You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@flex.apache.org by Raju Bitter <ra...@googlemail.com> on 2012/01/05 11:33:07 UTC
Falcon compiler source code / Falcon architecture
Alex,
I read your reply in the asdoctool dicussion:
2012/1/4 Alex Harui <ah...@adobe.com>:
> Keep in mind that Adobe hopes to contribute Falcon at some later point in
> time which seems to have a much cleaner code base to work from. Although it
> isn't available now. Also keep in mind that the compiler source may not be
> in the first drop to Apache.
Wouldn't it be important if the community had access to the Falcon
source at a very early stage? If a bunch of people start working on
the SDK tools, and that work will have to be redone once the Falcon
source code will be available, that doesn't seem to make much sense.
Is there any documentation available explaining the differences
between Falcon and the current version of the compiler? Or do you have
a developer wiki for Falcon which might be shared with us, even if we
don't have access to the source code, yet.
Does the Falcon team already have an understanding of how much more
work they want to do before the source code can be shared with Apache?
Raju
Re: Falcon compiler source code / Falcon architecture
Posted by Michael Schmalle <mi...@teotigraphix.com>.
> Is there any documentation available explaining the differences
> between Falcon and the current version of the compiler? Or do you have
> a developer wiki for Falcon which might be shared with us, even if we
> don't have access to the source code, yet.
>
> Does the Falcon team already have an understanding of how much more
> work they want to do before the source code can be shared with Apache?
>
> Raju
>
I am also interested in this and now that I have had time to sleep on it,
I really doubt I will get into any code with the compiler or asdoc before
there is a clear definition with what has changed or something is released.
I'm still not entirly clear on what I would plug into from the out side or
even have the ability to modify to create plugins for the outside.
I would be more interested in a simple explanation of the Falcon architecture
seeing that this was built from the ground up, I'm hoping there was some
thoughtful design put into it's core. :)
Mike
Re: Falcon compiler source code / Falcon architecture
Posted by jude <fl...@gmail.com>.
On Tue, Jan 17, 2012 at 6:48 PM, Alex Harui <ah...@adobe.com> wrote:
> For example, an MXML tag resulted in a UIComponentDescriptor for MX
> components and a tree of functions for Spark components. The plan was to
> have an MXML tag result in an entry in an Array that consisted of the class
> followed by property name/value pairs. The SDK would be modified to look
> for that array and call new and apply properties as needed. Prototypes
> showed that this was not only would this provide independence of SDK
> versions (since the mapping of MXML tag names and attributes to array
> entries is essentially version independent), but also faster because it
> takes advantage of the JIT better than running the Spark tree of individual
> functions once.
>
Can you expand on this? I've been using the model below
(ComponentDescription) to create a working component tree but I've been
planning to debating to extend UIComponentDescriptor.
What I'm using now is shown below:
/**
* Contains information about components that can be
* added or removed from the display list
* */
[Bindable]
public class ComponentDescription {
public function ComponentDescription(element:Object = null):void {
if (element) {
name = NameUtil.getUnqualifiedClassName(element);
className = getQualifiedClassName(element);
instance = element;
}
}
/**
* Unqualified class name
* */
public var name:String;
/**
* Qualified class name
* */
public var className:String;
/**
* Class used to create component instance
* */
public var classType:Object;
/**
* Class or path to icon
* */
public var icon:Object;
/**
* Default properties
* */
public var defaultProperties:Object;
/**
* Default styles
* */
public var defaultStyles:Object;
/**
* Properties
* */
public var properties:Object;
/**
* Styles
* */
public var styles:Object;
/**
* Instance of component. Optional.
* Used for display list
* */
public var instance:Object;
/**
* Children. Optional.
* Used for display in heiarchy view such as Tree.
* */
public var children:ArrayCollection;
/**
* Parent
* */
public var parent:ComponentDescription;
/**
* Skin class. If it doesn't exist the component
* may not be able to be added to the display list.
* */
public var skin:Class;
/**
* Used to store if the instance is visible.
* Does not check if an ancestor is visible
* @see parentVisible
* */
public var visible:Boolean = true;
/**
* Used to store if an ancestor is not visible.
* Manually set
* */
public var parentVisible:Boolean = true;
/**
*
* */
public var locked:Boolean = false;
}
Re: Falcon compiler source code / Falcon architecture
Posted by Alex Harui <ah...@adobe.com>.
On 1/17/12 4:18 PM, "peter.elst@gmail.com" <pe...@gmail.com> wrote:
> Am I reading it right that this means FB is pretty much useless as an IDE
> if Falcon isn't aligned with Apache Flex SDK releases?
It depends on what you want to change. Our goal for Falcon and the SDK back
before the "announcement" was to have MXML compilation generate less code so
that the kinds of things we typically changed from release to release would
all be in the SDK and not in the compiler, so the compiler would be
independent of SDK releases.
For example, an MXML tag resulted in a UIComponentDescriptor for MX
components and a tree of functions for Spark components. The plan was to
have an MXML tag result in an entry in an Array that consisted of the class
followed by property name/value pairs. The SDK would be modified to look
for that array and call new and apply properties as needed. Prototypes
showed that this was not only would this provide independence of SDK
versions (since the mapping of MXML tag names and attributes to array
entries is essentially version independent), but also faster because it
takes advantage of the JIT better than running the Spark tree of individual
functions once.
--
Alex Harui
Flex SDK Team
Adobe Systems, Inc.
http://blogs.adobe.com/aharui
Re: Falcon compiler source code / Falcon architecture
Posted by Peter Elst <pe...@gmail.com>.
> The version of FB that integrates Falcon will not be designed to use a
> different version of Falcon for different versions of the SDK. This is
> different than today, where FB uses the MXMLC from the SDK version for your
> project.
>
> This has some interesting implications. One is that we can't keep rolling
> out new versions of Falcon without factoring in backward compatible MXML
> code-gen for older SDKs. Another is that, if you muck with the language,
> the code-intelligence in FB will break or at minimum, won't be able to
> assist in those new language constructs.
Am I reading it right that this means FB is pretty much useless as an IDE
if Falcon isn't aligned with Apache Flex SDK releases?
If so, sounds like this could be serious a bottleneck for getting releases
out.
- Peter
RE: Falcon compiler source code / Falcon architecture
Posted by Gordon Smith <go...@adobe.com>.
> given the amount of work required to get Flex going as a project, I have to say it would be much less work to donate it [Falcon] within the Flex project
Good point!
- Gordon Smith, Adobe
-----Original Message-----
From: Alex Harui [mailto:aharui@adobe.com]
Sent: Tuesday, January 17, 2012 4:53 PM
To: flex-dev@incubator.apache.org; peter.elst@gmail.com; rajubitter@gmail.com
Subject: Re: Falcon compiler source code / Falcon architecture
On 1/17/12 4:38 PM, "Gordon Smith" <go...@adobe.com> wrote:
> Falcon can compile pure-AS apps that aren't built with the Flex
> framework so I don't think it makes sense as a subproject of Flex.
>
> - Gordon Smith, Adobe
If Falcon is going to have a distinct community then yes, it should have its own project, but given the amount of work required to get Flex going as a project, I have to say it would be much less work to donate it within the Flex project. However, I plan to worry more about that in April or May.
--
Alex Harui
Flex SDK Team
Adobe Systems, Inc.
http://blogs.adobe.com/aharui
Re: Falcon compiler source code / Falcon architecture
Posted by Alex Harui <ah...@adobe.com>.
On 1/17/12 4:38 PM, "Gordon Smith" <go...@adobe.com> wrote:
> Falcon can compile pure-AS apps that aren't built with the Flex framework so I
> don't think it makes sense as a subproject of Flex.
>
> - Gordon Smith, Adobe
If Falcon is going to have a distinct community then yes, it should have its
own project, but given the amount of work required to get Flex going as a
project, I have to say it would be much less work to donate it within the
Flex project. However, I plan to worry more about that in April or May.
--
Alex Harui
Flex SDK Team
Adobe Systems, Inc.
http://blogs.adobe.com/aharui
RE: Falcon compiler source code / Falcon architecture
Posted by Gordon Smith <go...@adobe.com>.
Falcon can compile pure-AS apps that aren't built with the Flex framework so I don't think it makes sense as a subproject of Flex.
- Gordon Smith, Adobe
-----Original Message-----
From: Peter Elst [mailto:peter.elst@gmail.com]
Sent: Tuesday, January 17, 2012 4:25 PM
To: flex-dev@incubator.apache.org; rajubitter@gmail.com
Subject: Re: Falcon compiler source code / Falcon architecture
> Do you think that it's still desirable for Falcon to be a separate ASF
> project?
If I understood correctly there is quite a dependency between the two - in which case perhaps a subproject would make more sense?
- Peter
Re: Falcon compiler source code / Falcon architecture
Posted by Peter Elst <pe...@gmail.com>.
> Do you think that it's still desirable for Falcon to be a separate ASF
> project?
If I understood correctly there is quite a dependency between the two - in
which case perhaps a subproject would make more sense?
- Peter
Re: Falcon compiler source code / Falcon architecture
Posted by Raju Bitter <ra...@googlemail.com>.
Thanks for the detailed update on Falcon. Sounds very promising, and
it's understandable that it will take some time until the Falcon code
base is at a stage where Adobe feels comfortable with contributing the
code. And mid-2012 sounds much better than the end of the year.
Do you think that it's still desirable for Falcon to be a separate ASF project?
- Raju
RE: Falcon compiler source code / Falcon architecture
Posted by Michael Schmalle <mi...@teotigraphix.com>.
Thanks for answering those questions Gordon.
