Front page | perl.perl5.porters |
Postings from May 2003
On Mon, May 26, 2003 at 02:41:48PM -0500, Ken Williams wrote:
> Hi,
>
> Below is a short test case in which I'm calling some callback perl code
> from XS and using the G_EVAL flag to trap exceptions. I then call
> croak(Nullch) to propagate the exception, and I get a segmentation
> fault. Does anyone see something wrong with it? The code is below.
> I'm running it under perl 5.6.1 on Mac OS X 10.2.6.
On 5.6.1 Doesn't look like a SEGV from here.
==20933== Memcheck, a.k.a. Valgrind, a memory error detector for x86-linux.
==20933== Copyright (C) 2002, and GNU GPL'd, by Julian Seward.
==20933== Using valgrind-1.9.6, a program instrumentation system for x86-linux.
==20933== Copyright (C) 2000-2002, and GNU GPL'd, by Julian Seward.
==20933== Estimated CPU clock rate is 1545 MHz
==20933== For more details, rerun with: -v
==20933==
Using /home/nick/Params/Validate/blib
Using /home/nick/Params/Validate/blib
Checking...
Will fail at bug.pl line 8.
Checking eval at bug.pl line 8.
Ref looks okay at bug.pl line 8.
==20933==
==20933== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)
==20933== malloc/free: in use at exit: 405518 bytes in 8761 blocks.
==20933== malloc/free: 17825 allocs, 9064 frees, 8529260 bytes allocated.
==20933== For a detailed leak analysis, rerun with: --leak-check=yes
==20933== For counts of detected errors, rerun with: -v
Looks like this die:
> if (SvROK(ERRSV)) {
> warn("Ref looks okay");
> croak(Nullch);
> }
> }
However, 5.8.0 does something different:
$ /usr/local/perl5.8.0/bin/perl5.8.0 bug.pl
Checking...
Will fail at bug.pl line 8.
Checking eval at bug.pl line 8.
Ref looks okay at bug.pl line 8.
main=HASH(0x814a600)
There's no newline at the end of that last line.
No, I don't know what's going on.
Nicholas Clark
Thread Previous
|
Thread Next