On Thu, 2013-08-29 at 12:20 -0400, David Malcolm wrote: > This patch is 110KB in size, so to avoid mailing-list size limits I've > uploaded it to: > > http://dmalcolm.fedorapeople.org/gcc/large-patches/a89d361b4f95dd216e1d29cb44fbaf90372c48b8-0003-Autogenerated-conversion-of-gimple-to-C.patch > > The ChangeLog entry and diffstat follow: > > gcc/ > > Patch autogenerated by refactor_gimple.py from > https://github.com/davidmalcolm/gcc-refactoring-scripts > revision 26aabff11750d29e6129930a426242a564538d1a
I've fixed the bug with OMP_TASK mentioned in http://gcc.gnu.org/ml/gcc-patches/2013-08/msg01934.html by regenerating this patch with a fixed refactor_gimple.py [1] Again, it's 110KB in size, exceeding the list limit, so I've uploaded it here: http://dmalcolm.fedorapeople.org/gcc/large-patches/733a06e20496464e373c1f48388cedf0b6935a4e-0003-Autogenerated-conversion-of-gimple-to-C.patch The only effective difference between the patches is this: +inline bool +is_a_helper <gimple_statement_omp_parallel>::test (gimple gs) +{ -+ return gs->code == GIMPLE_OMP_PARALLEL; ++ return gs->code == GIMPLE_OMP_PARALLEL || gs->code == GIMPLE_OMP_TASK; +} + (and the const equivalent). Successfully bootstrapped on x86_64-unknown-linux-gnu with plain "configure" and thus with checking enabled; all tests show same results as a unpatched control build (of r202029), fixing the omp ICEs described above. Dave [1] Specifically, this commit to my refactoring scripts: https://github.com/davidmalcolm/gcc-refactoring-scripts/commit/d55c7a3e3d1227bad36284ffdb67c83fb089b272