It sounds like this once settled in could be a tool developers dream
come true. I'm really excited about this and I'm sure it will be worth
the wait.
Having some experience in this area, with just the information you
gave(API and such) I can already start to think about some projects I
want to do to extend it or develop tools that execute it and use it's
DOM.
Thanks again!
Mike
Quoting Gordon Smith <go...@adobe.com>:
>> Does the ANTLR parser use the standard token stream if it's not
>> lexing the source?
>
> For AS, where we use JFlex to do the lexing (for better
> performance), we implement our own token stream.
>
> For CSS, where ANTLR is doing the lexing, we use its standard token stream.
>
>> Would there be a way to create a linked list token stream with what
>> the lexer creates?
>
> We don't buffer visited tokens for AS, but it would be
> straightforward to implement your own linked list token stream.
>
>> Is there any plan on adding FXG grammar and DOM?
>
> I forgot to mention that Falcon does support compiling FXG. But we
> didn't innovate much in this area... we mainly just brought over the
> FXG code from the old compiler, which you can take a look at. I
> believe it uses a SAX parser to parse the XML of the FXG file, and
> it builds a custom DOM with classes like RectNode and TextNode which
> it then uses for semantic analysis and code generation. One
> improvement is that the FXG compilation takes place in the same
> multithreaded fashion as other Falcon compilation.
>
> - Gordon
>
> -----Original Message-----
> From: Michael Schmalle [mailto:mike@teotigraphix.com]
> Sent: Thursday, January 19, 2012 4:48 PM
> To: flex-dev@incubator.apache.org
> Subject: RE: Falcon compiler source code / Falcon architecture
>
> Gordon, when you have time;
>
> Does the ANTLR parser use the standard token stream if it's not
> lexing the source?
>
> Would there be a way to create a linked list token stream with what
> the lexer creates?
>
> Is there any plan on adding FXG grammar and DOM (this would be
> awesome but maybe something the community could implement)?
>
> Thanks,
> Mike
>
>
> Quoting Gordon Smith <go...@adobe.com>:
>
>> ANTLR is used to parse (but not lex) AS, and to both lex and parse CSS.
>>
>> - Gordon.
>>
>> -----Original Message-----
>> From: Michael Schmalle [mailto:mike@teotigraphix.com]
>> Sent: Thursday, January 19, 2012 11:16 AM
>> To: flex-dev@incubator.apache.org
>> Subject: RE: Falcon compiler source code / Falcon architecture
>>
>> Gordon this is about the best news we could get.
>>
>> It's great to see this structure! I can't wait.
>>
>> What parts are using ANTLR grammar?
>>
>> Thanks,
>> Mike
>>
>> Quoting Gordon Smith <go...@adobe.com>:
>>
>>> Until we release specs, Javadoc, or code for Falcon, here is a very
>>> brief overview of Falcon's architecture:
>>>
>>> Falcon is designed from the beginning to support compiling multiple
>>> targets in multiple projects in a workspace (as you have in an IDE
>>> like Flash Builder). By contrast, asc was designed to compile a
>>> single AS file, and then mxmlc was built on top of that.
>>>
>>> Falcon uses multiple threads to compile multiple files at the same
>>> time. The more cores you have, the faster it goes.
>>>
>>> Critical data structures such as the symbol table (which stores
>>> information about which classes are known, what methods they have,
>>> etc.) are shared across the entire workspace, to minimize memory
>>> usage.
>>>
>>> Critical data structures are maintained in memory to support both
>>> compilation and IDS code intelligence in an efficient and consistent
>>> way. For example: If you open a file in the IDE, Falcon builds a
>>> syntax tree and symbol table for it to support intelligent editing.
>>> Compiling the file requires just one additional code generation step.
>>> By contrast, in the current Flash Builder which uses the compiler in
>>> the SDK, Flash Builder builds its own parse trees and symbol tables
>>> to support editing, and then when you compile the compiler in the SDK
>>> builds another set of parse trees and symbol tables. This is slow and
>>> a waste of memory.
>>>
>>> Falcon understands .as, .mxml, .css, and .properties files.
>>>
>>> The parse trees for AS consist of nodes from about 100 classes, such
>>> as LiteralNode, BinaryOperatorNode, FunctionNode, ClassNode. The
>>> parse trees for MXML consist of nodes from about 50 classes, such as
>>> MXMLDocumentNode, MXMLInstanceNode, MXMLScriptNode, etc. The
>>> ActionScript-y parts of MXML are represented by AS nodes inside of
>>> MXML nodes.
>>>
>>> The symbol table consists of objects representing the things your
>>> code defines, such as ClassDefiniition, FunctionDefinition, and
>>> VariableDefinition, arranged into a hierarchy of scope objects.
>>>
>>> Falcon uses 3rd-party grammars like JFlex, ANTLR, and JBurg to
>>> generate various lexers, parsers, and code generators.
>>>
>>> MXML is compiled directly to ABC, not to ActionScript source code or
>>> an ActionScript parse tree.
>>>
>>> - Gordon Smith, Adobe
>>>
>>> P.S. After working on the Flex framework since its inception for
>>> about
>>> 8 years, I joined the Falcon team 18 months ago.
>>>
>>>
>>>
>>> -----Original Message-----
>>> From: Raju Bitter [mailto:rajubitter@googlemail.com]
>>> Sent: Thursday, January 19, 2012 2:10 AM
>>> To: flex-dev@incubator.apache.org
>>> Subject: Re: Falcon compiler source code / Falcon architecture
>>>
>>> 2012/1/19 Alex Harui <ah...@adobe.com>:
>>>> I was in the Falcon code in November. It was way easier to figure
>>>> out how to make changes than with MXMLC. Even if it is not
>>>> documented, I think it will be easier to make progress that with the
>>>> Falcon code base.
>>> I agree, that sounds like it would be much better to use the Falcon
>>> code base. There are normally fewer community members with much
>>> experience in language/compiler design, and the cleaner the
>>> architecture is, the better for the community.
>>>
>>
>>
>>
>
>
>
RE: Falcon compiler source code / Falcon architecture
Posted by Gordon Smith <go...@adobe.com>.
> Does the ANTLR parser use the standard token stream if it's not lexing the source?
For AS, where we use JFlex to do the lexing (for better performance), we implement our own token stream.
For CSS, where ANTLR is doing the lexing, we use its standard token stream.
> Would there be a way to create a linked list token stream with what the lexer creates?
We don't buffer visited tokens for AS, but it would be straightforward to implement your own linked list token stream.
> Is there any plan on adding FXG grammar and DOM?
I forgot to mention that Falcon does support compiling FXG. But we didn't innovate much in this area... we mainly just brought over the FXG code from the old compiler, which you can take a look at. I believe it uses a SAX parser to parse the XML of the FXG file, and it builds a custom DOM with classes like RectNode and TextNode which it then uses for semantic analysis and code generation. One improvement is that the FXG compilation takes place in the same multithreaded fashion as other Falcon compilation.
- Gordon
-----Original Message-----
From: Michael Schmalle [mailto:mike@teotigraphix.com]
Sent: Thursday, January 19, 2012 4:48 PM
To: flex-dev@incubator.apache.org
Subject: RE: Falcon compiler source code / Falcon architecture
Gordon, when you have time;
Does the ANTLR parser use the standard token stream if it's not lexing the source?
Would there be a way to create a linked list token stream with what the lexer creates?
Is there any plan on adding FXG grammar and DOM (this would be awesome but maybe something the community could implement)?
Thanks,
Mike
Quoting Gordon Smith <go...@adobe.com>:
> ANTLR is used to parse (but not lex) AS, and to both lex and parse CSS.
>
> - Gordon.
>
> -----Original Message-----
> From: Michael Schmalle [mailto:mike@teotigraphix.com]
> Sent: Thursday, January 19, 2012 11:16 AM
> To: flex-dev@incubator.apache.org
> Subject: RE: Falcon compiler source code / Falcon architecture
>
> Gordon this is about the best news we could get.
>
> It's great to see this structure! I can't wait.
>
> What parts are using ANTLR grammar?
>
> Thanks,
> Mike
>
> Quoting Gordon Smith <go...@adobe.com>:
>
>> Until we release specs, Javadoc, or code for Falcon, here is a very
>> brief overview of Falcon's architecture:
>>
>> Falcon is designed from the beginning to support compiling multiple
>> targets in multiple projects in a workspace (as you have in an IDE
>> like Flash Builder). By contrast, asc was designed to compile a
>> single AS file, and then mxmlc was built on top of that.
>>
>> Falcon uses multiple threads to compile multiple files at the same
>> time. The more cores you have, the faster it goes.
>>
>> Critical data structures such as the symbol table (which stores
>> information about which classes are known, what methods they have,
>> etc.) are shared across the entire workspace, to minimize memory
>> usage.
>>
>> Critical data structures are maintained in memory to support both
>> compilation and IDS code intelligence in an efficient and consistent
>> way. For example: If you open a file in the IDE, Falcon builds a
>> syntax tree and symbol table for it to support intelligent editing.
>> Compiling the file requires just one additional code generation step.
>> By contrast, in the current Flash Builder which uses the compiler in
>> the SDK, Flash Builder builds its own parse trees and symbol tables
>> to support editing, and then when you compile the compiler in the SDK
>> builds another set of parse trees and symbol tables. This is slow and
>> a waste of memory.
>>
>> Falcon understands .as, .mxml, .css, and .properties files.
>>
>> The parse trees for AS consist of nodes from about 100 classes, such
>> as LiteralNode, BinaryOperatorNode, FunctionNode, ClassNode. The
>> parse trees for MXML consist of nodes from about 50 classes, such as
>> MXMLDocumentNode, MXMLInstanceNode, MXMLScriptNode, etc. The
>> ActionScript-y parts of MXML are represented by AS nodes inside of
>> MXML nodes.
