We've got the `$flexversion` compile-time constant, so with the new
compile-time evaluator, you can write:
if ($as3 && $flexversion == "10.1") {
// code path only for Flash Player 10.1
}
This will still not give you a lfc-version for 10.1, but at least makes
it easier for people who depend on 10.0 to generate their own lfc.
On 2/26/2011 8:46 PM, P T Withington wrote:
At this point, the most I am willing to do is make a release note. We don't
have compile-time support for conditionalizing on point versions, and I don't
think we want to generate separate runtimes for point versions. If someone
truly wants to support 10.0, they will have to build from source and set the
default to 10.0.
[Unless I am missing something? The flex compiler will not even let me compile
the 10.1 code unless I request the 10.1 runtime, so I don't see how I could
write code to dispatch dynamically at runtime based on the runtime point
release.]
On 2011-02-25, at 11:33, André Bargull wrote:
We've defaulted to 10.1, but users we're still able to change the required
minimum version by calling lz.embed.swf() with the second argument.
On 2/25/2011 5:22 PM, P T Withington wrote:
Whoops on the 'for' loop. I 'fixed' it from for each. Bleah.
I just did some research, at Amy's request. Apparently we have required 10.1
since 4.8.1. See LPP-9218, which updated embednew.js. So, I am not going to
worry about creating a 10.0 path.
On 2011-02-25, at 10:37, André Bargull wrote:
The default flex version is now 10.1
The change does not only change the default flex version to 10.1, it now
requires 10.1, so maybe we need/want two different code paths for 10.0 and 10.1
(btw flex version is a bad name, it's actually the flash player version. The flex version
we're using is 4.1.0.1607. And things will get more complicated as soon as we're using
the next flex version "Hero" which adds support for swf11, because starting
with swf11, the swf version and flash player version are no longer corresponding.)
for (var k in interfaces) {
for-loop instead of for-in to iterate over arrays
On 2/24/2011 2:06 AM, P T Withington wrote:
[Henry can you try profiling with this change in and report on the difference?]
Change ptw-20110223-UAD by [email protected] on 2011-02-23 19:53:21 EST
in /Users/ptw/OpenLaszlo/trunk-devo
for http://svn.openlaszlo.org/openlaszlo/trunk
Summary: Speed up `subclassof` on swf10
Bugs Fixed: LPP-9775 Improve `subclassof` for swf10
Technical Reviewer: [email protected] (pending)
QA Reviewer: [email protected] (pending)
Release Notes:
The default flex version is now 10.1
Overview:
Use Andre's suggestion of a custom query to the JSON introspection
routines to speed up the checking of class against interface for
$lzc$subclassof
Details:
LzRuntime: Use the new introspection tool
lps.properties, LPS: The default flex version is now 10.1
platform/flex/...: New runtime support
SWF9External: When compiling the LFC look for custom classes in
lfc/compiler/platform/flex.
Tests:
smokecheck, test/style/metasuite, Henry will profile lzmail
Files:
M WEB-INF/lps/lfc/compiler/LzRuntime.lzs
A WEB-INF/lps/lfc/compiler/platform/flex
A WEB-INF/lps/lfc/compiler/platform/flex/avmplus
A WEB-INF/lps/lfc/compiler/platform/flex/avmplus/JSONReflection.as
A WEB-INF/lps/lfc/compiler/platform/flex/README.txt
M WEB-INF/lps/config/lps.properties
M WEB-INF/lps/server/src/org/openlaszlo/sc/SWF9External.java
M WEB-INF/lps/server/src/org/openlaszlo/server/LPS.java
Changeset: http://svn.openlaszlo.org/openlaszlo/patches/ptw-20110223-UAD.tar
