This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Patch for gcc.dg/tree-ssa/ivopts-2.c failure in LP64 mode.


Steve Ellcey writes:
> It probably wouldn't hurt to run it on ILP32 platforms since I believe
> it will always pass there (because longs and ints are the same size),
> but it is failing on LP64 IA64 and X86_64 platforms.  If the failure on
> these platforms isn't a bug, then we shouldn't just xfail it.

I looked at x86_64 and the bug is in the scan-dump part of the
testcase.  We want to match the final candidate line before the line
about the final cost.  My regexp was not doing this correctly since I
used .* which could potentially match other candidate lines.

The patch below fixes this.  The test passes on mipsisa64-elf and
x86_64-linux but fails on mipsisa64-elf before the ivopts fix.

Steve, thanks for noticing this and sorry about the mistake!  Can you
check ia64?

OK to apply?

Adam

	* gcc.dg/tree-ssa/ivopts-2.c: Match final candidates line only.

Index: ivopts-2.c
===================================================================
--- ivopts-2.c	(revision 117733)
+++ ivopts-2.c	(working copy)
@@ -20,5 +20,5 @@ f ()
 }
 
 /* We should only have two IVs.  */
-/* { dg-final { scan-tree-dump-not "\\n  candidates .*,.*,.*\\n\[^\\n\]*\\nFinal cost" "ivopts" } }  */
+/* { dg-final { scan-tree-dump-not "\\n  candidates \\d+, \\d+,\[^\\n\]*\\n\[^\\n\]*\\nFinal cost" "ivopts" } }  */
 /* { dg-final { cleanup-tree-dump "ivopts" } }  */


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]