>>
>> The symbol table consists of objects representing the things your
>> code defines, such as ClassDefiniition, FunctionDefinition, and
>> VariableDefinition, arranged into a hierarchy of scope objects.
>>
>> Falcon uses 3rd-party grammars like JFlex, ANTLR, and JBurg to
>> generate various lexers, parsers, and code generators.
>>
>> MXML is compiled directly to ABC, not to ActionScript source code or
>> an ActionScript parse tree.
>>
>> - Gordon Smith, Adobe
>>
>> P.S. After working on the Flex framework since its inception for
>> about
>> 8 years, I joined the Falcon team 18 months ago.
>>
>>
>>
>> -----Original Message-----
>> From: Raju Bitter [mailto:rajubitter@googlemail.com]
>> Sent: Thursday, January 19, 2012 2:10 AM
>> To: flex-dev@incubator.apache.org
>> Subject: Re: Falcon compiler source code / Falcon architecture
>>
>> 2012/1/19 Alex Harui <ah...@adobe.com>:
>>> I was in the Falcon code in November. It was way easier to figure
>>> out how to make changes than with MXMLC. Even if it is not
>>> documented, I think it will be easier to make progress that with the
>>> Falcon code base.
>> I agree, that sounds like it would be much better to use the Falcon
>> code base. There are normally fewer community members with much
>> experience in language/compiler design, and the cleaner the
>> architecture is, the better for the community.
>>
>
>
>
Re: Falcon compiler source code / Falcon architecture
Posted by Rick Winscot <ri...@gmail.com>.
With regard to:
"It can compile an app against a library project without creating the SWC for that library on disk."
...sounds to me like Falcon could* serve as a replacement for the JavaScript interpreter. Interesting.
http://www.quilix.com/node/96
--
Rick Winscot
On Friday, January 20, 2012 at 1:34 PM, Gordon Smith wrote:
> Here is some other possibly-interesting info about Falcon:
>
> It was designed from the beginning for incremental compilation of a complete workspace, so it tracks dependencies between compilation units in various projects. (I.e., if you edit a file, it knows which other files it should recompile.)
>
> It can compile an app against a library project without creating the SWC for that library on disk.
>
> It tracks where in the source every token/tree node/definition came from, to support operations like go-to-definition in Flash Builder.
>
> It has the general concept of a "problem", where a problem can end up being considered an error, a warning, or ignored.
>
> It provides live problem highlighting to Flash Builder. When you pause during editing a file, the file is compiled and the problems show up as red squiggly underlines.
>
> It tries to continue after almost all problems, so that you don't necessarily have to fix every problem before you can get a SWF to run.
>
> And these are probably obvious:
>
> It includes code for reading and writing SWFs and SWCs.
>
> It includes code for reading and writing ActionScript byte code, at the level of individual instructions.
>
> It includes transcoders for the various types of embedded assets, which read, for example, a PNG file and convert it to the form required inside a SWF.
>
> - Gordon
>
> -----Original Message-----
> From: Gordon Smith
> Sent: Friday, January 20, 2012 10:17 AM
> To: flex-dev@incubator.apache.org (mailto:flex-dev@incubator.apache.org)
> Subject: RE: Falcon compiler source code / Falcon architecture
>
> > Does the ANTLR parser use the standard token stream if it's not lexing the source?
>
> For AS, where we use JFlex to do the lexing (for better performance), we implement our own token stream.
>
> For CSS, where ANTLR is doing the lexing, we use its standard token stream.
>
> > Would there be a way to create a linked list token stream with what the lexer creates?
>
> We don't buffer visited tokens for AS, but it would be straightforward to implement your own linked list token stream.
>
> > Is there any plan on adding FXG grammar and DOM?
>
> I forgot to mention that Falcon does support compiling FXG. But we didn't innovate much in this area... we mainly just brought over the FXG code from the old compiler, which you can take a look at. I believe it uses a SAX parser to parse the XML of the FXG file, and it builds a custom DOM with classes like RectNode and TextNode which it then uses for semantic analysis and code generation. One improvement is that the FXG compilation takes place in the same multithreaded fashion as other Falcon compilation.
>
> - Gordon
>
> -----Original Message-----
> From: Michael Schmalle [mailto:mike@teotigraphix.com]
> Sent: Thursday, January 19, 2012 4:48 PM
> To: flex-dev@incubator.apache.org (mailto:flex-dev@incubator.apache.org)
> Subject: RE: Falcon compiler source code / Falcon architecture
>
> Gordon, when you have time;
>
> Does the ANTLR parser use the standard token stream if it's not lexing the source?
>
> Would there be a way to create a linked list token stream with what the lexer creates?
>
> Is there any plan on adding FXG grammar and DOM (this would be awesome but maybe something the community could implement)?
>
> Thanks,
> Mike
>
>
> Quoting Gordon Smith <gosmith@adobe.com (mailto:gosmith@adobe.com)>:
>
> > ANTLR is used to parse (but not lex) AS, and to both lex and parse CSS.
> >
> > - Gordon.
> >
> > -----Original Message-----
> > From: Michael Schmalle [mailto:mike@teotigraphix.com]
> > Sent: Thursday, January 19, 2012 11:16 AM
> > To: flex-dev@incubator.apache.org (mailto:flex-dev@incubator.apache.org)
> > Subject: RE: Falcon compiler source code / Falcon architecture
> >
> > Gordon this is about the best news we could get.
> >
> > It's great to see this structure! I can't wait.
> >
> > What parts are using ANTLR grammar?
> >
> > Thanks,
> > Mike
> >
> > Quoting Gordon Smith <gosmith@adobe.com (mailto:gosmith@adobe.com)>:
> >
> > > Until we release specs, Javadoc, or code for Falcon, here is a very
> > > brief overview of Falcon's architecture:
> > >
> > > Falcon is designed from the beginning to support compiling multiple
> > > targets in multiple projects in a workspace (as you have in an IDE
> > > like Flash Builder). By contrast, asc was designed to compile a
> > > single AS file, and then mxmlc was built on top of that.
> > >
> > > Falcon uses multiple threads to compile multiple files at the same
> > > time. The more cores you have, the faster it goes.
> > >
> > > Critical data structures such as the symbol table (which stores
> > > information about which classes are known, what methods they have,
> > > etc.) are shared across the entire workspace, to minimize memory
> > > usage.
> > >
> > > Critical data structures are maintained in memory to support both
> > > compilation and IDS code intelligence in an efficient and consistent
> > > way. For example: If you open a file in the IDE, Falcon builds a
> > > syntax tree and symbol table for it to support intelligent editing.
> > > Compiling the file requires just one additional code generation step.
> > > By contrast, in the current Flash Builder which uses the compiler in
> > > the SDK, Flash Builder builds its own parse trees and symbol tables
> > > to support editing, and then when you compile the compiler in the SDK
> > > builds another set of parse trees and symbol tables. This is slow and
> > > a waste of memory.
> > >
> > > Falcon understands .as, .mxml, .css, and .properties files.
> > >
> > > The parse trees for AS consist of nodes from about 100 classes, such
> > > as LiteralNode, BinaryOperatorNode, FunctionNode, ClassNode. The
> > > parse trees for MXML consist of nodes from about 50 classes, such as
> > > MXMLDocumentNode, MXMLInstanceNode, MXMLScriptNode, etc. The
> > > ActionScript-y parts of MXML are represented by AS nodes inside of
> > > MXML nodes.
> > >
> > > The symbol table consists of objects representing the things your
> > > code defines, such as ClassDefiniition, FunctionDefinition, and
> > > VariableDefinition, arranged into a hierarchy of scope objects.
> > >
> > > Falcon uses 3rd-party grammars like JFlex, ANTLR, and JBurg to
> > > generate various lexers, parsers, and code generators.
> > >
> > > MXML is compiled directly to ABC, not to ActionScript source code or
> > > an ActionScript parse tree.
> > >
> > > - Gordon Smith, Adobe
> > >
> > > P.S. After working on the Flex framework since its inception for
> > > about
> > > 8 years, I joined the Falcon team 18 months ago.
> > >
> > >
> > >
> > > -----Original Message-----
> > > From: Raju Bitter [mailto:rajubitter@googlemail.com]
> > > Sent: Thursday, January 19, 2012 2:10 AM
> > > To: flex-dev@incubator.apache.org (mailto:flex-dev@incubator.apache.org)
> > > Subject: Re: Falcon compiler source code / Falcon architecture
> > >
> > > 2012/1/19 Alex Harui <aharui@adobe.com (mailto:aharui@adobe.com)>:
> > > > I was in the Falcon code in November. It was way easier to figure
> > > > out how to make changes than with MXMLC. Even if it is not
> > > > documented, I think it will be easier to make progress that with the
> > > > Falcon code base.
> > > >
> > >
> > > I agree, that sounds like it would be much better to use the Falcon
> > > code base. There are normally fewer community members with much
> > > experience in language/compiler design, and the cleaner the
> > > architecture is, the better for the community.
> > >
> >
> >
>
>
>
RE: Falcon compiler source code / Falcon architecture
Posted by Gordon Smith <go...@adobe.com>.
Here is some other possibly-interesting info about Falcon:
It was designed from the beginning for incremental compilation of a complete workspace, so it tracks dependencies between compilation units in various projects. (I.e., if you edit a file, it knows which other files it should recompile.)
It can compile an app against a library project without creating the SWC for that library on disk.
