This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: 988 GCC regressions, 971 new, with your patch on 2002-06-10T23:06:50Z.
- From: Roger Sayle <roger at eyesopen dot com>
- To: <gcc-patches at gcc dot gnu dot org>, <gcc-regression at gcc dot gnu dot org>
- Cc: <echristo at redhat dot com>, <hp at bitrange dot com>, <jakub at redhat dot com>, <ljrittle at acm dot org>, <rth at redhat dot com>, <zack at codesourcery dot com>
- Date: Mon, 10 Jun 2002 21:27:41 -0600 (MDT)
- Subject: Re: 988 GCC regressions, 971 new, with your patch on 2002-06-10T23:06:50Z.
On Tue, 11 Jun 2002, GCC regression checker wrote:
> With your recent patch, GCC has some regression test failures, which
> used to pass. There are 971 new failures, and 17 failures that existed
> before and after that patch; 2 failures have been fixed.
My sincerest apologies. I've now convinced myself that it was infact
me who has caused the recent disaster. Despite testing my change to
rest_of_compilation multiple times, and indeed getting Andreas to run
the SPECcpu2000 benchmarks on it, I noticed that I'd missed an "obvious"
change. Now that we were no longer running early if-conversion, surely
I could remove the test for "flag_if_conversion". I bootstrapped the
compiler to show that this was safe, but didn't actually run the GCC
testsuite. Mea culpa. Running the test-suite with "-fno-if-conversion"
would have shown the same failures. Obviously, we need to run cleanup_cfg
or find_basic_blocks at this point, even if we don't do if_conversion
or delete_null_pointer_checks.
To return the test suite to some form of normality, I've commited the
following reversion as obvious, but obviously broken. Jan can you come
up with the correct fix for this problem? I really hope this is the
true cause of the problem and I haven't made a fool of myself (again).
Waiting for the bootstraps and regression tests to complete to find out!
Sorry once again for any inconvenience this may have caused.
2002-06-10 Roger Sayle <roger@eyesopen.com>
* toplev.c (rest_of_compilation): Revert flag_if_conversion change.
Index: toplev.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/toplev.c,v
retrieving revision 1.647
diff -c -3 -p -r1.647 toplev.c
*** toplev.c 10 Jun 2002 22:33:07 -0000 1.647
--- toplev.c 11 Jun 2002 03:14:57 -0000
*************** rest_of_compilation (decl)
*** 2723,2729 ****
timevar_push (TV_JUMP);
! if (flag_delete_null_pointer_checks)
{
open_dump_file (DFI_null, decl);
find_basic_blocks (insns, max_reg_num (), rtl_dump_file);
--- 2723,2729 ----
timevar_push (TV_JUMP);
! if (flag_delete_null_pointer_checks || flag_if_conversion)
{
open_dump_file (DFI_null, decl);
find_basic_blocks (insns, max_reg_num (), rtl_dump_file);
Roger
--
Roger Sayle, E-mail: roger@eyesopen.com
OpenEye Scientific Software, WWW: http://www.eyesopen.com/
Suite 1107, 3600 Cerrillos Road, Tel: (+1) 505-473-7385
Santa Fe, New Mexico, 87507. Fax: (+1) 505-473-0833