Dave Rolsky [Thu, 28 Oct 2010 00:19:26 +0000]
Tidy lists of required modules
Dave Rolsky [Thu, 28 Oct 2010 00:18:41 +0000]
Bump Package::DeprecationManager dep to avoid Test::Exception in prereq chain
Ricardo Signes [Mon, 25 Oct 2010 11:52:48 +0000]
note the switch to Test::Fatal in changelog
Karen Etheridge [Sun, 24 Oct 2010 18:03:43 +0000]
convert all uses of Test::Exception to Test::Fatal.
Dave Rolsky [Mon, 18 Oct 2010 14:37:21 +0000]
Updates release # and date in Changes
Jesse Luehrs [Sat, 16 Oct 2010 00:29:23 +0000]
implement inlined access to the mop slot, to fix immutable anon classes
Jesse Luehrs [Sat, 16 Oct 2010 00:28:35 +0000]
move the __MOP__ stuff back into the instance metaclass
Dave Rolsky [Sat, 16 Oct 2010 01:41:37 +0000]
bump version to 1.10 so we can dep on it in Moose
Jesse Luehrs [Fri, 15 Oct 2010 22:08:37 +0000]
add inline_get_is_lvalue
Jesse Luehrs [Sat, 16 Oct 2010 00:30:28 +0000]
changelog
Jesse Luehrs [Fri, 1 Oct 2010 04:59:32 +0000]
the __MOP__ slot should be used for any weak class, not just anon
Jesse Luehrs [Fri, 1 Oct 2010 04:56:25 +0000]
docs for weaken
Jesse Luehrs [Fri, 1 Oct 2010 04:46:12 +0000]
weaken properly on reinitialize
Jesse Luehrs [Fri, 1 Oct 2010 02:37:07 +0000]
add Class::MOP::metaclass_is_weak
Jesse Luehrs [Fri, 1 Oct 2010 00:26:28 +0000]
don't let anon classes disappear while subclasses still exist
Jesse Luehrs [Fri, 1 Oct 2010 00:18:58 +0000]
make weakening of anon classes explicit
Jesse Luehrs [Thu, 30 Sep 2010 23:49:18 +0000]
add a bunch of anon class tests
Jesse Luehrs [Thu, 30 Sep 2010 22:36:50 +0000]
fix up the __MOP__ slot when reblessing to or from anon classes
Dave Rolsky [Tue, 5 Oct 2010 19:05:32 +0000]
# and date for 1.09
Karen Etheridge [Wed, 29 Sep 2010 19:24:54 +0000]
Revert "document the return value of add_attribute"
This reverts commit
392c350b99ac9271d8a2ef56b913daa6e0a77f0e.
Karen Etheridge [Wed, 29 Sep 2010 00:04:39 +0000]
document the return value of add_attribute
Karen Etheridge [Wed, 29 Sep 2010 00:03:42 +0000]
fix pod markup error
Jesse Luehrs [Mon, 27 Sep 2010 20:45:08 +0000]
let the method subclasses take care of their own hackery
Dave Rolsky [Mon, 27 Sep 2010 17:53:02 +0000]
require Class::MOP in MiniTrait - fixes some failure cases
Jesse Luehrs [Mon, 27 Sep 2010 15:07:54 +0000]
a few more spelling exceptions
Jesse Luehrs [Mon, 27 Sep 2010 15:06:10 +0000]
version bump ::Meta
Jesse Luehrs [Mon, 27 Sep 2010 15:05:44 +0000]
fix warning on earlier perls
Jesse Luehrs [Mon, 27 Sep 2010 08:05:03 +0000]
actually, why not make the meta method renameable
Jesse Luehrs [Mon, 27 Sep 2010 07:46:54 +0000]
ugh, need the hack for this too
Jesse Luehrs [Mon, 27 Sep 2010 01:14:42 +0000]
bootstrap the meta method in CMOP::Mixin too
Jesse Luehrs [Mon, 27 Sep 2010 01:05:58 +0000]
pass no_meta in the tests where we hardcode method metaclasses
Jesse Luehrs [Thu, 23 Sep 2010 08:34:47 +0000]
changelog
Jesse Luehrs [Thu, 23 Sep 2010 08:14:43 +0000]
move this back to HasMethods, since moose roles will need it too
Jesse Luehrs [Thu, 23 Sep 2010 07:59:09 +0000]
tests for meta method object
Jesse Luehrs [Thu, 23 Sep 2010 07:48:07 +0000]
make this a bit more extensible, for moose's benefit
Jesse Luehrs [Thu, 23 Sep 2010 07:31:52 +0000]
give CMOP::Object a real meta method, and simplify some things
Jesse Luehrs [Thu, 23 Sep 2010 05:43:14 +0000]
factor out meta method generation
Jesse Luehrs [Thu, 23 Sep 2010 05:34:43 +0000]
