On 7 August 2013 18:40, Nicholas Clark <n...@ccl4.org> wrote:
> If so, the fix would seem to be to change the new code to
>
> $runperl = $runperl . ($is_mswin ? ' <nul' : ' </dev/null');
>
> Does this make t/op/fork.t test 16 pass again on Win32?
> That now seems to be failing, and I don't know why.
>
I have that failing in blead anyway (i.e. without the changes in your
smoke-me branch), but (I thought) it only happens when I run the whole
test suite through: when I then run just op/fork.t afterwards to see
why it failed, it doesn't fail...
However, I've just tried out your branch plus the change you suggest
above as well and whilst it makes no difference to op/fork.t still
failing test 16, I have just found that I can reproduce it manually
afterwards now! Not sure if that's due to your changes or just random
luck, but here's the output while I see it anyway!:
C:\Dev\Git\perl\t>..\perl harness op\fork.t
op/fork.t .. 11/? PROG:
if (eval q{$pid = fork}) {
eval q{ die "parent died" };
print $@;
}
else {
eval q{ die "child died" };
print $@;
}
EXPECTED:
parent died at (eval 2) line 1.
child died at (eval 2) line 1.
GOT:
Attempt to free unreferenced scalar: SV 0x2d984c, Perl interpreter: 0x2d236c at
- line 1.
child died at (eval 2) line 1.
parent died at (eval 2) line 1.
# Failed test 16 - at op/fork.t line 16
op/fork.t .. Failed 1/26 subtests
(less 1 skipped subtest: 24 okay)
Test Summary Report
-------------------
op/fork.t (Wstat: 0 Tests: 26 Failed: 1)
Failed test: 16
Files=1, Tests=26, 13 wallclock secs ( 0.03 usr + 0.00 sys = 0.03 CPU)
Result: FAIL
