Pádraig Brady <[email protected]> writes: > > This is even more expensive as I just noticed a false failure on cfarm13 > due to the delay in sync'ing the whole file system. > So I'll apply this to both increase the timeout and exclude the check > unless RUN_VERY_EXPENSIVE_TESTS is enabled. > > cheers, > Padraig > > diff --git a/tests/misc/sync.sh b/tests/misc/sync.sh > index fa2ca86fa..df715ee6e 100755 > --- a/tests/misc/sync.sh > +++ b/tests/misc/sync.sh > @@ -61,8 +61,12 @@ if test "$fail" != '1'; then > # Ensure a fifo doesn't block > mkfifo_or_skip_ fifo > for opt in '' '-f' '-d'; do > - test "$opt" = '-f' && test "$RUN_EXPENSIVE_TESTS" != yes && continue > - returns_ 124 timeout 10 sync $opt fifo && fail=1 > + timeout=10 > + if test "$opt" = '-f'; then > + test "$RUN_VERY_EXPENSIVE_TESTS" = yes || continue > + timeout=60 > + fi > + returns_ 124 timeout $timeout sync $opt fifo && fail=1 > done > fi
Sounds good to me. I'll change my recently sent patch to be "very expensive" instead of "expensive" as well [1]. It is less finicky since I'm checking if sync is called, not that the program doesn't hang. Will hold it until after release regardless. Collin [1] https://lists.gnu.org/archive/html/coreutils/2026-02/msg00003.html
