[Commits] [fluid-project/infusion] 1d8f20: FLUID-4170: Proper event-style coordination for in...

noreply at github.com noreply at github.com
Fri Apr 1 09:48:49 UTC 2011

Branch: refs/heads/master
Home:   https://github.com/fluid-project/infusion

Commit: 1d8f205eb0b963ae56862a33587eb895caceaa9e
Author: Antranig Basman <antranig.basman at colorado.edu>
Date:   2011-04-01 (Fri, 01 Apr 2011)

Changed paths:
  M src/webapp/framework/core/js/Fluid.js
  M src/webapp/framework/core/js/FluidIoC.js
  M src/webapp/framework/core/js/FluidRequests.js
  M src/webapp/tests/framework-tests/core/html/FluidIoC-test.html
  M src/webapp/tests/framework-tests/core/js/FluidIoCTests.js
  M src/webapp/tests/framework-tests/core/js/FluidJSTests.js
  M src/webapp/tests/test-core/jqUnit/js/jqUnit.js
  M src/webapp/tests/test-core/utils/js/TestUtils.js

Log Message:
FLUID-4170: Proper event-style coordination for init lifecycle functions. Widespread upheaval as a longtime fault discovered in the options merging system - a mergePolicy of "function" has never really worked correctly, and was always patched over in the one place it was used so far, fluidRequests.js. Old-style "makeComponents" and the "all-in-one init functions" have been withdrawn and reorganisation of the FluidIoC tests to reflect new semantic - makeComponents() is now just a light wrapper for makeComponent() and will actually issue a fluid.defaults() call on your behalf. This functionality was only ever used in IoC testing. New testing utility "assertLeftHand" allows a convenient assertion of "deepEq by inclusion" - will not complain on discovering EXTRA members than those expected, now important since component options now invariably contain the lifecycle function firers and various other bits and pieces above those explicitly appearing in defaults. New jqUnit assertion,
  "assertNoValue". Reorganisation of file structure of Fluid.js (yet again) to a hopefully more logical order. Slight improvements to the base idiom of event.addListener in that it will now accept a global function name as well as a raw function, being what is natural to supply in the new declarative world. This was not met before since usually listener functions were contextualised and IoC-resolved rather than global static functions typical for lifecycle points.

More information about the Commits mailing list