run-tests: produce error on running a failing test
This patch fixes a regression recently introduced by a refactoring (see
92a6b16c9186 and about 200 previous changesets from Gregory Szorc). It produce
an error message everytime with a test filename which gets fail while testing
except at one condition when '--nodiff' option is enabled.
$ "$TESTDIR/hghave" execbit || exit 80
$ hg init
$ echo a > a
$ hg ci -Am'not executable'
adding a
$ chmod +x a
$ hg ci -m'executable'
$ hg id
79abf14474dc tip
Make sure we notice the change of mode if the cached size == -1:
$ hg rm a
$ hg revert -r 0 a
$ hg debugstate
n 0 -1 unset a
$ hg status
M a
$ hg up 0
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg id
d69afc33ff8a
$ test -x a && echo executable -- bad || echo not executable -- good
not executable -- good