flex-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Michael Schmalle <apa...@teotigraphix.com>
Subject Re: [FalconJx] Prototype ActionScript -> JavaScript compiler code up in svn
Date Fri, 14 Dec 2012 11:17:15 GMT

Quoting Daniel Wasilewski <devudesign@gmail.com>:

> Hi guys,
> I've lost a track on the Falcon JS for a while, (busy freelancing to  
> survive and pay my bills), but at time to time visiting the list.
> And don't quite understand the status. Is it like we have 3  
> different people working on their own/favourite implementation at  
> the same time?

No, We have the FlaconJS over bloated prototype as we call it. And we  
have what I just committed to svn. Erik is trying to work on the js  
framework level and actionscript component framework.

I know why he is getting upset, he needs to know what design pattern  
he is using because that determines how he has to "write" is js code.

I say stick with goog, if Frank wants to use something else he can. If  
in the future we find a better way, I will PERSONALLY help change  
every freaking line of code Erik. Just focus on goog!

> I've seen Michael done significant progress with his rewrite from scratch.
> But in my mind it wasn't excluding 'goog' approach for JS output  
> since he mentioned is not there yet.

Yes, I actually does spit out a bit of goog. I said I was changing  
that becasue I wanted and abstract class. Now that I think about it,  
Erik may have misinterpreted that as I was "refactoring" goog out,  
which is NOT what I said.

I said, I wanted the walker to be agnostic when it comes to javascript  
implementation. That is why I also said, this thing is going to spit  
out valid ActionScript code before JavaScript.

> I am sure bit and pieces of this puzzle can be put together. And  
> even if we have a prototype of Falcon in current form, Michael  
> decided to investigate different approach.

Yeah, this was out of survival. There is to much waste in this world  
because people half ass stuff, I am not one of those people. I take  
pride in my creation and design. If I'm wrong about the  
implementation, I don't even care, I will just learn new things. :)

> Some say it is not worth it, but I do appreciate his effort because  
> this project is on research state, not production ready. And this  
> very project needs solid foundations to start from. We don't want to  
> build on top of over-bloated solution only because it has been done.  
> I've seen many things done in my life, but quite useless as well.

Agreed, 2/3 of the things created these days are useless, just look at  
the millions of abandoned open source projects and this will prove  
that point.

> No effort in order to try out different solutions and approaches is wasted.
> And as I said, things can be put together especially by people have  
> learn something during research time and have more experience now.

That is why I took a week of MY time to get up in svn a valid  
alternate cross compiler, I want to give people hope in a new  
direction so it can galvanize a new effort.

What will be, will be.


> On 12/14/2012 8:33 AM, Erik de Bruin wrote:
>> So, basically, nobody loves the "goog" approach I spend the last weeks
>> working on (based mostly on feedback from the various discussion on
>> the list).
>> Or, let me rephrase, nobody cares enough to contribute to it?
>> EdB
>> On Fri, Dec 14, 2012 at 9:20 AM, Frank Wienberg <frank@jangaroo.net> wrote:
>>> This is great news, Mike! I will also try to dig into your code this
>>> weekend.
>>> In the meantime, I've been busy figuring out the "essence" of a new
>>> JavaScript runtime format that uses the principles described in my blog,
>>> but relies on RequireJS (not goog!) and ECMAScript 5 API, making it way
>>> more concise than the current Jangaroo Runtime. For IE8 and other non-ES5
>>> browsers, we would then use polyfills for all ES5 functions used.
>>> Let's see if I can get an approval from my company to contribute; if it
>>> takes too long, I'd blog about the concepts and you or someone else would
>>> have to implement them.
>>> Greetings
>>> -Frank-
>>> On Fri, Dec 14, 2012 at 1:31 AM, Michael Schmalle
>>> <apache@teotigraphix.com>wrote:
>>>> Not really,
>>>> I rebuilt everything from scratch. Yes I copied about half the code in
>>>> pieces. I purposely put it all back together myself so I knew what was
>>>> going on. So every class in the committed code was assembled by me, to
>>>> figure out it's function if relevant to the new design.
>>>> Besides most of it had either be deleted of changed because I am not
>>>> targeting SWF what so ever.
>>>> I tried to stick with the same base implementation so we kept the
>>>> multi-threaded Falcon parsing.
>>>> Take a look at the org.apache.flex.compiler.**internal.js.codegen package.
>>>> Specifically ASBlockWalker from that class alone you should see that this
>>>> is a completely different implementation.
>>>> A note to others looking at the code, in the ASBlockWalker I have mixed
>>>> some javascript emitting specific to the closure compiler. I want  
>>>> to change
>>>> this and have a base class not dependent on anything but to be able to
>>>> override it.
>>>> Case in point, most expressions and statements map the same in AS to JS,
>>>> so having a base implementation not tied to anything will be a positive
>>>> thing. I also don't like mixing design specific things in the base
>>>> traversing class, another reason why I want an abstract base or two.
>>>> Anyway, very prototype code and I reserve the right to yank things around.
>>>> :) I just wanted to get it up to show others there might be an easier and
>>>> more flexible way to get to where we need to go without the BURM.
>>>> Mike
>>>> Quoting Alex Harui <aharui@adobe.com>:
>>>>  I will try to look this weekend.
>>>>> Can you briefly describe the important files to look at?  Did  
>>>>> you copy the
>>>>> FalconJS files then do most of your work in a few of them?
>>>>> Thanks,
>>>>> -Alex
>>>>> On 12/13/12 3:37 PM, "Michael Schmalle" <apache@teotigraphix.com>
>>>>>> Hi,
>>>>>> Well, I spent the last 4 days working on this to where it was
>>>>>> something we all could start talking about.
>>>>>> Is it viable?, I really think so. I have spent a lot of time tinkering
>>>>>> with the framework, take a look. It's in my whiteboard for now under
>>>>>> Eclipse projects.
>>>>>> I know there was just a discussion about .project files but I
>>>>>> committed the .project and .classpath for both application and test
>>>>>> project, just like the rest of Falcon.
>>>>>> I'm working on more documentation. A thing to note about the code,
>>>>>> goal is to product ActionScript first, I will explain my thinking
>>>>>> later but, since I'm the one putting this together, that is what
>>>>>> decided was best for testing first. Once we get all ActionScript
>>>>>> generating, we start overriding things for JavaScript specific
>>>>>> implementations.
>>>>>> Source [0]
>>>>>> Right now I have 103 unit tests ALL passing for expressions and
>>>>>> statements. Its a good start.
>>>>>> Note; I have not don't a build file, if anybody wants to go for it.
>>>>>> Please, I hate them. :)
>>>>>> Peace,
>>>>>> Mike
>>>>>> - [0] https://svn.apache.org/repos/**asf/incubator/flex/whiteboard/**
>>>>>> mschmalle/<https://svn.apache.org/repos/asf/incubator/flex/whiteboard/mschmalle/>
>>>>> --
>>>>> Alex Harui
>>>>> Flex SDK Team
>>>>> Adobe Systems, Inc.
>>>>> http://blogs.adobe.com/aharui
>>>> --
>>>> Michael Schmalle - Teoti Graphix, LLC
>>>> http://www.teotigraphix.com
>>>> http://blog.teotigraphix.com

Michael Schmalle - Teoti Graphix, LLC

View raw message