On Tue, Jun 28 2022, Han Xin wrote: > t/t5329-no-lazy-fetch-with-commit-graph.sh | 53 ++++++++++++++++++++++ This fails "make test" since 5b92477f896 (builtin/gc.c: conditionally avoid pruning objects via loose, 2022-05-20), i.e. we have another t/t5329* test now. Per $subject the CI output for this is now a bit cryptic, it lands you on a failed run-build-and-test.sh step with: [...] => Run tests cat: 't/test-results/*.exit': No such file or directory === Failed test: * === The full logs are in the 'print test failures' step below. See also the 'failed-tests-*' artifacts attached to this run. cat: 't/test-results/*.markup': No such file or directory Error: Process completed with exit code 1. The last line of the suggested next step is then: Build job failed before the tests could have been run Going earlier and expanding the "Run tests" step we can see the issue: duplicate test numbers: t5329 make[1]: *** [Makefile:86: test-lint-duplicates] Error 1 make[1]: *** Waiting for unfinished jobs.... make[1]: Leaving directory '/home/runner/work/git/git/t' make: *** [Makefile:3065: test] Error 2 + res=2 + rm exit.status + end_group + test -n t + set +x Since the CI topic in $subject we've ran the "print failures" step separately from the "make" invocation, and therefore have to guess at why we fail, whereas before we'd get that output from "make" itself. Johannes, is this something you can fix? In any case, for this topic the fix is simple: The test needs to be renamed for a re-roll,
