Sergey Bronnikov <este...@gmail.com> writes:
> Imho test should FAIL in case coredump was detected.
I don't think that's a good idea; the current test (checking the exit
status of the child make process) looks enough to me. If a test doesn't
fail when a program dumps core then either the test should be fixed...
or it may do so on purpose.
> --- bsd.regress.mk_ Sun Jun 30 15:32:02 2013
> +++ bsd.regress.mk__ Sun Jun 30 15:46:04 2013
> @@ -98,7 +98,7 @@
> # XXX - we need a better method to see if a test fails due to timeout or just
> # normal failure.
> . if !defined(REGRESS_MAXTIME)
> - ${_SKIP_FAIL}if cd ${.CURDIR} && ${MAKE} ${RT}; then \
> + ${_SKIP_FAIL}if cd ${.CURDIR} && ${MAKE} ${RT} && ! (test -e *.core);
> then \
~$ ! (test -e a.core b.core); echo $?
ksh: test: b.core: unexpected operator/operand
0
~$
Also core dumps may not sit in the current directory.
> echo -n "SUCCESS " ${_REGRESS_OUT} ; \
> else \
> echo -n "FAIL " ${_REGRESS_OUT} ; \
Regards,
--
Jérémie Courrèges-Anglas
PGP Key fingerprint: 61DB D9A0 00A4 67CF 2A90 8961 6191 8FBF 06A1 1494