It tracks where in the source every token/tree node/definition came from, to support operations like go-to-definition in Flash Builder.
It has the general concept of a "problem", where a problem can end up being considered an error, a warning, or ignored.
It provides live problem highlighting to Flash Builder. When you pause during editing a file, the file is compiled and the problems show up as red squiggly underlines.
It tries to continue after almost all problems, so that you don't necessarily have to fix every problem before you can get a SWF to run.
And these are probably obvious:
It includes code for reading and writing SWFs and SWCs.
It includes code for reading and writing ActionScript byte code, at the level of individual instructions.
It includes transcoders for the various types of embedded assets, which read, for example, a PNG file and convert it to the form required inside a SWF.
- Gordon
-----Original Message-----
From: Gordon Smith
Sent: Friday, January 20, 2012 10:17 AM
To: flex-dev@incubator.apache.org
Subject: RE: Falcon compiler source code / Falcon architecture
> Does the ANTLR parser use the standard token stream if it's not lexing the source?
For AS, where we use JFlex to do the lexing (for better performance), we implement our own token stream.
For CSS, where ANTLR is doing the lexing, we use its standard token stream.
> Would there be a way to create a linked list token stream with what the lexer creates?
We don't buffer visited tokens for AS, but it would be straightforward to implement your own linked list token stream.
> Is there any plan on adding FXG grammar and DOM?
I forgot to mention that Falcon does support compiling FXG. But we didn't innovate much in this area... we mainly just brought over the FXG code from the old compiler, which you can take a look at. I believe it uses a SAX parser to parse the XML of the FXG file, and it builds a custom DOM with classes like RectNode and TextNode which it then uses for semantic analysis and code generation. One improvement is that the FXG compilation takes place in the same multithreaded fashion as other Falcon compilation.
- Gordon
-----Original Message-----
From: Michael Schmalle [mailto:mike@teotigraphix.com]
Sent: Thursday, January 19, 2012 4:48 PM
To: flex-dev@incubator.apache.org
Subject: RE: Falcon compiler source code / Falcon architecture
Gordon, when you have time;
Does the ANTLR parser use the standard token stream if it's not lexing the source?
Would there be a way to create a linked list token stream with what the lexer creates?
Is there any plan on adding FXG grammar and DOM (this would be awesome but maybe something the community could implement)?
Thanks,
Mike
Quoting Gordon Smith <go...@adobe.com>:
> ANTLR is used to parse (but not lex) AS, and to both lex and parse CSS.
>
> - Gordon.
>
> -----Original Message-----
> From: Michael Schmalle [mailto:mike@teotigraphix.com]
> Sent: Thursday, January 19, 2012 11:16 AM
> To: flex-dev@incubator.apache.org
> Subject: RE: Falcon compiler source code / Falcon architecture
>
> Gordon this is about the best news we could get.
>
> It's great to see this structure! I can't wait.
>
> What parts are using ANTLR grammar?
>
> Thanks,
> Mike
>
> Quoting Gordon Smith <go...@adobe.com>:
>
>> Until we release specs, Javadoc, or code for Falcon, here is a very
>> brief overview of Falcon's architecture:
>>
>> Falcon is designed from the beginning to support compiling multiple
>> targets in multiple projects in a workspace (as you have in an IDE
>> like Flash Builder). By contrast, asc was designed to compile a
>> single AS file, and then mxmlc was built on top of that.
>>
>> Falcon uses multiple threads to compile multiple files at the same
>> time. The more cores you have, the faster it goes.
>>
>> Critical data structures such as the symbol table (which stores
>> information about which classes are known, what methods they have,
>> etc.) are shared across the entire workspace, to minimize memory
>> usage.
>>
>> Critical data structures are maintained in memory to support both
>> compilation and IDS code intelligence in an efficient and consistent
>> way. For example: If you open a file in the IDE, Falcon builds a
>> syntax tree and symbol table for it to support intelligent editing.
>> Compiling the file requires just one additional code generation step.
>> By contrast, in the current Flash Builder which uses the compiler in
>> the SDK, Flash Builder builds its own parse trees and symbol tables
>> to support editing, and then when you compile the compiler in the SDK
>> builds another set of parse trees and symbol tables. This is slow and
>> a waste of memory.
>>
>> Falcon understands .as, .mxml, .css, and .properties files.
>>
>> The parse trees for AS consist of nodes from about 100 classes, such
>> as LiteralNode, BinaryOperatorNode, FunctionNode, ClassNode. The
>> parse trees for MXML consist of nodes from about 50 classes, such as
>> MXMLDocumentNode, MXMLInstanceNode, MXMLScriptNode, etc. The
>> ActionScript-y parts of MXML are represented by AS nodes inside of
>> MXML nodes.
>>
>> The symbol table consists of objects representing the things your
>> code defines, such as ClassDefiniition, FunctionDefinition, and
>> VariableDefinition, arranged into a hierarchy of scope objects.
>>
>> Falcon uses 3rd-party grammars like JFlex, ANTLR, and JBurg to
>> generate various lexers, parsers, and code generators.
>>
>> MXML is compiled directly to ABC, not to ActionScript source code or
>> an ActionScript parse tree.
>>
>> - Gordon Smith, Adobe
>>
>> P.S. After working on the Flex framework since its inception for
>> about
>> 8 years, I joined the Falcon team 18 months ago.
>>
>>
>>
>> -----Original Message-----
>> From: Raju Bitter [mailto:rajubitter@googlemail.com]
>> Sent: Thursday, January 19, 2012 2:10 AM
>> To: flex-dev@incubator.apache.org
>> Subject: Re: Falcon compiler source code / Falcon architecture
>>
>> 2012/1/19 Alex Harui <ah...@adobe.com>:
>>> I was in the Falcon code in November. It was way easier to figure
>>> out how to make changes than with MXMLC. Even if it is not
>>> documented, I think it will be easier to make progress that with the
>>> Falcon code base.
>> I agree, that sounds like it would be much better to use the Falcon
>> code base. There are normally fewer community members with much
>> experience in language/compiler design, and the cleaner the
>> architecture is, the better for the community.
>>
>
>
>
RE: Falcon compiler source code / Falcon architecture
Posted by Michael Schmalle <mi...@teotigraphix.com>.
Gordon, when you have time;
Does the ANTLR parser use the standard token stream if it's not lexing
the source?
Would there be a way to create a linked list token stream with what
the lexer creates?
Is there any plan on adding FXG grammar and DOM (this would be awesome
but maybe something the community could implement)?
Thanks,
Mike
Quoting Gordon Smith <go...@adobe.com>:
> ANTLR is used to parse (but not lex) AS, and to both lex and parse CSS.
>
> - Gordon.
>
> -----Original Message-----
> From: Michael Schmalle [mailto:mike@teotigraphix.com]
> Sent: Thursday, January 19, 2012 11:16 AM
> To: flex-dev@incubator.apache.org
> Subject: RE: Falcon compiler source code / Falcon architecture
>
> Gordon this is about the best news we could get.
>
> It's great to see this structure! I can't wait.
>
> What parts are using ANTLR grammar?
>
> Thanks,
> Mike
>
> Quoting Gordon Smith <go...@adobe.com>:
>
>> Until we release specs, Javadoc, or code for Falcon, here is a very
>> brief overview of Falcon's architecture:
>>
>> Falcon is designed from the beginning to support compiling multiple
>> targets in multiple projects in a workspace (as you have in an IDE
>> like Flash Builder). By contrast, asc was designed to compile a single
>> AS file, and then mxmlc was built on top of that.
>>
>> Falcon uses multiple threads to compile multiple files at the same
>> time. The more cores you have, the faster it goes.
>>
>> Critical data structures such as the symbol table (which stores
>> information about which classes are known, what methods they have,
>> etc.) are shared across the entire workspace, to minimize memory
>> usage.
>>
>> Critical data structures are maintained in memory to support both
>> compilation and IDS code intelligence in an efficient and consistent
>> way. For example: If you open a file in the IDE, Falcon builds a
>> syntax tree and symbol table for it to support intelligent editing.
>> Compiling the file requires just one additional code generation step.
>> By contrast, in the current Flash Builder which uses the compiler in
>> the SDK, Flash Builder builds its own parse trees and symbol tables to
>> support editing, and then when you compile the compiler in the SDK
>> builds another set of parse trees and symbol tables. This is slow and
>> a waste of memory.
>>
>> Falcon understands .as, .mxml, .css, and .properties files.
>>
>> The parse trees for AS consist of nodes from about 100 classes, such
>> as LiteralNode, BinaryOperatorNode, FunctionNode, ClassNode. The parse
>> trees for MXML consist of nodes from about 50 classes, such as
>> MXMLDocumentNode, MXMLInstanceNode, MXMLScriptNode, etc. The
>> ActionScript-y parts of MXML are represented by AS nodes inside of
>> MXML nodes.
>>
>> The symbol table consists of objects representing the things your code
>> defines, such as ClassDefiniition, FunctionDefinition, and
>> VariableDefinition, arranged into a hierarchy of scope objects.
>>
>> Falcon uses 3rd-party grammars like JFlex, ANTLR, and JBurg to
>> generate various lexers, parsers, and code generators.
>>
>> MXML is compiled directly to ABC, not to ActionScript source code or
>> an ActionScript parse tree.
>>
>> - Gordon Smith, Adobe
>>
>> P.S. After working on the Flex framework since its inception for about
>> 8 years, I joined the Falcon team 18 months ago.
