https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124574
--- Comment #2 from Christopher Albert <albert at tugraz dot at> --- I ran a full local GCC regression comparison for the proposed sparse-set patch. Both builds used: --enable-languages=c,c++,fortran,lto --disable-multilib --disable-bootstrap --enable-valgrind-annotations CFLAGS='-Og -g' CXXFLAGS='-Og -g' I then ran `make -j32 -k check` on: - patched tree: sparseset-valgrind-fix-v2 - clean baseline: upstream/master at a0d6c3f23cc Normalized FAIL/XPASS outcome comparison of the full logs: - patched-only outcomes: 0 - baseline-only outcomes: 1 The single baseline-only outcome was: gfortran.dg/coarray/send_array.f90 -fcoarray=lib -pthread -O2 -lcaf_shmem -lrt execution test That baseline-only failure was a timeout in the full run. I reran caf.exp=send_array.f90 directly afterward in both the patched and baseline builds, and it passed in both cases, so this looks flaky/environmental rather than patch-related. So from this full local GCC regression comparison I do not see any regression attributable to the sparse-set patch.
