On Sun, Apr 20, 2025 at 02:12:34PM -0700, Junio C Hamano wrote: > > So I think we either need to rewrite the "run" script's fallback code, > > or teach the GIT-BUILD-OPTIONS writer to avoid mentioning unset > > variables (which is the real source of the problem in 4638e8806e3a). > > ... this was raised. And then I completely forgot about the topic, > as nothing happened since then. > > > ...yes, this is definitely the tip of the iceberg. I don't mind doing > > this patch as an incremental step forward (and because it is an > > improvement in behavior even if 4638e8806e3a were reverted). But the > > issue is far from solved overall. > > I do not mind it as an incremental band-aid. Looks like this is in 'jch' now, but there's a mis-merge in e8cf2b99cd (Merge branch 'ps/meson-build-perf-bench' into jch, 2025-04-21). The original patch in this thread did something like: git_perf_settings=$(...pull GIT_PERF_* from env...) . ../test-lib.sh eval "$git_perf_settings" That is, we stash away the environment, then load test-lib.sh, which overwrites the environment, and then we restore (some of) the original values. In that merge, the test-lib.sh inclusion is moved (and in fact is now accompanied by an explicit inclusion of GIT-BUILD-OPTIONS), and we now have: . "$GIT_BUILD_DIR"/GIT-BUILD-OPTIONS . "$GIT_SOURCE_DIR"/t/test-lib.sh git_perf_settings=$(...) eval "$git_perf_settings" Which of course does nothing. We need to set $git_perf_settings before those other source lines (and the eval must remain after them). -Peff