>>
>>
>>
>> -----Original Message-----
>> From: Raju Bitter [mailto:rajubitter@googlemail.com]
>> Sent: Thursday, January 19, 2012 2:10 AM
>> To: flex-dev@incubator.apache.org
>> Subject: Re: Falcon compiler source code / Falcon architecture
>>
>> 2012/1/19 Alex Harui <ah...@adobe.com>:
>>> I was in the Falcon code in November. It was way easier to figure
>>> out how to make changes than with MXMLC. Even if it is not
>>> documented, I think it will be easier to make progress that with
>>> the Falcon code base.
>> I agree, that sounds like it would be much better to use the Falcon
>> code base. There are normally fewer community members with much
>> experience in language/compiler design, and the cleaner the
>> architecture is, the better for the community.
>>
>
>
>
RE: Falcon compiler source code / Falcon architecture
Posted by Gordon Smith <go...@adobe.com>.
ANTLR is used to parse (but not lex) AS, and to both lex and parse CSS.
- Gordon.
-----Original Message-----
From: Michael Schmalle [mailto:mike@teotigraphix.com]
Sent: Thursday, January 19, 2012 11:16 AM
To: flex-dev@incubator.apache.org
Subject: RE: Falcon compiler source code / Falcon architecture
Gordon this is about the best news we could get.
It's great to see this structure! I can't wait.
What parts are using ANTLR grammar?
Thanks,
Mike
Quoting Gordon Smith <go...@adobe.com>:
> Until we release specs, Javadoc, or code for Falcon, here is a very
> brief overview of Falcon's architecture:
>
> Falcon is designed from the beginning to support compiling multiple
> targets in multiple projects in a workspace (as you have in an IDE
> like Flash Builder). By contrast, asc was designed to compile a single
> AS file, and then mxmlc was built on top of that.
>
> Falcon uses multiple threads to compile multiple files at the same
> time. The more cores you have, the faster it goes.
>
> Critical data structures such as the symbol table (which stores
> information about which classes are known, what methods they have,
> etc.) are shared across the entire workspace, to minimize memory
> usage.
>
> Critical data structures are maintained in memory to support both
> compilation and IDS code intelligence in an efficient and consistent
> way. For example: If you open a file in the IDE, Falcon builds a
> syntax tree and symbol table for it to support intelligent editing.
> Compiling the file requires just one additional code generation step.
> By contrast, in the current Flash Builder which uses the compiler in
> the SDK, Flash Builder builds its own parse trees and symbol tables to
> support editing, and then when you compile the compiler in the SDK
> builds another set of parse trees and symbol tables. This is slow and
> a waste of memory.
>
> Falcon understands .as, .mxml, .css, and .properties files.
>
> The parse trees for AS consist of nodes from about 100 classes, such
> as LiteralNode, BinaryOperatorNode, FunctionNode, ClassNode. The parse
> trees for MXML consist of nodes from about 50 classes, such as
> MXMLDocumentNode, MXMLInstanceNode, MXMLScriptNode, etc. The
> ActionScript-y parts of MXML are represented by AS nodes inside of
> MXML nodes.
>
> The symbol table consists of objects representing the things your code
> defines, such as ClassDefiniition, FunctionDefinition, and
> VariableDefinition, arranged into a hierarchy of scope objects.
>
> Falcon uses 3rd-party grammars like JFlex, ANTLR, and JBurg to
> generate various lexers, parsers, and code generators.
>
> MXML is compiled directly to ABC, not to ActionScript source code or
> an ActionScript parse tree.
>
> - Gordon Smith, Adobe
>
> P.S. After working on the Flex framework since its inception for about
> 8 years, I joined the Falcon team 18 months ago.
>
>
>
> -----Original Message-----
> From: Raju Bitter [mailto:rajubitter@googlemail.com]
> Sent: Thursday, January 19, 2012 2:10 AM
> To: flex-dev@incubator.apache.org
> Subject: Re: Falcon compiler source code / Falcon architecture
>
> 2012/1/19 Alex Harui <ah...@adobe.com>:
>> I was in the Falcon code in November. It was way easier to figure
>> out how to make changes than with MXMLC. Even if it is not
>> documented, I think it will be easier to make progress that with the Falcon code base.
> I agree, that sounds like it would be much better to use the Falcon
> code base. There are normally fewer community members with much
> experience in language/compiler design, and the cleaner the
> architecture is, the better for the community.
>
RE: Falcon compiler source code / Falcon architecture
Posted by Michael Schmalle <mi...@teotigraphix.com>.
Gordon this is about the best news we could get.
It's great to see this structure! I can't wait.
What parts are using ANTLR grammar?
Thanks,
Mike
Quoting Gordon Smith <go...@adobe.com>:
> Until we release specs, Javadoc, or code for Falcon, here is a very
> brief overview of Falcon's architecture:
>
> Falcon is designed from the beginning to support compiling multiple
> targets in multiple projects in a workspace (as you have in an IDE
> like Flash Builder). By contrast, asc was designed to compile a
> single AS file, and then mxmlc was built on top of that.
>
> Falcon uses multiple threads to compile multiple files at the same
> time. The more cores you have, the faster it goes.
>
> Critical data structures such as the symbol table (which stores
> information about which classes are known, what methods they have,
> etc.) are shared across the entire workspace, to minimize memory
> usage.
>
> Critical data structures are maintained in memory to support both
> compilation and IDS code intelligence in an efficient and consistent
> way. For example: If you open a file in the IDE, Falcon builds a
> syntax tree and symbol table for it to support intelligent editing.
> Compiling the file requires just one additional code generation
> step. By contrast, in the current Flash Builder which uses the
> compiler in the SDK, Flash Builder builds its own parse trees and
> symbol tables to support editing, and then when you compile the
> compiler in the SDK builds another set of parse trees and symbol
> tables. This is slow and a waste of memory.
>
> Falcon understands .as, .mxml, .css, and .properties files.
>
> The parse trees for AS consist of nodes from about 100 classes, such
> as LiteralNode, BinaryOperatorNode, FunctionNode, ClassNode. The
> parse trees for MXML consist of nodes from about 50 classes, such as
> MXMLDocumentNode, MXMLInstanceNode, MXMLScriptNode, etc. The
> ActionScript-y parts of MXML are represented by AS nodes inside of
> MXML nodes.
>
> The symbol table consists of objects representing the things your
> code defines, such as ClassDefiniition, FunctionDefinition, and
> VariableDefinition, arranged into a hierarchy of scope objects.
>
> Falcon uses 3rd-party grammars like JFlex, ANTLR, and JBurg to
> generate various lexers, parsers, and code generators.
>
> MXML is compiled directly to ABC, not to ActionScript source code or
> an ActionScript parse tree.
>
> - Gordon Smith, Adobe
>
> P.S. After working on the Flex framework since its inception for
> about 8 years, I joined the Falcon team 18 months ago.
>
>
>
> -----Original Message-----
> From: Raju Bitter [mailto:rajubitter@googlemail.com]
> Sent: Thursday, January 19, 2012 2:10 AM
> To: flex-dev@incubator.apache.org
> Subject: Re: Falcon compiler source code / Falcon architecture
>
> 2012/1/19 Alex Harui <ah...@adobe.com>:
>> I was in the Falcon code in November. It was way easier to figure out
>> how to make changes than with MXMLC. Even if it is not documented, I
>> think it will be easier to make progress that with the Falcon code base.
> I agree, that sounds like it would be much better to use the Falcon
> code base. There are normally fewer community members with much
> experience in language/compiler design, and the cleaner the
> architecture is, the better for the community.
>
RE: Falcon compiler source code / Falcon architecture
Posted by Gordon Smith <go...@adobe.com>.
Until we release specs, Javadoc, or code for Falcon, here is a very brief overview of Falcon's architecture:
Falcon is designed from the beginning to support compiling multiple targets in multiple projects in a workspace (as you have in an IDE like Flash Builder). By contrast, asc was designed to compile a single AS file, and then mxmlc was built on top of that.
Falcon uses multiple threads to compile multiple files at the same time. The more cores you have, the faster it goes.
Critical data structures such as the symbol table (which stores information about which classes are known, what methods they have, etc.) are shared across the entire workspace, to minimize memory usage.
Critical data structures are maintained in memory to support both compilation and IDS code intelligence in an efficient and consistent way. For example: If you open a file in the IDE, Falcon builds a syntax tree and symbol table for it to support intelligent editing. Compiling the file requires just one additional code generation step. By contrast, in the current Flash Builder which uses the compiler in the SDK, Flash Builder builds its own parse trees and symbol tables to support editing, and then when you compile the compiler in the SDK builds another set of parse trees and symbol tables. This is slow and a waste of memory.
Falcon understands .as, .mxml, .css, and .properties files.
The parse trees for AS consist of nodes from about 100 classes, such as LiteralNode, BinaryOperatorNode, FunctionNode, ClassNode. The parse trees for MXML consist of nodes from about 50 classes, such as MXMLDocumentNode, MXMLInstanceNode, MXMLScriptNode, etc. The ActionScript-y parts of MXML are represented by AS nodes inside of MXML nodes.
The symbol table consists of objects representing the things your code defines, such as ClassDefiniition, FunctionDefinition, and VariableDefinition, arranged into a hierarchy of scope objects.
Falcon uses 3rd-party grammars like JFlex, ANTLR, and JBurg to generate various lexers, parsers, and code generators.
MXML is compiled directly to ABC, not to ActionScript source code or an ActionScript parse tree.
- Gordon Smith, Adobe
P.S. After working on the Flex framework since its inception for about 8 years, I joined the Falcon team 18 months ago.