don't pass %options through again here, metaclass compat will handle it
Jesse Luehrs [Mon, 27 Sep 2010 07:37:55 +0000]
ugh, hacks
Jesse Luehrs [Mon, 27 Sep 2010 06:59:52 +0000]
push this stuff back into HasMethods and HasAttributes
Jesse Luehrs [Mon, 27 Sep 2010 06:41:09 +0000]
simplify
Jesse Luehrs [Mon, 27 Sep 2010 04:09:57 +0000]
push a bunch of details about metaclass compat into CMOP::Object
Dave Rolsky [Mon, 27 Sep 2010 01:01:40 +0000]
This CMOP conflicts with Moose
Dave Rolsky [Mon, 27 Sep 2010 01:00:32 +0000]
bump version to 1.09
Dave Rolsky [Mon, 27 Sep 2010 00:58:27 +0000]
Add some whitespace and NEXT
Jesse Luehrs [Mon, 27 Sep 2010 00:12:56 +0000]
reorder these methods into a more "proper" order
Jesse Luehrs [Mon, 27 Sep 2010 00:07:03 +0000]
handle reblessing metaclasses with attributes properly
Jesse Luehrs [Sun, 26 Sep 2010 19:44:07 +0000]
changelog
Jesse Luehrs [Sun, 26 Sep 2010 11:33:01 +0000]
make sure accessors are properly removed on reinitialize
Jesse Luehrs [Sun, 26 Sep 2010 09:56:10 +0000]
better tests
Jesse Luehrs [Sun, 26 Sep 2010 09:04:53 +0000]
basic implementation of preserving attrs/methods across reinitialization
Dave Rolsky [Sun, 26 Sep 2010 15:51:46 +0000]
Add inline_rebless_instance to metaclass
Dave Rolsky [Sun, 26 Sep 2010 05:16:27 +0000]
When inlining constructor, don't access meta instance directly, use metaclass to do inlining
Dave Rolsky [Sun, 26 Sep 2010 05:09:03 +0000]
Implement inlining code in CMOP::Attribute
This can mostly be reused in Moose, and ensures all attributes have a consistent API
Shawn M Moore [Sun, 26 Sep 2010 03:19:13 +0000]
Use AuthorRequires to force authors to run all tests
Dave Rolsky [Thu, 23 Sep 2010 17:14:43 +0000]
Add CMOP::MiniTrait
This just takes the logic that we use when applying the immutability trait and
puts it into a separate package for reuse by other core classes.
Ricardo Signes [Thu, 23 Sep 2010 15:11:06 +0000]
correct link to Perl6-ObjectSpace
Jesse Luehrs [Thu, 23 Sep 2010 04:59:40 +0000]
also allow suppressing the meta method during CMOP::Class->create
Jesse Luehrs [Thu, 23 Sep 2010 04:50:38 +0000]
docs and changelog
Jesse Luehrs [Thu, 23 Sep 2010 01:26:07 +0000]
make lack of a meta method testable
Jesse Luehrs [Thu, 23 Sep 2010 00:50:24 +0000]
don't both "use metaclass" and then "use base 'Class::MOP::Class'"
this breaks several assumptions, including metaclass compatibility (use
base can't recheck that) and the existence of a 'meta' method (if
no_meta => 1 is passed to use metaclass). neither of these are an issue
here specifically, but better to not open up the issue (and it's not
necessary anyway).
Jesse Luehrs [Tue, 6 Jul 2010 05:50:15 +0000]
allow not importing a meta method
Jesse Luehrs [Wed, 22 Sep 2010 18:17:28 +0000]
use get_or_add_package_symbol where appropriate
Jesse Luehrs [Wed, 22 Sep 2010 16:43:25 +0000]
_get_method_map was renamed to _method_map a while ago
Florian Ragwitz [Mon, 20 Sep 2010 18:47:34 +0000]
Depend on a known good version of Package::Stash
Dave Rolsky [Fri, 17 Sep 2010 20:19:54 +0000]
remove debugging warn
Dave Rolsky [Thu, 16 Sep 2010 16:37:26 +0000]
Remove pointless concatenation
Dave Rolsky [Wed, 15 Sep 2010 02:04:56 +0000]
Use Perl::Tidy to tidy dump of generate source code if we can.
Also check source dumping env var each time we generate a sub. This makes it
possible to turn it on selectively, at the cost of a small speed hit (which I
really doubt matters next to the cost of eval'ing a block of code!)
