One last tweak:
obj.hasOwnProperty("$lsc$Fcn_dependencies")
should be:
"$lsc$Fcn_dependencies" in obj
I think. Because the dependency function might be defined in a
superclass of obj?
This idiom is effectively Lisp :send-if-handles
On 2008-03-04, at 10:43 EST, Donald Anderson wrote:
Change 20080304-dda-Q by [EMAIL PROTECTED] on 2008-03-04 10:32:20 EST
in /Users/dda/laszlo/src/svn/openlaszlo/trunk
for http://svn.openlaszlo.org/openlaszlo/trunk
Summary: In dependency calculation function, use bracket syntax
rather than cast
New Features:
Bugs Fixed: LPP-5335
Technical Reviewer: ptw (pending)
QA Reviewer: (pending)
Doc Reviewer: (pending)
Documentation:
Release Notes:
Details:
Previously, the new Compiler.dependenciesForExpression would
produce an expression like:
obj.hasOwnProperty("$lsc$Fcn_dependencies") ? (obj cast Object).
$lsc$Fcn_dependencies(this, obj, arg1)
For consistency with other cases that must do 'typeless'
references, this has been changed
to use the bracket notation:
obj.hasOwnProperty("$lsc$Fcn_dependencies") ? obj['$lsc
$Fcn_dependencies'](this, obj, arg1)
Tests:
Checked regressions: (smokecheck, weather, lzpix) x (swf8, dhtml)
The feature itself cannot be easily checked until this function is
used
within the new constraint framework.
Files:
M WEB-INF/lps/server/src/org/openlaszlo/sc/
ReferenceCollector.java
Changeset: http://svn.openlaszlo.org/openlaszlo/patches/20080304-dda-Q.tar
--
Don Anderson
Java/C/C++, Berkeley DB, systems consultant
voice: 617-547-7881
email: [EMAIL PROTECTED]
www: http://www.ddanderson.com