flex-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alex Harui <aha...@adobe.com>
Subject Re: [FlexJS] FDT responds to latest developments
Date Thu, 06 Mar 2014 16:28:33 GMT


On 3/6/14 6:32 AM, "Erik de Bruin" <erik@ixsoftware.nl> wrote:

>Hi,
>
>I just got a response to my earlier bump to the FDT devs (in <comment>
>below). Maybe someone can address these points, as I'm not familiar with
>the FlexJS SDK construction by the Installer, nor the inner workings of
>the
>AS side of the FlexJS SDK...
>
><comment>
>Hi Erik de Bruin,
>
>I was able to create the FlexJS SDK.
>Also I was able to create a test project DataBindingTest.
>and it compiled into some java script code.
>
>Two points about the FlexJS:
>FDT is using internally the compiler.jar of any Flex SDK.
>The compiler.jar of the Flex JS SDK asks for some locals
>(frameworks/locale/en_US)
>I copied them by hand into he Flex JS SDK so that the compilation runs
>through.
>I think it would be the best if your installer also copy these files and
>so
>creates a normal functional Flex SDK so that the debug session for Flex JS
>does
>not differ from normal Action Script Session with SWFs. This will make the
>adaption far more easier.
I think this may be a project assumption by FDT.  Could they be adding
that library into the library path when calling compiler.jar?  FlexJS does
not use Flex SDK ResourceManager.  I am not seeing this in FB.

>
>Another point is this exception of the generated SWF during a debug
> session:
>
>  Exception fault: ReferenceError: Error #1069: Property updateParent not
>found on mx.binding.PropertyWatcher and there is no default value.
>        at _MyInitialViewWatcherSetupUtil/setup()
>        at MyInitialView()[C:\Work\runtime-New_configuration-
>Maven\FlexJSTest\src\MyInitialView.mxml:23]
>        at DataBindingTest/_DataBindingTest_MyInitialView1_c()
>        at DataBindingTest()[C:\Work\runtime-New_configuration-
>Maven\FlexJSTest\src\DataBindingTest.mxml:8]
>
>Maybe you can give a hint to solve this issue.
I think IntelliJ will also face this.  If you look at the FB launch
configurations, they pass a bunch of new MXML options to the compiler like
mxml.children-as-data and others that swap out PropertyChangeEvent for
ValueChangeEvent.  In FB, I have not found a way to make these options be
a default (by putting in -config.xml for example) such that when
MXMLC/compiler.jar does a compilation during editing, MXMLC doesn't choke
on those unknown options.  That's why FB currently has these launch
configs.  Someday, we may be brave enough to not package MXMLC in a FlexJS
SDK and totally go with Falcon and then these issues may go away.  But for
now, the IDE vendors also need a way to launch the Falcon compiler with
different options before running a SWF.

-Alex


Mime
View raw message