don't die when looking for 'DOES' on perl < 5.10 (RT#80402)

[gitmo/Role-Tiny.git] / Changes

  - t/does-Moo.t moved to 'xt' (RT#80290)
  - don't die when looking for 'DOES' on perl < 5.10 (RT#80402)

1.002000 - 2012-10-19
  - load class in addition to roles when using create_class_from_roles
  - fix module name in Makefile.PL (RT#78591)
  - when classes consume roles, override their DOES method (RT#79747)
  - method modifiers can be used for 'does' and 'DOES'

1.001005 - 2012-07-18
  - localize UNIVERSAL::can change to avoid confusing TB2
  - properly report roles consumed by superclasses

1.001004 - 2012-07-12
  - remove strictures.pm from the test supplied by mmcleric so we install again
  - when applying runtime roles include roles from original class in new class
    ( fixes ::does_role checks)

1.001003 - 2012-06-19
  - correctly apply modifiers with role composition
  - check for conflicts during role-to-object application (test from mmcleric)
  - add an explicit return to all exported subs so people don't accidentally
    rely on the return value
  - store coderefs as well as their refaddrs to protect against crazy

1.001002 - 2012-05-05
  - alter duplication test to not provoke Class::Method::Modifiers loading

1.001001 - 2012-04-27
  - remove strictures from one last test file

1.001000 - 2012-04-27
  - Documentation improvements, no code changes

1.000_901 - 2012-04-12
  - Fix MANIFEST inclusion of Role::Basic composition

1.000_900 - 2012-04-11
  - Add composition with tests stolen from Role::Basic

1.000001 - 2012-04-03
  - Document that Class::Method::Modifiers must be depended on separately
  - Update tests so that they skip correctly without C::M::M
  - Add a SEE ALSO section

1.000000 - 2012-03-29
  - Remove redundant code in create_class_with_roles
  - Minor doc fix to does_role
  - Split Role::Tiny out into its own dist

Changes below this line are from when Role::Tiny was still bundled with Moo:

  - Fix a bug where coercions weren't called on lazy default/builder returns
  - Switch Moo::Utils to using Module::Runtime, and add the 5.8 %INC
    leakage fix into Role::Tiny's _load_module to provide partial parity
  - Update incompatibilities with Moose documentation
  - Remove Sub::Quote's outstanding queue since it doesn't actually slow
    things down to do it this way and makes debugging easier.
  - Revert 'local $@' around require calls to avoid triggering Unknown Error
  - Explicitly require Role::Tiny in Role::Tiny::With (RT#70446)
  - Fix spurious 'once' warnings under perl -w

0.009013 - 2011-12-23
  - fix up Class::XSAccessor version check to be more robust
  - improved documentation
  - fix failures on perls < 5.8.3
  - fix test failures on cygwin

0.009012 - 2011-11-15
  - make Method::Generate::Constructor handle $obj->new
  - fix bug where constants containing a reference weren't handled correctly
    (ref(\[]) is 'REF' not 'SCALAR', ref(\v1) is 'VSTRING')

0.009011 - 2011-10-03
  - add support for DEMOLISH
  - add support for BUILDARGS

0.009010 - 2011-07-20
  - missing new files for Role::Tiny::With

0.009009 - 2011-07-20
  - remove the big scary warning because we seem to be mostly working now
  - perl based getter dies if @_ > 1 (XSAccessor already did)
  - add Role::Tiny::With for use in classes
  - automatically generate constructors in subclasses when required so that
    subclasses with a BUILD method but no attributes get it honoured
  - add coerce handling

0.009008 - 2011-06-03
  - transfer fix to _load_module to Role::Tiny and make a note it's an inline
  - Bring back 5.8.1 compat

0.009007 - 2011-02-25
  - I botched the copyright. re-disting.

0.009006 - 2011-02-25
  - handle non-lazy default and builder when init_arg is undef
  - add copyright and license info for downstream packagers
  - weak ref checking for Sub::Quote to avoid bugs on refaddr reuse
  - Switch composed role names to be a valid package name

0.9.5 Tue Jan 11 2011
  - Fix clobberage of runtime-installed wrappers by Sub::Defer
  - Fix nonMoo constructor firing through multiple layers of Moo
  - Fix bug where nonMoo is mistakenly detected given a Moo superclass
    with no attributes (and hence no own constructor)

0.9.4 Mon Dec 13 2010
  - Automatic detection on non-Moo superclasses

0.9.3 Sun Dec 5 2010
  - Fix _load_module to deal with pre-existing subpackages

0.9.2 Wed Nov 17 2010
  - Add explanation of Moo's existence
  - Change @ISA setting mechanism to deal with a big in 5.10.0's get_linear_isa
  - Change 5.10 checks to >= to not try and load MRO::Compat on 5.10.0
  - Make 'perl -Moo' DTRT

0.9.1 Tue Nov 16 2010
  - Initial release