Re: Segfault with MooseX::Declare and Regexp::Grammars

On win32 I get the same, but this also disappears if I use Data::Dump or Data::Dump::Streamer instead of Data::Dumper

Running under the debugger, I get

Signal SEGV at (reeval 422)[pm.890409.pl:30] line 1
[download]

on the 10th iteration.

With Devel::Trace it happens after

>> (eval 334)[C:/perl/site/5.10.1/lib/MSWin32-x86-multi-thread/Class/M
+OP/Method/Generated.pm:53]:1: sub {exists $_[0]->{"actual_body"}}
>> C:/perl/site/5.10.1/lib/MSWin32-x86-multi-thread/Class/MOP/Method/W
+rapped.pm:48:             return $around->{cache}->(@_);
>> accessor actual_body defined at C:/perl/site/5.10.1/lib/MooseX/Meth
+od/Signatures/Meta/Method.pm:101: >> accessor actual_body defined at 
+C:/perl/site/5.10.1/lib/MooseX/Method/Signatures/Meta/Method.pm:101: 
+>> C:/perl/site/5.10.1/lib/MooseX/Method/Signatures/Meta/Method.pm:15
+3:         goto &{ $actual_body };
>> C:/perl/site/5.10.1/lib/MooseX/Method/Signatures/Meta/Method.pm:153
+:         goto &{ $actual_body };
>> pm.890409.pl:34:     method BUILD {

>> pm.890409.pl:35:         use Data::Dumper;        print 'AH ', Dump
+er \@_;

>> C:/perl/5.10.1/lib/MSWin32-x86-multi-thread/Data/Dumper.pm:553:   r
+eturn Data::Dumper->Dump([@_]);
>> C:/perl/5.10.1/lib/MSWin32-x86-multi-thread/Data/Dumper.pm:190:    
+ return &Dumpxs
>> C:/perl/5.10.1/lib/MSWin32-x86-multi-thread/Data/Dumper.pm:64:   my
+($c, $v, $n) = @_;
>> C:/perl/5.10.1/lib/MSWin32-x86-multi-thread/Data/Dumper.pm:66:   cr
+oak "Usage:  PACKAGE->new(ARRAYREF, [ARRAYREF])" 
>> C:/perl/5.10.1/lib/MSWin32-x86-multi-thread/Data/Dumper.pm:68:   $n
+ = [] unless (defined($n) && (ref($n) eq 'ARRAY'));
>> C:/perl/5.10.1/lib/MSWin32-x86-multi-thread/Data/Dumper.pm:70:   my
+($s) = { 
>> C:/perl/5.10.1/lib/MSWin32-x86-multi-thread/Data/Dumper.pm:97:   if
+ ($Indent > 0) {
>> C:/perl/5.10.1/lib/MSWin32-x86-multi-thread/Data/Dumper.pm:98:     
+$s->{xpad} = "  ";
>> C:/perl/5.10.1/lib/MSWin32-x86-multi-thread/Data/Dumper.pm:99:     
+$s->{sep} = "\n";
>> C:/perl/5.10.1/lib/MSWin32-x86-multi-thread/Data/Dumper.pm:101:   r
+eturn bless($s, $c);
>> generated method (unknown origin):30: >> (reeval 421)[pm.890409.pl:
+30]:1: >> (reeval 421)[pm.890409.pl:30]:2: >> (reeval 418)[pm.890409.
+pl:30]:2: >> C:/perl/site/5.10.1/lib/Regexp/Grammars.pm:556:     my (
+$stack_ref, $key, $original_name, $value) = @_;
[download]

So I think it could be a bug in Dumpxs, or the perl regexp engine, or both.

Comment on Re: Segfault with MooseX::Declare and Regexp::Grammars Select or Download Code

Replies are listed 'Best First'.
Re^2: Segfault with MooseX::Declare and Regexp::Grammars by betterworld (Curate) on Feb 27, 2011 at 14:06 UTC
This cannot be a bug in Data::Dumper: If I remove the dumping code (so BUILD is empty), the segmentation fault still happens. Sorry I did not mention this. I was so accustomed to Data::Dumper that I did not suspect a bug there :)	[reply]
Re^3: Segfault with MooseX::Declare and Regexp::Grammars by Anonymous Monk on Feb 27, 2011 at 14:30 UTC
You're not the only one to load Data::Dumper :) `print "AH Data::Dumper::VERSION $Data::Dumper::VERSION @_\n" +;` [download] `AH Data::Dumper::VERSION 2.128 Foo=HASH(0x1bc0a2c) HASH(0x1bc0afc)` [download] Data::Dumper may or may not be responsible, directly or otherwise ... someone needs to bust out a debugger and track down where the segfault occurs	[reply] [d/l] [select]


laziness, impatience, and hubris
	PerlMonks