[Bug middle-end/52650] [4.8 Regression] FAIL: gcc.dg/torture/pr51106-2.c * (internal compiler error)

[rguenther at suse dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52650

--- Comment #4 from rguenther at suse dot de <rguenther at suse dot de> 2012-03-21 13:38:28 UTC ---
On Wed, 21 Mar 2012, ro at CeBiTec dot Uni-Bielefeld.DE wrote:

> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52650
> 
> --- Comment #3 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> 2012-03-21 13:32:31 UTC ---
> > I don't yet know if they are related.
> 
> Now I do: I've just built a 32-bit C-only compiler on
> sparc-sun-solaris2.11 at r185563 (i.e. immediately before the culprit
> patch), and both the ICE in gcc.dg/torture/pr51106-2.c and the other
> execution failures reported are gone.

If they are related (which I doubt), removing the call to cleanup_cfg
in cfgexpand.c should fix them.  That's the only real change in
this patch - previously the cleanup_cfg call was conditional on a
present EH region tree (thus, all of these issues are latent and
just require a slightly different testcase)