Revision history for MooseX-AttributeShortcuts 0.024 2014-05-02 12:50:57 PDT-0700 * Make the undocumented isa_* options "deprecated" ...which seems like better karma than no warning, as people really shouldn't be using these anyways, but better nice than to have it done to me someday :) 0.023 2014-04-04 21:23:33 PDT-0700 * Add isa_instance_of attribute option. 0.022 2013-09-29 00:08:39 PDT-0700 * No code changes -- released with refreshed test suite 0.021 2013-09-08 23:32:09 PDT-0700 * Drop MXCPA entirely, and incorporate the subtyping functionality 0.020_01 2013-08-26 13:33:44 PDT-0700 * Keep MXCPA quiet, for now. This is evil (hence the dev release), but should keep things status-quo until we properly integrate MXCPA functionality into MXAS proper. 0.020 2013-08-19 21:51:02 PDT-0700 * Bump CoercePerAttribute version requirement to 1.000 and update docs 0.019 2013-04-20 21:50:05 PDT-0700 * Better support inline subtyping and coercions (people who are inclined to shoot themselves in their foot will do it regardless) * Additional tests for coercions 0.018 2013-01-09 10:20:25 PST8PDT * ~~ TRIAL ~~ * If a constraint coderef is given and coercion is requested, we now copy the parent type's coercions to the anonymous child type. * Better document the constraint option and how it interacts with simple coercion. 0.017 2012-10-28 20:53:52 PST8PDT * make actually make classes immutable during tests * add initial primitive anonymous inline subtyping support; this needs to be evaluated with the outstanding pull request (issue #6) but I need this Right Now 0.016 2012-09-08 21:41:38 America/Los_Angeles * provide for "negative shortcuts" for clearer and predicate, but don't document for the moment 0.015 2012-08-26 11:48:29 America/Los_Angeles * add 'builder => sub { ... }' shortcut [gh #4] 0.014 2012-08-16 17:26:09 America/Los_Angeles * explictly test for definedness rather than truthiness when determining if someone has also specified a default along with is => lazy 0.013 2012-07-15 22:11:37 America/Los_Angeles * Restore tests lost from (d11e67a) -- no functional / code changes! 0.012 2012-05-02 11:02:10 America/Los_Angeles * TRIAL/dev release * Don't try to guess what sort of metaclass we need; just die if we don't have one * Misc cleanups 0.011 2012-04-30 13:41:05 America/Los_Angeles * TRIAL/dev release * Handle the no-metaclass case in our init_meta() 0.010 2012-04-06 18:25:42 America/Los_Angeles * Minor doc updates; no functional changes from 0.009 (TRIAL) 0.009 2012-03-26 23:46:58 America/Los_Angeles * Drop 'init_arg => undef' from our 'ro => "lazy"' shortcut. This was done in coordination with the author of Moo, based on consistently surprised user feedback. 0.008 2012-01-10 23:05:34 America/Los_Angeles * Prebuild our roles, for better caching (and fewer warnings, at least) 0.007 2012-01-10 17:30:57 America/Los_Angeles * Allow default to be also be specified when is => lazy 0.006 2011-10-27 10:15:07 America/Los_Angeles * We now handle "trigger => 1", as well * All our tests are wrapped in Test::Moose::with_immutable(), just to make sure that we immutablize properly 0.005 2011-08-18 20:54:02 America/Los_Angeles * handle attribute extension/cloning, for realz this time * we now handle the case of attributes defined in roles by properly applying our trait to the applied_attribute metaclass 0.004 2011-08-02 23:44:20 America/Los_Angeles * add lazy_build => 'private' * our laziness was being ignored when an attribute defined in one class was being extended/cloned in another class -- this could not stand 0.003 2011-07-03 16:40:19 America/Los_Angeles * handle predicate => 1, clearer => 1 as well 0.002 2011-03-30 23:43:11 America/Los_Angeles * Convert our trait to a parameterized role, to allow for different build and writer prefixes (e.g. '_set_' vs '_') * Cause 'is => "lazy"' to behave the way it does in Moo * Change the default writer prefix from _ to _set_; it was pointed out that this is the expected way to do things (and makes more sense, really). 0.001 2011-03-27 08:12:11 America/Los_Angeles * initial release