Re: [PATCH] Test RegexpRef type using SvRX or checking PERL_magic_qr

2011-01-29 Thread David Leadbeater
Hi, On 31 Dec 2010, at 21:46, David Leadbeater wrote: > With pluggable regexp engines checking "ref" of a qr// isn't correct; > it's also possible for a normal Regex to be blessed into another > class. [...] There's now a cleaned up version of this sitting in

Re: $some_role->DOES

2011-01-06 Thread David Leadbeater
On Thu, Jan 06, 2011 at 06:58:35AM -0500, Stevan Little wrote: > I think currently DOES support is located in Moose::Object, seems like a > good idea to add support for it in the Meta Role too. The behaviour depends on if UNIVERSAL::DOES is implemented by the version of perl too. On 5.12 Ovid's s

[PATCH] Test RegexpRef type using SvRX or checking PERL_magic_qr

2010-12-31 Thread David Leadbeater
With pluggable regexp engines checking "ref" of a qr// isn't correct; it's also possible for a normal Regex to be blessed into another class. Additionally an 'Object' subtype previously excluded objects that were ->isa('Regexp'), this was inconsistent with the RegexpRef handling (i.e. neither Rege

Re: [PATCH] Test for RegexpRef type constraint via ->isa

2010-12-31 Thread David Leadbeater
On 31 Dec 2010, at 18:41, Florian Ragwitz wrote: > The specialness of regexp refs mostly comes from either having an SV > type of SVt_REGEXP on new perls, or from PERL_MAGIC_qr on older perls. That's what I was getting at with my reftype comment, I've just realised re::is_regexp exists which is p

[PATCH] Test for RegexpRef type constraint via ->isa

2010-12-31 Thread David Leadbeater
With pluggable regexp engines a Regexp is actually a subclass of Regexp, not Regexp itself. (Another way to do this would be to use Scalar::Util::reftype and check for REGEXP; however this only works since regexps became a first class type in perl 5.12.) I've not been able to run the tests, lots