-----Original Message-----
From: Raju Bitter [mailto:rajubitter@googlemail.com]
Sent: Thursday, January 19, 2012 2:10 AM
To: flex-dev@incubator.apache.org
Subject: Re: Falcon compiler source code / Falcon architecture
2012/1/19 Alex Harui <ah...@adobe.com>:
> I was in the Falcon code in November. It was way easier to figure out
> how to make changes than with MXMLC. Even if it is not documented, I
> think it will be easier to make progress that with the Falcon code base.
I agree, that sounds like it would be much better to use the Falcon code base. There are normally fewer community members with much experience in language/compiler design, and the cleaner the architecture is, the better for the community.
Re: Falcon compiler source code / Falcon architecture
Posted by Raju Bitter <ra...@googlemail.com>.
2012/1/19 Alex Harui <ah...@adobe.com>:
> I was in the Falcon code in November. It was way easier to figure out how
> to make changes than with MXMLC. Even if it is not documented, I think it
> will be easier to make progress that with the Falcon code base.
I agree, that sounds like it would be much better to use the Falcon
code base. There are normally fewer community members with much
experience in language/compiler design, and the cleaner the
architecture is, the better for the community.
Re: Falcon compiler source code / Falcon architecture
Posted by Andras Csizmadia <an...@vpmedia.eu>.
There is one open-source ActionScript -> JavaScript compiler available
called Jangaroo,
I'm not sure how close it is from production quality (most of the FP10 API's
are not implemented regarding to release notes):
http://www.jangaroo.net/home/
https://github.com/CoreMedia/jangaroo-tools
https://github.com/CoreMedia/jangaroo-tools/wiki/
https://github.com/CoreMedia/jangaroo-tools/wiki/Release-Notes
http://blogs.adobe.com/bparadie/2011/12/01/jangaroo-and-falconjs/
--
Also i'm testing HaxeNME in my free time (http://www.haxenme.org/) which
uses Jeash (http://jeash.com/) as an extension to compile Haxe code to HTML5
Canvas application.
But there are some missing things which hold up to port any major AS3
library to Haxe (like overriding native super getter/setter properties).
I think the improvement of the Haxe language should be much faster, there
are things which were requested years ago in their mailing list and it's
still open.
Best regards,
Andrew
-----Original Message-----
From: Scott Delamater
Sent: Wednesday, January 18, 2012 3:36 PM
To: flex-dev@incubator.apache.org
Subject: Re: Falcon compiler source code / Falcon architecture
On Wed, Jan 18, 2012 at 12:59 AM, David Arno <da...@davidarno.org> wrote:
> Should we just ignore Falcon and instead consider a major overhaul of
mxmlc,
> or even a completely new compiler?
It sounds like, at best, we might be able to follow Falcon development
updates (and thanks in advance to Alex for keeping us apprised).
But if it's not being done in the open, it doesn't feel like the kind of
thing we should (or can?) use to steer decisions. That'd be a very tenuous
coupling.
An alternate compiler would be very interesting, at least.
Re: Falcon compiler source code / Falcon architecture
Posted by Scott Delamater <sc...@gmail.com>.
On Wed, Jan 18, 2012 at 12:59 AM, David Arno <da...@davidarno.org> wrote:
> Should we just ignore Falcon and instead consider a major overhaul of
mxmlc,
> or even a completely new compiler?
It sounds like, at best, we might be able to follow Falcon development
updates (and thanks in advance to Alex for keeping us apprised).
But if it's not being done in the open, it doesn't feel like the kind of
thing we should (or can?) use to steer decisions. That'd be a very tenuous
coupling.
An alternate compiler would be very interesting, at least.
RE: Falcon compiler source code / Falcon architecture
Posted by Gordon Smith <go...@adobe.com>.
> I guess it depends on how long we have to wait to get access to the mxmlc source
Well, you don't have to wait until it gets into Apache. It's been available for years at http://opensource.adobe.com/flex.
> I don't like the idea that we will have to wait six months for a half-working compiler, presumably with no APIs, documentation etc. provided beforehand.
I'll try to push the idea the we make the Javadoc for Falcon available sooner.
- Gordon Smith, Adobe
-----Original Message-----
From: David Arno [mailto:david@davidarno.org]
Sent: Wednesday, January 18, 2012 12:00 AM
To: flex-dev@incubator.apache.org
Subject: RE: Falcon compiler source code / Falcon architecture
From: Alex Harui [mailto:aharui@adobe.com]
Sent: 18 January 2012 00:08
>
> 1) Falcon is intended to serve not only as the compiler for FB, but as
> the code-model service for FB's code intelligence.
> 2) The version of FB that integrates Falcon will not be designed to
> use a different version of Falcon for different versions of the SDK.
> This is different than today, where FB uses the MXMLC from the SDK
> version for your project.
> ... That work will likely not be fully complete at the time of donation.
The more I read about Falcon, the more concerned I become about it. I don't like the idea that we will have to wait six months for a half-working compiler, presumably with no APIs, documentation etc provided beforehand.
There is huge risks associated with this: there is a great danger that we will lose too much community momentum in that time and we would have to put a great deal of trust into Adobe to deliver during a period of significant upheaval and ventures into the unknown..
Should we just ignore Falcon and instead consider a major overhaul of mxmlc, or even a completely new compiler? I guess it depends on how long we have to wait to get access to the mxmlc source and docs as to what the best way forward is.
David.
Re: Falcon compiler source code / Falcon architecture
Posted by Alex Harui <ah...@adobe.com>.
On 1/17/12 11:59 PM, "David Arno" <da...@davidarno.org> wrote:
>
> The more I read about Falcon, the more concerned I become about it. I don't
> like the idea that we will have to wait six months for a half-working
> compiler, presumably with no APIs, documentation etc provided beforehand.
> There is huge risks associated with this: there is a great danger that we
> will lose too much community momentum in that time and we would have to put
> a great deal of trust into Adobe to deliver during a period of significant
> upheaval and ventures into the unknown..
Well, I guess you can say it will be "half-working", but really, the AS
portion should be fully working, and the MXML portion should be working for
simple cases.
I suppose Apache Flex has the option of writing its own MXML-to-AS front
end, but I'm not sure that will perform as well. Falcon goes MXML-to-ABC
and I believe that is one of the reasons it is faster. For us to start from
scratch and not leverage the finished AS work would probably still take
longer. I will say, though, that the MXML portion of Falcon is still
generating MX- and Spark-specific code. An alternative plan would be to
alter the SDK code to handle a single kind of codegen. That would make MXML
compiling simpler and potentially faster, but isn't backward-compatible and
could break some apps.
>
> Should we just ignore Falcon and instead consider a major overhaul of mxmlc,
> or even a completely new compiler? I guess it depends on how long we have to
> wait to get access to the mxmlc source and docs as to what the best way
> forward is.
I was in the Falcon code in November. It was way easier to figure out how
to make changes than with MXMLC. Even if it is not documented, I think it
will be easier to make progress that with the Falcon code base.
--
Alex Harui
Flex SDK Team
Adobe Systems, Inc.
http://blogs.adobe.com/aharui
RE: Falcon compiler source code / Falcon architecture
Posted by David Arno <da...@davidarno.org>.
From: Alex Harui [mailto:aharui@adobe.com]
Sent: 18 January 2012 00:08
>
> 1) Falcon is intended to serve not only as the compiler for FB,
> but as the code-model service for FB's code intelligence.
> 2) The version of FB that integrates Falcon will not be designed
> to use a different version of Falcon for different versions of the SDK.
> This is different than today, where FB uses the MXMLC from the
> SDK version for your project.
> ... That work will likely not be fully complete at the time of donation.
The more I read about Falcon, the more concerned I become about it. I don't
like the idea that we will have to wait six months for a half-working
compiler, presumably with no APIs, documentation etc provided beforehand.
There is huge risks associated with this: there is a great danger that we
will lose too much community momentum in that time and we would have to put
a great deal of trust into Adobe to deliver during a period of significant
upheaval and ventures into the unknown..
Should we just ignore Falcon and instead consider a major overhaul of mxmlc,
or even a completely new compiler? I guess it depends on how long we have to
wait to get access to the mxmlc source and docs as to what the best way
forward is.
David.
Re: Falcon compiler source code / Falcon architecture
Posted by Alex Harui <ah...@adobe.com>.
While we're on the subject, I learned a few things in my meetings today that
may not have occurred to folks wanting to change the compiler and language.
1) Falcon is intended to serve not only as the compiler for FB, but as the
code-model service for FB's code intelligence.
2) The version of FB that integrates Falcon will not be designed to use a
different version of Falcon for different versions of the SDK. This is
different than today, where FB uses the MXMLC from the SDK version for your
project.
This has some interesting implications. One is that we can't keep rolling
out new versions of Falcon without factoring in backward compatible MXML
code-gen for older SDKs. Another is that, if you muck with the language,
the code-intelligence in FB will break or at minimum, won't be able to
assist in those new language constructs.
For the first issue, the plan of record was to have Falcon convert MXML
constructs to data structures instead of code, and modify the SDK framework
to interpret those data structures. That work will likely not be fully
complete at the time of donation.
Anyway, keeps these facts in mind as you plan the future for Falcon.
-Alex
On 1/17/12 2:37 PM, "Jonathan Campos" <jo...@gmail.com> wrote:
> On Tue, Jan 17, 2012 at 4:35 PM, Roland Zwaga <ro...@stackandheap.com>wrote:
>
>> I agree, sorry if I sounded impatient :) And I understand the legal battles
>> that Alex et al have to wage
>> to get the code out there earlier than the original timeline.
>>
>
> "Soon" can't be soon enough in my book. The earlier the better.
--
Alex Harui
Flex SDK Team
Adobe Systems, Inc.
http://blogs.adobe.com/aharui
Re: Falcon compiler source code / Falcon architecture
Posted by Jonathan Campos <jo...@gmail.com>.
