Front page | perl.perl5.porters |
Postings from September 2013
On Fri, Sep 20, 2013 at 02:22:09PM +0200, Andreas Koenig wrote:
> "Father Chrysostomos via RT" <perlbug-followup@perl.org> writes:
>
> > I get the bug on 5.8.8, but not 5.8.9. Bisecting will take a long time,
> > as it requires a human slave.
>
> No need for slaves, the answer is perl-5.8.8-990-g5f42405.
Or this one in blead
commit 48dbb59ed9af112b035bdcc063c1471a22ec25a2
Author: Steffen Ullrich <coyote.frank@gmx.net>
Date: Tue Feb 26 20:43:00 2008 +0100
Re: interrupting system() with signal depends on signal handler
Message-ID: <47C45DB4.9060306@gmx.net>
p4raw-id: //depot/perl@33408
diff --git a/util.c b/util.c
index 36166fb..82f147e 100644
--- a/util.c
+++ b/util.c
@@ -3011,6 +3011,7 @@ Perl_wait4pid(pTHX_ Pid_t pid, int *statusp, int flags)
#endif
if (result < 0 && errno == EINTR) {
PERL_ASYNC_CHECK();
+ errno = EINTR; /* reset in case a signal handler changed $! */
}
return result;
}
However, there is still a long hang at the end of the test script at that
revision, whereas blead exits quickly.
I hope I can find that one.
Nicholas Clark
Thread Previous
|
Thread Next