[patch] logical OR loses side-effect

[Stuart Hastings]

It's rare I get such a nice bug.  This excellent testcase comes from a 
co-worker here at Apple:

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: 20050125-1.c
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20050126/d11a5eae/attachment.c>
-------------- next part --------------


It fails on TOT.

The failure seems to be in the gimplifier.  This patch:

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: gcc.fsf.diffs.txt
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20050126/d11a5eae/attachment.txt>
-------------- next part --------------


fixes the problem.

Bootstrapped C,C++,ObjC on powerpc-apple-darwin, DejaGnu, no 
regressions.

2005-01-25  Stuart Hastings  <stuart@apple.com>

	* gcc/gimplify.c (shortcut_cond_expr): Re-compute side-effects.
	* gcc/testsuite/gcc.c-torture/execute/20050125-1.c: New.

I'd guess the testcase would logically go to 
gcc/testsuite/gcc.c-torture/execute/20050125-1.c, or similar.

Do I need to file a PR in Bugzilla?

O.K. to commit?

stuart hastings
Apple Computer