https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120591
--- Comment #1 from David Malcolm <dmalcolm at gcc dot gnu.org> --- Looking at a recent gcc.sum on my own machine for that test, I see: PASS: gcc.dg/sarif-output/bad-pragma.c (test .sarif output against SARIF 2.1 schema) PASS: gcc.dg/sarif-output/bad-pragma.c (test .sarif output for UTF-8-encoded parseable JSON) PASS: gcc.dg/sarif-output/bad-pragma.c (test for excess errors) PASS: gcc.dg/sarif-output/bad-pragma.c bad-pragma.py::test_messages_have_embedded_url I'm not sure if I'm reading comment #0 correctly, but it looks like in both configurations the "name" of the test result doesn't contain the builddir. It looks like the test "names" are: gcc.dg/sarif-output/bad-pragma.c bad-pragma.py::test_messages_have_embedded_urls - assert "unknown opti...s.html)' .. and gcc.dg/sarif-output/bad-pragma.c bad-pragma.py::test_messages_have_embedded_urls - assert "unknown opti...s.html)' ki.. where seem to be the same up to to the trailing ' character. Is the ellipsizing of the message happening in compare_tests, or in what's in the .sum or .log file. Do the .sum and .log files have the full message? Is the second result perhaps saying "killed"? Can you attach a .sum and .log file giving more information? Thanks
