You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@flex.apache.org by Harbs <ha...@gmail.com> on 2016/05/01 14:50:07 UTC
Re: [FalconJX] compiling non-browser javascript
Here’s my stab at producing ActionScript files from the OMV files: https://github.com/unhurdle/omv2as
The output is actually pretty good. I get error-free output on InDesign files with the exception of File types because I don’t yet have the core types linked. Photoshop output is not as good, for the most part because the OMV files types are not all true types.
I do have a question though (before I even got to the point where I’m trying to use this to cross-compile code): When I run the base classes through the app, I get a bunch of classes which do not compile into a SWC very well. At least part of the problem is due to the fact that they confluct with core classes, and I’m not sure how to best handle this. Here’s a link of the as code: https://www.dropbox.com/s/pziyrqj7k1ob9p7/ExtendScript.zip?dl=0
I’m not sure how to best handle this. If anyone has good ideas, please let me know.
On Apr 25, 2016, at 9:28 PM, Harbs <ha...@gmail.com> wrote:
> I was guessing that the release would probably work. I am concerned about debugging though.
>
> I will probably try this suggestion next week and see how far I can get without further help. Chances are I’ll be back here before I’m successful though… ;-)
>
> Thanks!
> Harbs
>
> On Apr 25, 2016, at 6:27 PM, Alex Harui <ah...@adobe.com> wrote:
>
>>
>>
>> On 4/25/16, 8:16 AM, "Josh Tynjala" <jo...@gmail.com> wrote:
>>
>>> In the bin/js-release directory, all of the generated JavaScript is
>>> concatenated into a single file, so it no longer uses goog.require(). That
>>> should work in environments that cannot load multiple scripts.
>>
>> I was about to suggest that as well. By default, the single-file output
>> is minified so is hard to debug. You can add
>> -js-compiler-option="--compilation_level WHITESPACE_ONLY"
>>
>> to the cross-compile and I think you'll still get a single file without
>> goog.require but it will be debuggable.
>>
>> These options are handled by the compiler code in a Publisher.
>> MXMLFlexJSPublisher has this default behavior. You can subclass it and
>> create a different js-output-type get it to spit a single-file to the
>> js-debug and a minified single-file to js-release. It will take a long
>> time, though, as gathering in a single file is done by the Google Closure
>> Compiler. But you don't to know much about compilers to make a custom
>> Publisher. Everything is compiled at that point and you are basically
>> dealing with files and configs for GCC.
>>
>> A harder task is to make the goog.require replaceable with some other
>> pattern.
>>
>> -Alex
>>
>
Re: [FalconJX] compiling non-browser javascript
Posted by Harbs <ha...@gmail.com>.
OK. Looks like that’s correct.
I just did a quick test:
function doSomething(in){
alert(in);
}
and Chrome flagged the function parameter as an unexpected token.
I was having a hard time figuring out what an “Identifier” in ECMAScript is, but at the very least, Chrome considers it illegal. That’s enough for me.
I’ll change reserved words in the generator.
Thanks!
Harbs
On May 2, 2016, at 7:28 AM, Josh Tynjala <jo...@gmail.com> wrote:
> I'll double check the ECMAScript spec to see if reserved words are allowed
> as function parameters. It's only in very specific places where they began
> to be allowed in ES5. This may not be one of those places. If my memory
> serves, I suspect it isn't one of those places, and Falcon is behaving
> correctly by treating that as an error.
>
> - Josh
> We've recently made changes to allow keywords in more places. I
> personally don't enjoy doing that work, but maybe Josh will.
>
> -Alex
>
> On 5/1/16, 12:34 PM, "Harbs" <ha...@gmail.com> wrote:
>
>> I get the same error for other reserved words:
>> in,
>> include,
>> with,
>> etc…
>>
>> On May 1, 2016, at 10:28 PM, Harbs <ha...@gmail.com> wrote:
>>
>>> I manually deleted most of the core classes to get it to compile.
>>>
>>> I’m now getting an error which I don’t know if it’s valid or a bug in
>>> Falcon:
>>>
>>> public function findKeyStrings(for:String):String{return
> null;}
>>> public function
> translateKeyString(for:String):String{return null;}
>>>
>>> When trying to compile a class which contains code like this, I get:
>>> ERROR
>>> /Users/harbs/Desktop/InDesign10.2/src/com/adobe/indesign/Application.as[6
>>> 6:33]:
>>> 'for' is not allowed here
>>>
>>> Is “for” really not allowed as the name of a parameter, or i it a bug
>>> that the compiler thinks it’s a for loop?
>>>
>>> Harbs
>>>
>>> On May 1, 2016, at 3:50 PM, Harbs <ha...@gmail.com> wrote:
>>>
>>>> Here’s my stab at producing ActionScript files from the OMV files:
>>>> https://github.com/unhurdle/omv2as
>>>>
>>>> The output is actually pretty good. I get error-free output on
>>>> InDesign files with the exception of File types because I don’t yet
>>>> have the core types linked. Photoshop output is not as good, for the
>>>> most part because the OMV files types are not all true types.
