Fixed (or at least returned to 3.4.1 state), the sublest regression ever, related to Y.cloning Attribute driven objects with Node values (e.g. Anim), on IE. Fundamentally, Y.cloning nodes is broken in IE and has been since 3.3.0. If an object holds multiple references to a Node instance, the first reference encountered while cloning, will be a damaged clone. Subsequent references encountered will not clone (and hence be OK). After the Attribute refactor, the value stored in State was the first reference encountered, resulting in permanent damage. Before the refactor, a transient event based reference was the first reference encountered, and hence was less likely to cause real-world problems. * Added unit test cases to cover cloning Attribute based objects, with node values * Tested on IE6, 7, 8, 9. * Ran some core unit tests (App, Plugin, Widget, Attribute, Base, Node) through Yeti on IE8, Chrome, Safari, FF. * Spot tested a couple of examples (Widget/App/Attribute). Wanted to get this in before testfest tomorrow. Fixes #2531929 Longer term fix for Y.clone(node) on IE is still required.

More line level coverage for Attribute

Added path to test valueFn:string

Fixed bugs in attribute._getAttrCfg and state.removeAll, found while bumping up unit test coverage for attribute-extras and attribute-core. Fixes #2531891. Fixes #2531892 Added standalone state unit tests also.

Fixed regression - Y.augment(node, Y.Attribute) not firing valueChange events. Fixes #2531851

Fixed failing Attribute tests

Fixed failing Attribute tests

More adhoc attr tests - enabled/disabled

Initial drop for ad-hoc constructor attribute support. See #2528968

* Added BaseCore Tests * Fixed AttributeCore to handle Base and BaseCore references (thanks to tests) * Fixed whitespace in Attribute docs.

Prepped tests for coverage reporting

Added param support to constructor. Fixes #2530093. Added unit tests too (just like Ryan does)

Pulled in modified tests (names, no console) from 3.1.0+ branch

Added writeOnce:'initOnly' support, to allow users to define attributes which can only be set during construction

Added intl unit tests, fixed root lang handling to return instead of null for root lang. Fixed yui- to yui3- in attribute tests

Updated to account for the "newVal" property of attribute change events passing through defined getter.

Added support to allow ATTRS function refs (setter, getter, validator, valueFn) to be specified by instance method name. Avoids anonymous fn hop and promotes best practice of encapsulating logic in instance methods

Attribute should test against raw value stored in state, as opposed to massaged value from getter, when determining if state actually changed in order to notify after listeners. Fixes #2528248. It's a fairly rare case (rawNewVal === massagedPrevVal would not invoke after listeners) for which this would actually break in it's currently buggy state, but it's the right thing for Attribute to do, so wanted to make the fix for GA. * Attribute unit tests all pass, including new ones added for this use case. * No example regressions found, apart from one - in Overlay/Widget-StdMod set content before/after examples. * Widget-StdMod was broken because of the funky way it was handling set content changes, when inserting content before or after the existing content. Cleaning this up fixed the issue.

Added 2nd level merge for Node augmented with Attribute use case, to protect not only the attrs hash, but also each config. Code change is only used for the Node augmented with Attribute use case. Base use case is unchanged.

Attribute currently will not invoke after listeners if the value of an attribute has not changed (using a newVal === prevVal test) during a set. This commit limits this behavior to primitives, so that attributes with Object values can be set to the same Object reference, but with modified Object state (e.g. property change, array elements added etc.) and still have after listeners notified to react to the changed Object state.

Fixed typo. base-complex -> attribute-complex in loader meta-data. Also added attribute as a widget dependendency adding attribute-complex support OOTB.

Added support for _stateProxy (to be used by Node). Also renamed private internal '_conf' property to '_state' to match. _stateProxy usage still needs testing, but normal Attribute usage tests all pass.

1). Fixed hasImpl method on Base.build classes. Fixes #2527997. 2). Allow ad-hoc attributes (attribute which have not be added/configured) to be set/get. 3). Allow PluginHost to be used on classes without _getClasses() impl (in preparation to make it independant of Base). 4). Minor kweight tweaks.

Minor perf/kweight refactor. Use local var for clazz.ATTRS. Also renamed the _lazyAttrInit flag to _lazyAddAttrs to match addAttrs API method. Currently only drag.js is using it to disable lazy attribute init (and it's not planning to for beta 1)

Fix for silent init support - forgot to add VALUE key to state.add(INITIALIZED, VALUE, val) call

Lazy Attribute Support. For a 22 Attribute Test Object : Lazy construction (no attributes configured) = 2.15ms -> 1.45ms (FF3), 6.25 -> 3.9 (IE6). Lazy construction + all attributes configured lazily (expected to be a little slower) = 2.5ms -> 2.65ms (FF3), 8.6ms -> 7ms (IE6). Lazy attribute support is configurable on an instance level. All Attribute tests pass. No regressions in Node, Anim, DD, Widget, Overlay, Slider examples aside for DD group and DD scroll support (due to the fact that setters were being used to set state in other parts of the object).

Added plughost, plugin, Base.plug, Base.unplug unit tests

Added ability to modify a subset of the attribute configuration properties, after intial configuration. The attribute properties which are modifiable are ones which if changed, don't have the potential to conflict with the stored value. These are getter, readOnly, writeOne and broadcast. Fixes #2527834

Attributes for which user provided initial value fails validation will revert to default configured value. Fixes #2527830

Added broadcast support for attribute events (Y.Focus requirement)

Added basic construction perf test to attribute unit tests

Attribute getter, setter, validator now receive attribute name as 2nd arg (value, name). Fixes #2527872

Base Sugar: Added ability to subscribe to on/after event moments through the constructor. Minor perf tweak - use String.splice instead of String.split when we only need the attribute name from a complex ('x.y.z') set

Attribute: Init on demand feature. Initial commit. Needs optimization/cleanup pass

Re-added dynamic:false support to Base.build. Fixed compressor var warnings

Only notify after listeners if attribute value has really changed (uses === for comparison). Sub Attribute value changes made using the . notation (set(x.y.z, 5)) will always notify after listeners, since we don't want to do a deep compare for perf reasons. Fixes #2527833

Added support for setters to prevent a set by returning Attribute.INVALID_VALUE. Added test case for it also

Remove logLevel setting from Console

Removed EventTarget wrappers in Base. Using EventTarget prefix instead

Attribute beta1 fixes and API changes - cfg.set -> cfg.setter, cf.get -> cfg.getter, xxxAtts -> xxxAttrs, validator before set, over-ride support for static defaults, removed CLONE support. Attribute tests pass. Anim, DD, Widget, Overlay, Slider examples pass.

First Attribute pass on beta1 candidate - still have an issue to fix with static object values

House cleaning..