Comparing http://gcc.gnu.org/ml/gcc-testresults/2008-04/msg02061.html and http://gcc.gnu.org/ml/gcc-testresults/2008-04/msg02120.html shows the following new failures on powerpc-apple-darwin*: FAIL: gcc.dg/memcpy-1.c scan-tree-dump-times optimized "nasty_local" 0 FAIL: gcc.dg/pr35729.c scan-rtl-dump-times loop2_invariant "Decided to move invariant" 0 ERROR: gcc.dg/var-expand3.c: error executing dg-final: no files matched glob pattern "var-expand3.c.[0-9][0-9][0-9]r.loop2_unroll" UNRESOLVED: gcc.dg/var-expand3.c: error executing dg-final: no files matched glob pattern "var-expand3.c.[0-9][0-9][0-9]r.loop2_unroll" between between revisions 134696 and 134717. The possible culprits seem to be 134697 (most likely) Author: uros Date: Sat Apr 26 13:43:13 2008 UTC (4 days ago) Log Message: * config/i386/i386.md (fix_trunc<mode>_i387_fisttp_with_temp): Use 'X' constraint for operand 2 when operand 0 is memory operand. (fix_truncdi_i387_with_temp): : Use 'X' constraint for operand 4 when operand 0 is memory operand. (fix_trunc<mode>_i387_with_temp): Ditto. (*floatsi<mode>2_vector_mixed_with_temp): Use 'X' constraint for operand 2 when operand 1 is memory operand. (*float<SSEMODEI24:mode><MODEF:mode>2_mixed_with_temp): Ditto. (*floatsi<mode>2_vector_sse_with_temp): Ditto. (*float<SSEMODEI24:mode><MODEF:mode>2_sse_with_temp): Ditto. (*float<SSEMODEI24:mode><X87MODEF:mode>2_i387_with_temp): Ditto. (floatdi<X87MODEF:mode>2_i387_with_xmm): Use 'X' constraint for operands 2,3 and 4 when operand 1 is memory operand. (fistdi2_with_temp): Use 'X' constraint for operand 2 when operand 0 is memory operand. (fistdi2_floor_with_temp): Ditto. (fist<mode>2_floor_with_temp): Ditto. (fistdi2_ceil_with_temp): Ditto. (fist<mode>2_ceil_with_temp): Ditto. (*truncdfsf_fast_mixed): Merge alternatives 0 and 1. or 134714 Author: simonb Date: Sat Apr 26 16:59:38 2008 UTC (3 days, 20 hours ago) Log Message: PR c/35652 * builtins.c (c_strlen): Suppressed multiple warnings that can occur with propagated string constants.
After reading again my post, I realized that rev. 134697 deals with i386 only and should not affect powerpc and I don't see how 134714 can be the cause. All the other revisions deal with branches or gfortran.
Those tests have been failing for quite awhile and are reported through http://gcc.gnu.org/ml/gcc-testresults/2008-04/msg01958.html. Starting in http://gcc.gnu.org/ml/gcc-testresults/2008-04/msg01978.html we see: ERROR: tcl error sourcing /Users/regress/tbox/svn-gcc/gcc/testsuite/gcc.dg/dg.exp. ERROR: unmatched open brace in list When that tcl error was fixed, the tests apparently started running again and the failures were reported again. It's a problem that the tests fail, but we should close this entry because the failures are unrelated. The gcc.dg/memcpy-1.c failure is reported in 31677.
>ERROR: gcc.dg/var-expand3.c: error executing dg-final: no files matched glob I reported this to the developer before: http://gcc.gnu.org/ml/gcc-patches/2008-03/msg01229.html
For gcc.dg/var-expand3.c I've got a patch to scandump.exp to get rid of the ERROR and report FAIL if the requested dump file doesn't exist. The test still fails because of the issue in comment #3, but more cleanly.
I also see the pr35729.c failure on hppa*-*-*. FAIL: gcc.dg/pr35729.c scan-rtl-dump-times loop2_invariant "Decided to move invariant" 0
Janis -- I don't understand this PR. It sounds like this should just be closed out, since the failures reported are long-standing, and, if necessary, new PRs opened for the other issues reported here. Is that correct? Thanks, -- Mark
Mark, the tests started failing because -fdump-rtl-loop2 used to produce dump files for all loop2_* passes. The compiler could be fixed to do that again, or the tests mentioned here could be changed to use -fdump-rtl-loop2_unroll and -fdump-rtl-loop2_invariant instead of -fdump-rtl-loop2. I was expecting someone to fix the compiler after Andrew complained about the change, but I wouldn't mind modifying the tests and closing this PR.
Subject: Re: [4.4 Regression] test failures between revs. 134696 and 134717 janis at gcc dot gnu dot org wrote: > ------- Comment #7 from janis at gcc dot gnu dot org 2008-08-08 23:34 ------- > Mark, the tests started failing because -fdump-rtl-loop2 used to produce dump > files for all loop2_* passes. The compiler could be fixed to do that again, or > the tests mentioned here could be changed to use -fdump-rtl-loop2_unroll and > -fdump-rtl-loop2_invariant instead of -fdump-rtl-loop2. I was expecting > someone to fix the compiler after Andrew complained about the change, but I > wouldn't mind modifying the tests and closing this PR. OK, I understand much better now, thanks! Do we have a separate PR open for the problem as well, or shall I retitle this one to more accurately reflect the problem? Thanks,
Subject: Bug 36087 Author: janis Date: Tue Aug 12 21:18:54 2008 New Revision: 139037 URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=139037 Log: PR testsuite/36087 * gcc.dg/var-expand3.c: Fix name of dump file. Modified: trunk/gcc/doc/invoke.texi trunk/gcc/testsuite/ChangeLog trunk/gcc/testsuite/gcc.dg/var-expand3.c
Closing as fixed. The memcpy issue is tracked separately, the pr35729.c failure should be filed in a less confusing bugreport.