Created attachment 857411 [details] build.log.xz (ppc, 12.2.1_p20230121-r1) === gcc Summary === # of expected passes 15778 # of unexpected failures 46 # of unexpected successes 24 # of expected failures 227 # of unresolved testcases 1 [...] === gcc Summary === # of expected passes 17895 # of unexpected failures 51 # of unexpected successes 3 # of expected failures 268 # of unresolved testcases 3 [...] === gcc Summary === # of expected passes 21685 # of unexpected failures 33 # of unexpected successes 1 # of expected failures 216 # of unresolved testcases 1 [...] === gcc Summary === # of expected passes 18633 # of unexpected failures 111 # of expected failures 131 # of unsupported tests 460 [...] === gcc Summary === # of expected passes 18262 # of unexpected failures 86 # of unexpected successes 5 # of expected failures 237 # of unresolved testcases 7 [...] === gcc Summary === # of expected passes 30030 # of unexpected failures 88 # of unexpected successes 1 # of expected failures 146 # of unsupported tests 354 [...] === gcc Summary === # of expected passes 28049 # of unexpected failures 49 # of expected failures 196 # of unsupported tests 634
Created attachment 857413 [details] emerge --info
Created attachment 857433 [details] build.log.xz (ppc, 11.3.1_p20230120-r1)
It's expected to see test failures (we haven't yet made some 'reasonable subset' of tests that should always pass), as the GCC suite contains some wishful thinking tests. What really ought to be tested with GCC is regressions, hence the GCC_TESTS_NO_IGNORE_BASELINE=1 suggestion from the message.
Running contrib/compare_tests on gcc 11 vs 12 may be interesting.
(In reply to Sam James from comment #4) > Running contrib/compare_tests on gcc 11 vs 12 may be interesting. Hm, what's the best way to do that?
(In reply to ernsteiswuerfel from comment #5) > (In reply to Sam James from comment #4) > > Running contrib/compare_tests on gcc 11 vs 12 may be interesting. > Hm, what's the best way to do that? If you have the workdir of both, go into one of them (ideally 11), then run contrib/compare_tests /path/to/11/workdir /path/to/12/workdir.
Created attachment 857695 [details] compare_tests results (ppc, 11.3.1_p20230120-r1 vs. 12.2.1_p20230121-r1) (In reply to Sam James from comment #6) > If you have the workdir of both, go into one of them (ideally 11), then run > contrib/compare_tests /path/to/11/workdir /path/to/12/workdir. This does indeed show some interesting differences: # Comparing directories ## Dir1=/var/tmp/portage/sys-devel/gcc-11.3.1_p20230120-r1/work/: 8 sum files ## Dir2=/var/tmp/portage/sys-devel/gcc-12.2.1_p20230121-r1/work/: 8 sum files # Comparing 8 common sum files ## /bin/sh ./contrib/compare_tests /tmp/gxx-sum1.20865 /tmp/gxx-sum2.20865 Tests that now fail, but worked before (15 tests): c-c++-common/Wstringop-overflow-2.c -Wc++-compat (test for warnings, line 226) c-c++-common/Wstringop-overflow-2.c -Wc++-compat (test for warnings, line 227) c-c++-common/Wstringop-overflow-2.c -std=gnu++14 (test for warnings, line 226) c-c++-common/Wstringop-overflow-2.c -std=gnu++14 (test for warnings, line 227) c-c++-common/Wstringop-overflow-2.c -std=gnu++17 (test for warnings, line 226) c-c++-common/Wstringop-overflow-2.c -std=gnu++17 (test for warnings, line 227) c-c++-common/Wstringop-overflow-2.c -std=gnu++98 (test for warnings, line 226) c-c++-common/Wstringop-overflow-2.c -std=gnu++98 (test for warnings, line 227) g++.dg/warn/Warray-bounds-16.C -std=gnu++14 pr102690 (test for bogus messages, line 22) g++.dg/warn/Warray-bounds-16.C -std=gnu++17 pr102690 (test for bogus messages, line 22) g++.dg/warn/Warray-bounds-16.C -std=gnu++98 pr102690 (test for bogus messages, line 22) gcc.dg/Wstringop-overflow-43.c (test for warnings, line 176) gcc.dg/guality/pr54693-2.c -O2 -flto -fuse-linker-plugin -fno-fat-lto-objects -DPREVENT_OPTIMIZATION line 21 x == 10 - i gcc.dg/uninit-15.c (test for warnings, line 13) gcc.target/powerpc/float128-cmp2-runnable.c (test for excess errors) Tests that now work, but didn't before (129 tests): [...]
