coreutils-8.10 sparse-to-pipe failure

[Gilles Espinasse]

sparse-to-pipe test may fail when

cat pipe > copy &

did not finish before

cmp sparse copy

This happen randomly to me with debian 2.6.26 kernel (x86 32 bits) but I haven't had a failure on 3 other machines with different kernels (even 2.6.27 or 2.6.18)

I don't know what is the key to reproduce the test failure. I have two machines with mostly same CPU :

one with XP 3000+ and kernel 2.6.27 that never fail, one with XP 2800+ and kernel debian 2.6.26 that fail 80% of attempts.

Tested with

for i in {1..20}; do make -s check -C tests TESTS=cp/sparse-to-pipe 2>/dev/null;done | grep '1 test'

Upstream fix is to wait before cmp. A sed would be good enought to fix that

sed -i 's/^cmp sparse copy/wait; cmp sparse copy/' tests/cp/sparse-to-pipe

I let the milestone to 6.8 as the fix will be needed only for coreutils-8.10. I don't know why this happen on only one machine.

[bdubbs@…]

I'm inclined to close this as WONTFIX. We really can't chase every upstream bug. I'd be more inclined to make a fix for a security problem. This problem seems to be pretty esoteric and only in a test script.

[Gilles Espinasse]

I have no problem for that. I just reported this issue may happen and a simple fix is available. The previous test issue is more current. If you have ext3, the issue will happen.

There is another way to workaround the 2 test issue, by disabling the faulty tests.

sed -i -e '/cp\/sparse\-/d' -e '/cp\/fiemap/d' tests/Makefile.in

[bdubbs@…]

Fixed in revision 9497.

[(none)]

Milestone 6.9 deleted