[Commits] svn commit: 9995

[www-data]

------------------------------------------------------------------------
r9995 | antranig at caret.cam.ac.uk | 2010-08-26 04:04:49 -0400 (Thu, 26 Aug 2010) | 10 lines
Changed paths:
   M /fluid/infusion/trunk/src/webapp/components/undo/js/Undo.js
   M /fluid/infusion/trunk/src/webapp/framework/core/js/DataBinding.js
   M /fluid/infusion/trunk/src/webapp/framework/core/js/Fluid.js
   M /fluid/infusion/trunk/src/webapp/framework/core/js/FluidIoC.js
   M /fluid/infusion/trunk/src/webapp/framework/renderer/js/RendererUtilities.js
   M /fluid/infusion/trunk/src/webapp/framework/renderer/js/fluidParser.js
   M /fluid/infusion/trunk/src/webapp/framework/renderer/js/fluidRenderer.js
   M /fluid/infusion/trunk/src/webapp/tests/framework-tests/renderer/js/RendererUtilitiesTests.js

FLUID-3681, FLUID-3659, FLUID-3683, FLUID-3688:
Widespread changes and improvements to IoC system driven by requirements for CSPACE-2637
Integrated "initRendererComponent" as a new top-level initialiser - "grade" work and "2-pass rendering" is still to come but basic renderer workflow is now available as an integrated package
Alignment of fluid.find with fluid.each, fluid.keyForValue and fluid.transform (TESTS REQUIRED!)
Fix for breakage of Undo component by change in contract of fluid.container (null container now means no component for initView) - this should really use fluid.initLittleComponent which did not exist at time of writing. This entire family needs to be updated for IoC.
Fix for "decorator with valuebinding" FLUID-3659 issue which is now properly understood - the issue actually affected a component which had "value" specified as an expandable EL expression rather than a plain path in "valuebinding". This mode should be supported for consistency and is frequently seen in CollectionSpace.
Further upgrades for fluid.fetchResources - now supports caching of resources with "forceCache" flag. This currently seems somewhat rickety since a "not modified" response from the seems to blow everything up. Ironically a stop-gap solution is to tell the server *NOT* to cache the request for the (subsequently) cacheable resource. More investigation required.
Numerous fixes to IoC - now has support for "static resolution in dynamic context" by appending the static environment onto the thatStack during resolution - required for bundling "test configurations" in CSpace without intruding on the basic component code. Corrected resolution of non-string arguments to resolve recursively as expected. Fix to "noexpand" expander which with "noCopy" semantic was simply causing noexpand material to disappear.
"Unusual" fix to renderer that allows it to locate repetitive leaves in parent contexts rather than infuriatingly giving the "downmap is undefined" error. This needs to look forward to proper work on scoping of names in the renderer.
Puzzled comment on "globalAccept" in DataBinding - why on earth is this here? Must go before we start using this stuff in Kettle.
------------------------------------------------------------------------