Mailing-List: contact flex-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: flex-dev@incubator.apache.org
Received-SPF: pass (athena.apache.org: domain of aharui@adobe.com designates
 64.18.1.183 as permitted sender)
From: Alex Harui <aharui@adobe.com>
To: "flex-dev@incubator.apache.org" <flex-dev@incubator.apache.org>
Date: Thu, 6 Dec 2012 11:45:47 -0800
Subject: Re: [ASJS] Some information on "templates"
Thread-Topic: [ASJS] Some information on "templates"
Thread-Index: Ac3T6YYFg0ih8AUxR+yIHriXsEjbuwAAMNVT
Message-ID: <CCE635EB.499E2%aharui@adobe.com>
In-Reply-To: <20121206143942.99854rj0sybwriwu@franklin.liquidweb.com>
Accept-Language: en-US
Content-Language: en-US
user-agent: Microsoft-Entourage/13.13.0.120411
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0


On 12/6/12 11:39 AM, "Michael Schmalle" <apache@teotigraphix.com> wrote:

> Erik,
>=20
> I'm trying to get clear on something.
>=20
> What exact classes or interfaces get translated from .as to .js from
> the framework directories?
IMO, that's in flux.  I haven't looked at Erik's refactoring, but in my
version each component on the AS side was broken into MVC "beads".  I didn'=
t
do that on the JS side, but right now I think we should, then the M & C
beads can just be cross-compiled.  The V beads are where all flash APIs wil=
l
be used, so equivalent JS V beads will be created by hand.


>=20
> What files are going to be maintained by hand and tests? I'm still a
> bit confused to the flow.
>=20
> You have FlexObject.js, FlexGlobals.js which obviously has no
> correlation in .as. What about that?
>=20
>=20
> Mike
>=20
>=20
>=20
>=20
> Quoting Erik de Bruin <erik@ixsoftware.nl>:
>=20
>> That looks very promising! Some tweaking is needed, and some of the
>> output (the bottom part mostly, I guess that is for some kind of
>> introspection that is in my template provided by other methods) isn't
>> needed, at least not from what I understand.
>>=20
>> I'll look into the details of the new output tomorrow and indicate if
>> and what changes would be nice ;-)
>>=20
>> EdB
>>=20
>>=20
>> On Thu, Dec 6, 2012 at 7:01 PM, Michael Schmalle
>> <apache@teotigraphix.com> wrote:
>>> Hi,
>>>=20
>>> Right now after about 5 hours of researching and messing around I can g=
et
>>> the following using the closure compiler flag;
>>>=20
>>> //=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>>> AS CODE
>>>=20
>>>=20
>>> package com.example.components
>>> {
>>>=20
>>> import flash.display.Sprite;
>>>=20
>>> public class MyTextButton extends Sprite
>>> {
>>>     public function MyTextButton()
>>>=20
>>>     {
>>>         super();
>>>      }
>>>=20
>>>     private var _privateVar:String =3D "do ";
>>>=20
>>>     public var publicProperty:Number =3D 100;
>>>=20
>>>     public function myFunction(value: String): String
>>>     {
>>>         return "Don't " + _privateVar + value;
>>>     }
>>> }
>>> }
>>>=20
>>>=20
>>> //=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>>> JS CODE
>>>=20
>>> /** @preserve CROSS-COMPILED BY MXMLJSC (329449.1) ON 2012-12-06 12:55:=
52
>>>  */
>>> /**
>>>  * CROSS-COMPILED BY MXMLJSC (329449.1) ON 2012-12-06 12:52:19
>>>  *
>>>  * Class: com.example.components.MyTextButton
>>>  * @constructor
>>>  * @extends flash.display.Sprite
>>>  */
>>>=20
>>> // Constructor
>>>=20
>>>=20
>>> /**
>>>  * Constructor: com.example.components.MyTextButton()
>>>  * @constructor
>>>  * @this {com.example.components}
>>>  */
>>> com.example.components.MyTextButton =3D function()
>>> {
>>>         var self =3D this;
>>>                 self.publicProperty /* : Number */ =3D 100;
>>>                 goog.base(this);
>>>                 return self;
>>> }
>>>=20
>>>=20
