I got the same thing recently.
I was passing a string as an argument to the 'first' delegation method
and not a code ref (which is what first expects).
Not sure if that is your issue too, but thought I would throw in my 2
cents.
(if this is the correct issue perhaps a bug to List::Util is in order,
segfaulting is not acceptable behavior when encountering bad input).
- Stevan
On Nov 11, 2009, at 1:45 PM, Ian Sillitoe wrote:
Haven't solved this but thought I would update this email with some
extra
information.
After some help from #moose (thanks rafl) - I've got a gdb traceback
which
points the finger towards some kind of interaction with
List::Util::first
Program received signal SIGSEGV, Segmentation fault.
0x00002b1a1bb0e254 in XS_List__Util_first (cv=0x9d16f40) at
ListUtil.xs:337
337 PUSH_MULTICALL(cv);
(gdb) bt
#0 0x00002b1a1bb0e254 in XS_List__Util_first (cv=0x9d16f40) at
ListUtil.xs:337
#1 0x0000000000498cc1 in Perl_pp_entersub () at pp_hot.c:2877
#2 0x0000000000477bc2 in Perl_runops_debug () at dump.c:1459
#3 0x000000000041ff74 in S_run_body (oldscope=1) at perl.c:2366
#4 0x000000000041fa86 in perl_run (my_perl=0x9b72010) at perl.c:2283
#5 0x000000000041bc14 in main (argc=4, argv=0x7fff1ee8c728,
env=0x7fff1ee8c750) at perlmain.c:99
Strange thing is that I use List::Util all over the place in my code
but
only get the seg fault when I implement native traits. I also can't
seem to
get a test case to seg fault without getting Moose involved.
I'm using latest version of List::Util and I've tried on a bunch of
different perl builds (and machines) at work with the same backtrace.
Obviously Moose gets a clean bill of health on cpan as does
List::Util so
maybe it is a weirdness with the setup here.
If anyone else has any suggestions for other possible avenues of
exploration
then please do let me know.
Cheers,
Ian