On Tue, Jan 17, 2012 at 4:35 PM, Roland Zwaga <ro...@stackandheap.com>wrote:
> I agree, sorry if I sounded impatient :) And I understand the legal battles
> that Alex et al have to wage
> to get the code out there earlier than the original timeline.
>
"Soon" can't be soon enough in my book. The earlier the better.
--
Jonathan Campos
Re: Falcon compiler source code / Falcon architecture
Posted by Roland Zwaga <ro...@stackandheap.com>.
I agree, sorry if I sounded impatient :) And I understand the legal battles
that Alex et al have to wage
to get the code out there earlier than the original timeline.
> Wow, the summer, that's still quite a while away.
>
>
> Sure, but given the initial timeline that was given at the Flex Summit
> having something by summer would be a huge win IMHO.
>
> In the worst case and community participation doesn't help speed up getting
> a stable release of Falcon out, it still gives the community exposure to
> whats going on.
>
Re: Falcon compiler source code / Falcon architecture
Posted by Peter Elst <pe...@gmail.com>.
> Wow, the summer, that's still quite a while away.
Sure, but given the initial timeline that was given at the Flex Summit
having something by summer would be a huge win IMHO.
In the worst case and community participation doesn't help speed up getting
a stable release of Falcon out, it still gives the community exposure to
whats going on.
- Peter
Re: Falcon compiler source code / Falcon architecture
Posted by Roland Zwaga <ro...@stackandheap.com>.
Wow, the summer, that's still quite a while away. That is, if you get the
proposal approved.
Good luck with that Alex, until that time we'll all be reading up on
compiler theory :)
Okay, obviously still campaigning for open development on this.
>
> I don't have any official timeline for Falcon at this time. Hopefully
> there will be a more official timeline in about a week. I was in two
> meetings today on this very topic. The proposal we are trying to get
> approved would have the community getting the unfinished Falcon code in the
> summer.
>
RE: Falcon compiler source code / Falcon architecture
Posted by "Michael A. Labriola" <la...@digitalprimates.net>.
Okay, obviously still campaigning for open development on this.
-----Original Message-----
From: Alex Harui [mailto:aharui@adobe.com]
Sent: Tuesday, January 17, 2012 4:08 PM
To: flex-dev@incubator.apache.org
Subject: Re: Falcon compiler source code / Falcon architecture
On 1/17/12 1:01 PM, "Michael A. Labriola" <la...@digitalprimates.net>
wrote:
> Alex any movement on this?
>
> My hope is still that Adobe will donate this thing and finish
> development in the open. It would send a strong commitment to Adobe
> wanting to work with the community.
I don't have any official timeline for Falcon at this time. Hopefully there will be a more official timeline in about a week. I was in two meetings today on this very topic. The proposal we are trying to get approved would have the community getting the unfinished Falcon code in the summer.
-Alex
Notice: This transmission is intended only for the use of the individual or entity to which it is addressed and may contain information that is privileged or confidential. Any dissemination, distribution or copying of this transmission by anyone other than the intended recipient is strictly prohibited. If you have received this transmission in error, please notify the sender immediately by e-mail or telephone and delete the original transmission. Thank you.
Re: Falcon compiler source code / Falcon architecture
Posted by Alex Harui <ah...@adobe.com>.
On 1/17/12 1:01 PM, "Michael A. Labriola" <la...@digitalprimates.net>
wrote:
> Alex any movement on this?
>
> My hope is still that Adobe will donate this thing and finish development in
> the open. It would send a strong commitment to Adobe wanting to work with the
> community.
I don't have any official timeline for Falcon at this time. Hopefully there
will be a more official timeline in about a week. I was in two meetings
today on this very topic. The proposal we are trying to get approved would
have the community getting the unfinished Falcon code in the summer.
-Alex
RE: Falcon compiler source code / Falcon architecture
Posted by "Michael A. Labriola" <la...@digitalprimates.net>.
Alex any movement on this?
My hope is still that Adobe will donate this thing and finish development in the open. It would send a strong commitment to Adobe wanting to work with the community.
-----Original Message-----
From: Alex Harui [mailto:aharui@adobe.com]
Sent: Thursday, January 05, 2012 1:29 PM
To: flex-dev@incubator.apache.org
Subject: Re: Falcon compiler source code / Falcon architecture
The goal is to donate Falcon before it is done, but it won't be before we want to cut some release from our Flex podling. Yes, Falcon looks like it could make your build script run twice as fast, but it won't be the 10x the early prototypes showed for incremental compiles from the tool.
One the other hand, MXMLC could compile twice as fast if it was compiling a more modular framework and less stuff to think about.
I fully expect Falcon to become available and a major factor in the long run.
On 1/5/12 11:19 AM, "Roland Zwaga" <ro...@stackandheap.com> wrote:
> From what I remember from the Flex summit talks, there is a fair
> chance that Adobe will release the compiler front end earlier than planned.
> In that case it would be entirely possible that someone or someones in
> the community will build their own MXML frontend into the compiler.
> Adobe can still build their own MXML frontend and use it in
> Flashbuilder 2012, so they wouldn't necessarily jeopardise their
> releases. All in all, they projected they'd have the entire thing
> finished in 2013! We really don't want to wait THAT long, with the
> community's support this could be sped up quite a bit, plus we'll be
> able to build extensions that Adobe hasn't had time to dream up :) The
> fact that we can't use the compiler in Flashbuilder also doesn't
> concern me too much, if we can hook it up using build scripts I'd be
> more than happy to go that route for a while.
> So I agree competely with the fact that the new compiler is
> potentially a lot more important than the SDK, perhaps not in
> short-term, but in the long-run it is going to make all the difference.
>
> cheers,
>
> Roland
>
>
> On 5 January 2012 19:17, Michael Schmalle <mi...@teotigraphix.com> wrote:
>
>> Quoting Alex Harui <ah...@adobe.com>:
>>
>> The compiler IS the Flex SDK and to empower developers to get
>> involved
>>>> would be a huge new driving force for the architecture built to use it.
>>>>
>>>
>> I don't agree with this. I can use the compiler to generate code
>> for a
>>> whole new framework if I want to.
>>>
>>> --
>>> Alex Harui
>>> Flex SDK Team
>>> Adobe Systems, Inc.
>>> http://blogs.adobe.com/aharui
>>>
>>
>> Ok,
>>
>> Poorly chosen words for exactly what you said. My point is the
>> compiler spawns the empowerment I am speaking of, take the Flex SDK
>> out of it. My main driving force of understanding is "the thingy that
>> reads ActionScript3 source code and puts it into a model AST
>> hopefully that a developer like myself could use to create anything".
>>
>> This parallels what you said Alex, "I can use the compiler to
>> generate code for a whole new framework if I want to."
>>
>> I have been vocal on this because I just can imagine what we can do
>> with a pure AS3 model compiled by the Flacon parser. I am not to
>> concerned about MXML because that is actually XML that could be
>> converted into AST that is used by the actual parser, I have done this in my own projects.
>>
>> I also get that Adobe has plans for it, so I will say no more on this
>> topic at this time.
>>
>> Mike
>>
>>
>>
>>
>>
>
--
Alex Harui
Flex SDK Team
Adobe Systems, Inc.
http://blogs.adobe.com/aharui
Notice: This transmission is intended only for the use of the individual or entity to which it is addressed and may contain information that is privileged or confidential. Any dissemination, distribution or copying of this transmission by anyone other than the intended recipient is strictly prohibited. If you have received this transmission in error, please notify the sender immediately by e-mail or telephone and delete the original transmission. Thank you.
Re: Falcon compiler source code / Falcon architecture
Posted by Alex Harui <ah...@adobe.com>.
The goal is to donate Falcon before it is done, but it won't be before we
want to cut some release from our Flex podling. Yes, Falcon looks like it
could make your build script run twice as fast, but it won't be the 10x the
early prototypes showed for incremental compiles from the tool.
One the other hand, MXMLC could compile twice as fast if it was compiling a
more modular framework and less stuff to think about.
I fully expect Falcon to become available and a major factor in the long
run.
On 1/5/12 11:19 AM, "Roland Zwaga" <ro...@stackandheap.com> wrote:
> From what I remember from the Flex summit talks, there is a fair chance
> that Adobe will release the compiler front end earlier than planned.
> In that case it would be entirely possible that someone or someones in the
> community will build their own MXML frontend into the compiler. Adobe can
> still build their own MXML frontend and use it in Flashbuilder 2012, so
> they wouldn't necessarily jeopardise their releases. All in all, they
> projected they'd have the entire thing finished in 2013! We really don't
> want to wait THAT long, with the community's support this could be sped up
> quite a bit, plus we'll be able to build extensions that Adobe hasn't had
> time to dream up :)
> The fact that we can't use the compiler in Flashbuilder also doesn't
> concern me too much, if we can hook it up using build scripts I'd be more
> than happy to go that route for a while.
> So I agree competely with the fact that the new compiler is potentially a
> lot more important than the SDK, perhaps not in short-term, but in the
> long-run it is going to make all the difference.
>
> cheers,
>
> Roland
>
>
> On 5 January 2012 19:17, Michael Schmalle <mi...@teotigraphix.com> wrote:
>
>> Quoting Alex Harui <ah...@adobe.com>:
>>
>> The compiler IS the Flex SDK and to empower developers to get involved
>>>> would be a huge new driving force for the architecture built to use it.
>>>>
>>>
>> I don't agree with this. I can use the compiler to generate code for a
>>> whole new framework if I want to.
