Bug #96310 for Moose: Moose::Exception subclassing bug

Sat Jun 07 12:48:45 2014 xenoterracide [...] gmail.com - Ticket created

Subject:	Moose::Exception subclassing bug
Date:	Sat, 7 Jun 2014 11:48:36 -0500
To:	bugs-moose [...] rt.cpan.org
From:	Caleb Cushing <xenoterracide [...] gmail.com>

Can't call method "as_string" on an undefined value at /home/xenoterracide/perl5/perlbrew/perls/perl-5.18.2/lib/site_perl/5.18.2/x86_64-linux/Moose/Exception.pm I extended Moose::Exception and was overriding trace, this was unintentional on my part, bu if you look at Moose::Exception you'll fine that the overload stringification, calls $self->trace->as_string, it should instead check to see if trace is defined, in the event that a subclasser does this. At the discretion of the fixer how it handles undefined, though I would suggest an exception of some sort. -- Caleb Cushing http://xenoterracide.com Calendar: https://www.google.com/calendar/embed?src=xenoterracide%40gmail.com&ctz=America/Chicago

Wed Jun 11 18:34:25 2014 ether [...] cpan.org - Correspondence added

On 2014-06-07 09:48:45, XENO wrote: Show quoted text

> Can't call method "as_string" on an undefined value at > /home/xenoterracide/perl5/perlbrew/perls/perl- > 5.18.2/lib/site_perl/5.18.2/x86_64-linux/Moose/Exception.pm

Given that the 'trace' method comes from an attribute, declared thusly: has 'trace' => ( is => 'ro', isa => 'Devel::StackTrace', ... ); trace could ever become undefined. Even if something else was altering how 'trace' behaves, the original contract of "trace must always return something defined" ought to be honoured. Show quoted text

> I extended Moose::Exception and was overriding trace, this was > unintentional on my part, bu if you look at Moose::Exception you'll > fine that the overload stringification, calls $self->trace->as_string, > it should instead check to see if trace is defined, in the event that > a subclasser does this. At the discretion of the fixer how it handles > undefined, though I would suggest an exception of some sort.

Well, it did throw an exception, just not an objectified one :) Since the prospect of subclassing Moose::Exception has been shown elsewhere to be dubious, I'm closing this ticket.

Wed Jun 11 18:34:26 2014 The RT System itself - Status changed from 'new' to 'open'

Wed Jun 11 18:34:27 2014 ether [...] cpan.org - Status changed from 'open' to 'rejected'

Wed Jun 11 18:35:31 2014 ether [...] cpan.org - Correspondence added

On 2014-06-11 15:34:25, ETHER wrote: Show quoted text

> trace could ever become undefined.

Sorry, typo: ... trace SHOULD NEVER become undefined.