On Fri, Dec 05, 2008 at 01:23:06PM -0800, Chip Salzenberg wrote: > What say we make named parameters part of Perl 5.12? (Finally?!) > > A quick check of CPAN suggests that Schwern's code in Method::Signature > defines the state of this particular art. No. The state of the art is defined by: rafl's signatures.pm rhesa's Method::Signatures::Simple rafl's MooseX::Method::Signatures which are all being used for production code[0] - the last one is based on gaal's Perl6::Signatures work although we'll probably move it to its own custom parser at some point. This is what's likely to be supported by default in future Catalyst/DBIx::Class stuff and MooseX::Declare. Schwern's stuff is basically a prototyping land and has shown a lack of stability - Method::Signatures::Simple was a reaction to it failing to install on win32 when the Data::Alias stuff was first introduced and a desire to keep things relatively minimal until Moose-land has had a chance to really shake out what we want from a larger feature set. However, I'm not sure this is how we should be looking at the problem. What I'd really prefer to see is a way to allow pluggable signature parsers as a pragma (which I think is roughly where rafl's signatures.pm is going) and provide a simple default along the lines of Method::Signatures::Simple's syntax. At the very least, we should do that -first- and keep an eye on what stabilises out on CPAN before considering bringing anything more complex near the core. [0] admittedly in the case of MX::Method::Signatures only under controlled conditions with a subset of the syntax so far - the documented syntax in that module is considered largely stable now, although we're not encouraging wide use yet. -- Matt S Trout Need help with your Catalyst or DBIx::Class project? Technical Director http://www.shadowcat.co.uk/catalyst/ Shadowcat Systems Ltd. Want a managed development or deployment platform? http://chainsawblues.vox.com/ http://www.shadowcat.co.uk/servers/Thread Previous | Thread Next