>>>
>>> --
>>> Alex Harui
>>> Flex SDK Team
>>> Adobe Systems, Inc.
>>> http://blogs.adobe.com/aharui
>>>
>>
>> Ok,
>>
>> Poorly chosen words for exactly what you said. My point is the compiler
>> spawns the empowerment I am speaking of, take the Flex SDK out of it. My
>> main driving force of understanding is "the thingy that reads ActionScript3
>> source code and puts it into a model AST hopefully that a developer like
>> myself could use to create anything".
>>
>> This parallels what you said Alex, "I can use the compiler to generate
>> code for a whole new framework if I want to."
>>
>> I have been vocal on this because I just can imagine what we can do with a
>> pure AS3 model compiled by the Flacon parser. I am not to concerned about
>> MXML because that is actually XML that could be converted into AST that is
>> used by the actual parser, I have done this in my own projects.
>>
>> I also get that Adobe has plans for it, so I will say no more on this
>> topic at this time.
>>
>> Mike
>>
>>
>>
>>
>>
>
--
Alex Harui
Flex SDK Team
Adobe Systems, Inc.
http://blogs.adobe.com/aharui
Re: Falcon compiler source code / Falcon architecture
Posted by Roland Zwaga <ro...@stackandheap.com>.
>From what I remember from the Flex summit talks, there is a fair chance
that Adobe will release the compiler front end earlier than planned.
In that case it would be entirely possible that someone or someones in the
community will build their own MXML frontend into the compiler. Adobe can
still build their own MXML frontend and use it in Flashbuilder 2012, so
they wouldn't necessarily jeopardise their releases. All in all, they
projected they'd have the entire thing finished in 2013! We really don't
want to wait THAT long, with the community's support this could be sped up
quite a bit, plus we'll be able to build extensions that Adobe hasn't had
time to dream up :)
The fact that we can't use the compiler in Flashbuilder also doesn't
concern me too much, if we can hook it up using build scripts I'd be more
than happy to go that route for a while.
So I agree competely with the fact that the new compiler is potentially a
lot more important than the SDK, perhaps not in short-term, but in the
long-run it is going to make all the difference.
cheers,
Roland
On 5 January 2012 19:17, Michael Schmalle <mi...@teotigraphix.com> wrote:
> Quoting Alex Harui <ah...@adobe.com>:
>
> The compiler IS the Flex SDK and to empower developers to get involved
>>> would be a huge new driving force for the architecture built to use it.
>>>
>>
> I don't agree with this. I can use the compiler to generate code for a
>> whole new framework if I want to.
>>
>> --
>> Alex Harui
>> Flex SDK Team
>> Adobe Systems, Inc.
>> http://blogs.adobe.com/aharui
>>
>
> Ok,
>
> Poorly chosen words for exactly what you said. My point is the compiler
> spawns the empowerment I am speaking of, take the Flex SDK out of it. My
> main driving force of understanding is "the thingy that reads ActionScript3
> source code and puts it into a model AST hopefully that a developer like
> myself could use to create anything".
>
> This parallels what you said Alex, "I can use the compiler to generate
> code for a whole new framework if I want to."
>
> I have been vocal on this because I just can imagine what we can do with a
> pure AS3 model compiled by the Flacon parser. I am not to concerned about
> MXML because that is actually XML that could be converted into AST that is
> used by the actual parser, I have done this in my own projects.
>
> I also get that Adobe has plans for it, so I will say no more on this
> topic at this time.
>
> Mike
>
>
>
>
>
--
regards,
Roland
--
Roland Zwaga
Senior Consultant | Stack & Heap BVBA
+32 (0)486 16 12 62 | roland@stackandheap.com | http://www.stackandheap.com
Re: Falcon compiler source code / Falcon architecture
Posted by Michael Schmalle <mi...@teotigraphix.com>.
Quoting Alex Harui <ah...@adobe.com>:
>> The compiler IS the Flex SDK and to empower developers to get involved
>> would be a huge new driving force for the architecture built to use it.
> I don't agree with this. I can use the compiler to generate code for a
> whole new framework if I want to.
>
> --
> Alex Harui
> Flex SDK Team
> Adobe Systems, Inc.
> http://blogs.adobe.com/aharui
Ok,
Poorly chosen words for exactly what you said. My point is the
compiler spawns the empowerment I am speaking of, take the Flex SDK
out of it. My main driving force of understanding is "the thingy that
reads ActionScript3 source code and puts it into a model AST hopefully
that a developer like myself could use to create anything".
This parallels what you said Alex, "I can use the compiler to generate
code for a whole new framework if I want to."
I have been vocal on this because I just can imagine what we can do
with a pure AS3 model compiled by the Flacon parser. I am not to
concerned about MXML because that is actually XML that could be
converted into AST that is used by the actual parser, I have done this
in my own projects.
I also get that Adobe has plans for it, so I will say no more on this
topic at this time.
Mike
Re: Falcon compiler source code / Falcon architecture
Posted by Alex Harui <ah...@adobe.com>.
> The compiler IS the Flex SDK and to empower developers to get involved
> would be a huge new driving force for the architecture built to use it.
I don't agree with this. I can use the compiler to generate code for a
whole new framework if I want to.
>>> Wouldn't it be important if the community had access to the Falcon
>>> source at a very early stage?
Yes, but the fact is, Adobe is building Falcon for its own commercial
reasons and based on this experience as to how distracting it is to donate a
project to Apache, it is hard to make an argument that they should risk
their schedules to donate early. But we are pushing to find a drop point
before they are done.
Another thing to keep in mind is that Falcon and its performance improvement
is still mainly leveraged by Flash Builder and incremental compiles. The
current Flash Builder is not designed to take in a replacement compiler at
all so even if you could get the code now, you wouldn't be able to leverage
it until long after we want to cut some releases from this project, so I'm
trying to make sure folks stay focused on finding key things that will make
a difference sooner.
A final note: Falcon's MXML handling is still up in the air. The focus is
on getting AS compiling as that is the direction of use for Falcon within
Adobe. The early drop may not handle MXML at all.
On 1/5/12 4:34 AM, "Michael Schmalle" <mi...@teotigraphix.com> wrote:
> Quoting David Arno <da...@davidarno.org>:
>
>> +1
>>
>>> Wouldn't it be important if the community had access to the Falcon
>>> source at a very early stage?
>>
>> There are some great ideas here around creating an API for the
>> compiler to support formatting and documentation tools for example.
>> Also I'd like to experiment with adding custom generics support to
>> the compiler. I'd love to get involved in this stuff, but see no
>> point in starting it with the mxmlc compiler source since we know
>> falcon is coming at some point. So please can those with the falcon
>> team's ear badger them into releasing to Apache as soon as possible
>> (before it's complete.)
>>
>> David.
>>
>>
>
> I will +1 this as well since I will almost certainly be a major
> contributor/committer to this branch of development. One of my goals
> personally dealing with the Falcon compiler is to create a place that
> I will document everything I do as a notepad for others. So I would
> like to have as much documentation on the developing the compiler as
> say extending a component.
>
> The compiler IS the Flex SDK and to empower developers to get involved
> would be a huge new driving force for the architecture built to use it.
>
> I see this as a very important step in getting Apache Flex into the
> melting pot viewed fairly along side other competing technologies. I
> have had a few blog posts about the transformation of Flex and really,
> this is exactly where it is going to happen on the low level.
>
> Mike
>
>
--
Alex Harui
Flex SDK Team
Adobe Systems, Inc.
http://blogs.adobe.com/aharui
Re: Falcon compiler source code / Falcon architecture
Posted by Fréderic Cox <fr...@exuvis.com>.
>I see this as a very important step in getting Apache Flex into the
>melting pot viewed fairly along side other competing technologies. I
>have had a few blog posts about the transformation of Flex and really,
>this is exactly where it is going to happen on the low level.
>
>Mike
+1
RE: Falcon compiler source code / Falcon architecture
Posted by Michael Schmalle <mi...@teotigraphix.com>.
Quoting David Arno <da...@davidarno.org>:
> +1
>
>> Wouldn't it be important if the community had access to the Falcon
>> source at a very early stage?
>
> There are some great ideas here around creating an API for the
> compiler to support formatting and documentation tools for example.
> Also I'd like to experiment with adding custom generics support to
> the compiler. I'd love to get involved in this stuff, but see no
> point in starting it with the mxmlc compiler source since we know
> falcon is coming at some point. So please can those with the falcon
> team's ear badger them into releasing to Apache as soon as possible
> (before it's complete.)
>
> David.
>
>
I will +1 this as well since I will almost certainly be a major
contributor/committer to this branch of development. One of my goals
personally dealing with the Falcon compiler is to create a place that
I will document everything I do as a notepad for others. So I would
like to have as much documentation on the developing the compiler as
say extending a component.
The compiler IS the Flex SDK and to empower developers to get involved
would be a huge new driving force for the architecture built to use it.
I see this as a very important step in getting Apache Flex into the
melting pot viewed fairly along side other competing technologies. I
have had a few blog posts about the transformation of Flex and really,
this is exactly where it is going to happen on the low level.
Mike
RE: Falcon compiler source code / Falcon architecture
Posted by David Arno <da...@davidarno.org>.
+1
>Wouldn't it be important if the community had access to the Falcon source at a very early stage?
There are some great ideas here around creating an API for the compiler to support formatting and documentation tools for example. Also I'd like to experiment with adding custom generics support to the compiler. I'd love to get involved in this stuff, but see no point in starting it with the mxmlc compiler source since we know falcon is coming at some point. So please can those with the falcon team's ear badger them into releasing to Apache as soon as possible (before it's complete.)
David.