GTK has a some tests that produce output like
1..227
...
ok 57 /FinalizeObject/GtkConstraint
ok 58 /FinalizeObject/GtkConstraintGuide
ok 59 /FinalizeObject/GtkConstraintLayoutChild # SKIP Skipping GtkConstraintLayoutChild
ok 60 /FinalizeObject/GtkConstraintLayout
ok 61 /FinalizeObject/GtkCssProvider
...
ok 227 /FinalizeObject/GtkWindowHandle
that is, 227 ok's and an exit code of 0.
And meson 0.55.0 on my local system says:
3/3 gtk:gtk+object / objects-finalize FAIL 11.85s (exit status 0)
If I shorten the test to leave out all the cases that emit a # SKIP comment, meson 0.55 is happy again.
0.54 in our ci thinks everything is just fine.
I also think that everything with our testsuite is just fine, and meson regressed.
Bisection says this regressed in 0c3bb153.
Despite the fix being included in 0.55.1, this still doesn't seem to work. Any skipped tests still cause the test suite to fail. I believe this should be reopened.
Despite the above, it seems to have been fixed eventually in 0.55.3, so :shrug:
Despite the above, it seems to have been fixed eventually in 0.55.3, so 馃し
Yes, the discussion happened on https://github.com/mesonbuild/meson/pull/7618.
Most helpful comment
Despite the fix being included in 0.55.1, this still doesn't seem to work. Any skipped tests still cause the test suite to fail. I believe this should be reopened.