On Linux/x86, revision 206621 gave: FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -fcilkplus -g -O0 -std=c99 (internal compiler error) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -fcilkplus -g -O0 -std=c99 (test for excess errors) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -fcilkplus (internal compiler error) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -fcilkplus -O2 -std=c99 (internal compiler error) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -fcilkplus -O2 -std=c99 (test for excess errors) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -fcilkplus -O3 -std=c99 (internal compiler error) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -fcilkplus -O3 -std=c99 (test for excess errors) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -fcilkplus -std=c99 (internal compiler error) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -fcilkplus -std=c99 (test for excess errors) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -fcilkplus (test for excess errors) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -g -fcilkplus (internal compiler error) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -g -fcilkplus (test for excess errors) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -g -O2 -ftree-vectorize -fcilkplus (internal compiler error) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -g -O2 -ftree-vectorize -fcilkplus (test for excess errors) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -O1 -fcilkplus (internal compiler error) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -O1 -fcilkplus (test for excess errors) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -O2 -fcilkplus (internal compiler error) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -O2 -fcilkplus (test for excess errors) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -O3 -fcilkplus (internal compiler error) FAIL: c-c++-common/cilk-plus/AN/rank_mismatch2.c -O3 -fcilkplus (test for excess errors) Revision 206615 is OK.
It is caused by r206620.
That is known, I've said that explicitly in the patch submission that AN is broken and will need to be fixed.
Given Cilk+ is new in 4.9 the test failures are not a regression.
(In reply to Jeffrey A. Law from comment #3) > Given Cilk+ is new in 4.9 the test failures are not a regression. Meh, I see you subscribe to the release-maintainer-definition; still they passed before, so it's a regression on trunk, even if not compared to a release. (An autotester doesn't care whether the reason they passed is right or wrong, it just keeps shouting at me...) Anyway, apparently all targets see this, pain dutily shared. Thanks to H.J. for the PR.
It's certainly my mindset right now.. But fully understand the complaints you're getting from the autotester :-)
Author: bviyer Date: Thu Jan 23 17:00:53 2014 New Revision: 206991 URL: http://gcc.gnu.org/viewcvs?rev=206991&root=gcc&view=rev Log: Fix for PR c/59825. 2014-01-23 Balaji V. Iyer <balaji.v.iyer@intel.com> PR c/59825 * c-array-notation.c (expand_array_notation_exprs): Rewrote this function to use walk_tree and moved a lot of its functionality to expand_array_notations. (expand_array_notations): New function. Modified: trunk/gcc/c/ChangeLog trunk/gcc/c/c-array-notation.c
Should be resolved by Balaji's trunk commit.