenth at gcc dot gnu dot
|dot org |org
Status|NEW |ASSIGNED
Last reconfirmed|2010-04-23 16:03:35 |2010-07-20 14:21:28
date| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42451
