tests: don't wait forever on GNU/Hurd

* tests/cp/parent-perm-race.sh: Add timeouts so that the test does
not wait forever on GNU/Hurd.  This does not fix the underlying
bug but at least lets the tests make progress.
Problem reported by Nelson H.F. Beebe in: http://bugs.gnu.org/22624
Also, check that 'cp' actually copies the data; this is a better
test anyway, and simplifies the test code.
Also, avoid unlikely race if 'pid' is set in the environment.
This commit is contained in:
Paul Eggert committed 2016-02-12 10:59:54 -08:00
1 parent 62e7af0326
commit 632eda520f
1 file changed
+8 -14
+8 -14
View File
@@ -26,7 +26,8 @@ umask 002
mkdir mode ownership d || framework_failure_
chmod g+s d 2>/dev/null # The cp test is valid either way.
# Terminate any background cp process
# Terminate any background cp process.
pid=
cleanup_() { kill $pid 2>/dev/null && wait $pid; }
for attr in mode ownership
@@ -35,21 +36,14 @@ do
# Copy a fifo's contents. That way, we can examine d/$attr's
# state while cp is running.
cp --preserve=$attr -R --copy-contents --parents $attr d & pid=$!
timeout 10 cp --preserve=$attr -R --copy-contents --parents $attr d & pid=$!
(
# Now 'cp' is reading the fifo.
# Check the permissions of the temporary destination
# directory that 'cp' has made.
ls -ld d/$attr >d/$attr.ls
# Check the permissions of the destination directory that 'cp' has made.
# 'ls' won't start until after 'cp' has made the destination directory
# $d/attr and has started to read the source file $attr/fifo.
timeout 10 sh -c "ls -ld d/$attr >$attr/fifo" || fail=1
# Close the fifo so that "cp" can continue. But output first,
# before exiting, otherwise some shells would optimize away the file
# descriptor that holds the fifo open.
echo foo
) >$attr/fifo
ls_output=$(cat d/$attr.ls) || fail=1
ls_output=$(cat d/$attr/fifo) || fail=1
case $attr,$ls_output in
ownership,d???--[-S]--[-S]* | \
mode,d????-??-?* | \