>>> goog.inherits(com.example.components.MyTextButton, flash.display.Sprite=
);
>>>=20
>>> /**
>>>  * Member: com.example.components.MyTextButton.prototype._CLASS
>>>  * @const
>>>  * @type {com.example.components.MyTextButton}
>>>  */
>>> com.example.components.MyTextButton.prototype._CLASS =3D
>>> com.example.components.MyTextButton;
>>> ;
>>>=20
>>> /**
>>>  * Member: com.example.components.MyTextButton._privateVar
>>>=20
>>>  * @private
>>>  * @type {string}
>>>  */
>>> com.example.components.MyTextButton.prototype._privateVar /* : String *=
/ =3D
>>> "do ";
>>> ;
>>>=20
>>> /**
>>>  * Member: com.example.components.MyTextButton.publicProperty
>>>  * @type {number}
>>>  */
>>> com.example.components.MyTextButton.prototype.publicProperty /* : Numbe=
r */
>>> =3D 100;
>>> ;
>>>=20
>>>=20
>>> /**
>>>  * Method: com.example.components.MyTextButton.myFunction()
>>>  * @this {com.example.components.MyTextButton}
>>>  * @param {string} value
>>>  * @return {string}
>>>  */
>>> com.example.components.MyTextButton.prototype.myFunction =3D function(v=
alue /*
>>> : String */) /* : String */
>>> {
>>>                 /** @type {com.example.components.MyTextButton} */
>>>                 var self =3D this;
>>>                 return (("Don't " + self._privateVar) + value);
>>> }
>>>=20
>>> /**
>>>  * Member: com.example.components.MyTextButton._PACKAGE
>>>  * @const
>>>  * @type {com.example.components}
>>>  */
>>> com.example.components.MyTextButton._PACKAGE =3D com.example.components=
;
>>>=20
>>>=20
>>> /**
>>>  * Member: com.example.components.MyTextButton._NAME
>>>  * @const
>>>  * @type {string}
>>>  */
>>> com.example.components.MyTextButton._NAME =3D "MyTextButton";
>>>=20
>>> /**
>>>  * Member: com.example.components.MyTextButton._FULLNAME
>>>  * @const
>>>  * @type {string}
>>>  */
>>> com.example.components.MyTextButton._FULLNAME =3D
>>> "com.example.components.MyTextButton";
>>>=20
>>> /**
>>>  * Member: com.example.components.MyTextButton._SUPER
>>>  * @const
>>>  * @type {flash.display.Sprite}
>>>  */
>>> com.example.components.MyTextButton._SUPER =3D flash.display.Sprite;
>>>=20
>>> /**
>>>  * Member: com.example.components.MyTextButton._NAMESPACES
>>>  * @const
>>>  * @type {Object}
>>>  */
>>> com.example.components.MyTextButton._NAMESPACES =3D {
>>>         "_privateVar::7:com.example.components.MyTextButton" : true,
>>>         "publicProperty::2" : true,
>>>         "myFunction::2" : true
>>> }
>>>=20
>>> adobe.classes["com.example.components.MyTextButton"]  =3D
>>> com.example.components.MyTextButton;
>>>=20
>>> /CODE
>>> //=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>>>=20
>>>=20
>>> Is this a base we can work off of Erik?
>>>=20
>>> Let me know what you would want changed within that scope.
>>>=20
>>> I know how I can get the imports (IE require()) in. There is some funky
>>> business going on with the constructor that will need to be sorted out.
>>>=20
>>> JSDoc comments are not hard to add/remove either with tags etc..
>>>=20
>>>=20
>>> Mike
>>>=20
>>>=20
>>>=20
>>>=20
>>>=20
>>>=20
>>> Quoting Erik de Bruin <erik@ixsoftware.nl>:
>>>=20
>>>> Correct.
>>>>=20
>>>> I have created 'tests' directories on both the AS and JS frameworks,
>>>> and plan to use FlexUnit for AS and Jasmine for JS unit testing. That
>>>> covers the frameworks. I imagine we build pre- and postcompile tests
>>>> for the project code as well.
>>>>=20
>>>> And leaving out the intermediate step (i.e. have FalconJS directly
>>>> output minified JS, or have FalconJS call the Closure Compiler
>>>> directly as a post compile step) will only increase the difficulty in
>>>> testing for and more importantly, debugging any issues in the output
>>>> JS.