Dave Rolsky [Mon, 13 Sep 2010 15:54:04 +0000]
Bump version to 1.08
Dave Rolsky [Mon, 13 Sep 2010 15:53:57 +0000]
Release date for 1.08
Karen Etheridge [Sat, 11 Sep 2010 17:49:32 +0000]
add conflict for earlier namespace::autoclean, which uses deprecated method_map calls
Dave Rolsky [Mon, 6 Sep 2010 15:00:40 +0000]
Doc changes from gfx
Dave Rolsky [Mon, 6 Sep 2010 14:59:45 +0000]
Clarify comment for new code
Dave Rolsky [Mon, 6 Sep 2010 14:59:24 +0000]
Tidy code
gfx [Mon, 6 Sep 2010 04:56:35 +0000]
Fix a bug that get_method_list() thrown the error for subroutine stubs
Karen Etheridge [Tue, 24 Aug 2010 20:34:47 +0000]
use Test::Requires in tests
Florian Ragwitz [Tue, 24 Aug 2010 19:26:16 +0000]
Make M::I::ExtraTests fail more obvious
Dave Rolsky [Tue, 24 Aug 2010 14:38:07 +0000]
Small changes wording tweaks
Dave Rolsky [Tue, 24 Aug 2010 14:37:46 +0000]
bump version to 1.07
Dave Rolsky [Tue, 24 Aug 2010 14:37:25 +0000]
Changes for 1.07
Dave Rolsky [Tue, 24 Aug 2010 14:36:32 +0000]
Fix a mysterious failure reported by Piers Cawley
Unfortunately, neither of us could think of a simple test for this, but Piers
confirmed that the patch fixes the failure he saw in production. May be
related to the use of MX::Declare and/or TryCatch.
Dave Rolsky [Mon, 23 Aug 2010 13:34:57 +0000]
Restore the call to check_conflicts
Dave Rolsky [Mon, 23 Aug 2010 13:34:04 +0000]
Revert "Use check_conflicts from Module::Install"
This reverts commit
3f3e009e602ec89c1f724e3fabef114ddf68e6d3.
Using this seems to break make disttest and I don't have time to figure this
out.
Dave Rolsky [Mon, 23 Aug 2010 13:20:38 +0000]
Remove extra use of check_conflicts
Dave Rolsky [Mon, 23 Aug 2010 13:18:47 +0000]
bump version to 1.06
Dave Rolsky [Mon, 23 Aug 2010 13:13:29 +0000]
Changes for 1.06
Dave Rolsky [Mon, 23 Aug 2010 13:13:17 +0000]
Add a test for a constant with an undef value
Dave Rolsky [Mon, 23 Aug 2010 13:08:21 +0000]
Constants _should_ be reported as methods.
This makes no fucking sense, but we used to do this, so changing it now may
break existing (boneheaded) code.
Dave Rolsky [Sun, 22 Aug 2010 16:31:57 +0000]
Use check_conflicts from Module::Install
Dave Rolsky [Sun, 22 Aug 2010 15:31:14 +0000]
Release date for 1.05
Dave Rolsky [Sun, 22 Aug 2010 15:31:02 +0000]
Changes for 1.05
Dave Rolsky [Fri, 20 Aug 2010 12:10:46 +0000]
Bump version to 1.05
I want to do this now, since the latest Moose requires the latest CMOP, and I want to install them locally for various bits of testing.
Dave Rolsky [Thu, 19 Aug 2010 14:02:06 +0000]
Only get ISA once in superclasses method
Dave Rolsky [Thu, 19 Aug 2010 04:29:08 +0000]
Don't make a method object for calls to has_method, just for get_method
This speeds things up a bit more, since we can check for the existence of a
method without actually inflating a code ref into a method object.
Dave Rolsky [Sun, 15 Aug 2010 09:17:57 +0000]
Make sure that get_method_list and _get_local_methods handle constants in the namespace safely
Dave Rolsky [Sat, 14 Aug 2010 17:28:33 +0000]
Add _get_local_methods method that returns method objects directly.
Using this in various places can make our code much more efficient, since we
avoid calling get_method twice.
Dave Rolsky [Sat, 14 Aug 2010 16:54:30 +0000]
Big optimization for get_method_list
By checking for a CODE slot in the glob, we can avoid a lot of calls to
->has_method, which is quite expensive.
This seems to provide a speedup of approximately 10% when compiling
Markdent::Simple::Document (and all the classes it loads in turn).
Dave Rolsky [Sat, 14 Aug 2010 16:35:32 +0000]
Simplify code a bit by returning immediately from _check_metaclass_compatibility when possible
Dave Rolsky [Sat, 14 Aug 2010 16:29:30 +0000]
Clarify comment after refactoring
Robert Buels [Tue, 10 Aug 2010 22:50:32 +0000]
docpatch to Class::MOP::Class to make it more clear that get_attribute_list does not traverse the inheritance hierarchy