>>>>=20
>>>> Finally, not having intermediate (annotated) JS files will make it
>>>> impossible to let the Closure Builder do it's magic. That magic is an
>>>> important part of having a highly optimised final output. The Builder
>>>> calculates the dependencies between the project code, the FlexJS JS
>>>> framework and the Closure Library, allowing it to remove unused code
>>>> (no penalty for using a large library of which you only use one
>>>> method) and to minify all internal references, while the annotation
>>>> prevents the renaming or even deletion of public interfaces.
>>>>=20
>>>> All in all, the "intermediate" step, at least for the first few
>>>> versions looks to me like a valuable tool and a way to allow for
>>>> speedy development of both projects as well as the frameworks.
>>>>=20
>>>> EdB
>>>>=20
>>>>=20
>>>>=20
>>>> On Thu, Dec 6, 2012 at 4:56 PM, Michael Schmalle
>>>> <apache@teotigraphix.com> wrote:
>>>>>=20
>>>>>=20
>>>>> Quoting Carol Frampton <cframpto@adobe.com>:
>>>>>=20
>>>>>>=20
>>>>>>=20
>>>>>> On 12/6/12 7 :49AM, "Erik de Bruin" <erik@ixsoftware.nl> wrote:
>>>>>>=20
>>>>>>> Mike,
>>>>>>>=20
>>>>>>> After I'm done fixing the mess I made in SVN, I'll start work on th=
e
>>>>>>> 'template', but here is the basic idea to get you started: I would
>>>>>>> like the compiler to output "intermediate" JS, by which I mean "hum=
an
>>>>>>> readable". My Publisher then takes this intermediate code and puts =
it
>>>>>>> through the Google Closure Builder, which optimises and minifies it=
.
>>>>>>> The advantage of having the "intermediate" step is that it makes
>>>>>>> debugging much (MUCH) easier. It will allow us to write tests,
>>>>>>> something that minified code won't. And it will let us run the code=
 in
>>>>>>> the various browser based tooling and have the output make sense.
>>>>>>=20
>>>>>>=20
>>>>>>=20
>>>>>> If you use an intermediate form to test how to you know there aren't
>>>>>> bugs
>>>>>> introduced by the publishing process?
>>>>>>=20
>>>>>> Carol
>>>>>=20
>>>>>=20
>>>>>=20
>>>>> Isn't the publishing process the closure compiler which google has te=
sts
>>>>> for
>>>>> when it compiles out the minified optimized js?
>>>>>=20
>>>>> I probably have more time into investigating the corss compile code I=
n
>>>>> FalconJS and I have to say as it stands we don't even know if our
>>>>> compiler
>>>>> is creating bugs in the output js. It seems Adobe was unit testing th=
e js
>>>>> but we don't have any of those tests.
>>>>>=20
>>>>> So its reasonable to think we need to have tests for js that is compi=
led
>>>>> by
>>>>> FalconJS in the intermediate stage.
>>>>>=20
>>>>>=20
>>>>>=20
>>>>>=20
>>>>> Mike
>>>>>=20
>>>>> --
>>>>> Michael Schmalle - Teoti Graphix, LLC
>>>>> http://www.teotigraphix.com
>>>>> http://blog.teotigraphix.com
>>>>>=20
>>>>=20
>>>>=20
>>>>=20
>>>> --
>>>> Ix Multimedia Software
>>>>=20
>>>> Jan Luykenstraat 27
>>>> 3521 VB Utrecht
>>>>=20
>>>> T. 06-51952295
>>>> I. www.ixsoftware.nl
>>>>=20
>>>=20
>>> --
>>> Michael Schmalle - Teoti Graphix, LLC
>>> http://www.teotigraphix.com
>>> http://blog.teotigraphix.com
>>>=20
>>=20
>>=20
>>=20
>> --
>> Ix Multimedia Software
>>=20
>> Jan Luykenstraat 27
>> 3521 VB Utrecht
>>=20
>> T. 06-51952295
>> I. www.ixsoftware.nl
>>=20

--=20
Alex Harui
Flex SDK Team
Adobe Systems, Inc.
http://blogs.adobe.com/aharui