>>>>
>>>> I do have a question though (before I even got to the point where I’m
>>>> trying to use this to cross-compile code): When I run the base classes
>>>> through the app, I get a bunch of classes which do not compile into a
>>>> SWC very well. At least part of the problem is due to the fact that
>>>> they confluct with core classes, and I’m not sure how to best handle
>>>> this. Here’s a link of the as code:
>>>> https://www.dropbox.com/s/pziyrqj7k1ob9p7/ExtendScript.zip?dl=0
>>>>
>>>> I’m not sure how to best handle this. If anyone has good ideas, please
>>>> let me know.
>>>>
>>>> On Apr 25, 2016, at 9:28 PM, Harbs <ha...@gmail.com> wrote:
>>>>
>>>>> I was guessing that the release would probably work. I am concerned
>>>>> about debugging though.
>>>>>
>>>>> I will probably try this suggestion next week and see how far I can
>>>>> get without further help. Chances are I’ll be back here before I’m
>>>>> successful though… ;-)
>>>>>
>>>>> Thanks!
>>>>> Harbs
>>>>>
>>>>> On Apr 25, 2016, at 6:27 PM, Alex Harui <ah...@adobe.com> wrote:
>>>>>
>>>>>>
>>>>>>
>>>>>> On 4/25/16, 8:16 AM, "Josh Tynjala" <jo...@gmail.com> wrote:
>>>>>>
>>>>>>> In the bin/js-release directory, all of the generated JavaScript is
>>>>>>> concatenated into a single file, so it no longer uses
>>>>>>> goog.require(). That
>>>>>>> should work in environments that cannot load multiple scripts.
>>>>>>
>>>>>> I was about to suggest that as well. By default, the single-file
>>>>>> output
>>>>>> is minified so is hard to debug. You can add
>>>>>> -js-compiler-option="--compilation_level WHITESPACE_ONLY"
>>>>>>
>>>>>> to the cross-compile and I think you'll still get a single file
>>>>>> without
>>>>>> goog.require but it will be debuggable.
>>>>>>
>>>>>> These options are handled by the compiler code in a Publisher.
>>>>>> MXMLFlexJSPublisher has this default behavior. You can subclass it
>>>>>> and
>>>>>> create a different js-output-type get it to spit a single-file to the
>>>>>> js-debug and a minified single-file to js-release. It will take a
>>>>>> long
>>>>>> time, though, as gathering in a single file is done by the Google
>>>>>> Closure
>>>>>> Compiler. But you don't to know much about compilers to make a
>>>>>> custom
>>>>>> Publisher. Everything is compiled at that point and you are
>>>>>> basically
>>>>>> dealing with files and configs for GCC.
>>>>>>
>>>>>> A harder task is to make the goog.require replaceable with some other
>>>>>> pattern.
>>>>>>
>>>>>> -Alex
>>>>>>
>>>>>
>>>>
>>>
>>
Re: [FalconJX] compiling non-browser javascript
Posted by Josh Tynjala <jo...@gmail.com>.
I'll double check the ECMAScript spec to see if reserved words are allowed
as function parameters. It's only in very specific places where they began
to be allowed in ES5. This may not be one of those places. If my memory
serves, I suspect it isn't one of those places, and Falcon is behaving
correctly by treating that as an error.
- Josh
We've recently made changes to allow keywords in more places. I
personally don't enjoy doing that work, but maybe Josh will.
-Alex
On 5/1/16, 12:34 PM, "Harbs" <ha...@gmail.com> wrote:
>I get the same error for other reserved words:
>in,
>include,
>with,
>etc…
>
>On May 1, 2016, at 10:28 PM, Harbs <ha...@gmail.com> wrote:
>
>> I manually deleted most of the core classes to get it to compile.
>>
>> I’m now getting an error which I don’t know if it’s valid or a bug in
>>Falcon:
>>
>> public function findKeyStrings(for:String):String{return
null;}
>> public function
translateKeyString(for:String):String{return null;}
>>
>> When trying to compile a class which contains code like this, I get:
>> ERROR
>>/Users/harbs/Desktop/InDesign10.2/src/com/adobe/indesign/Application.as[6
>>6:33]:
>> 'for' is not allowed here
>>
>> Is “for” really not allowed as the name of a parameter, or i it a bug
>>that the compiler thinks it’s a for loop?
>>
>> Harbs
>>
>> On May 1, 2016, at 3:50 PM, Harbs <ha...@gmail.com> wrote:
>>
>>> Here’s my stab at producing ActionScript files from the OMV files:
>>>https://github.com/unhurdle/omv2as
>>>
>>> The output is actually pretty good. I get error-free output on
>>>InDesign files with the exception of File types because I don’t yet
>>>have the core types linked. Photoshop output is not as good, for the
>>>most part because the OMV files types are not all true types.
>>>
>>> I do have a question though (before I even got to the point where I’m
>>>trying to use this to cross-compile code): When I run the base classes
>>>through the app, I get a bunch of classes which do not compile into a
>>>SWC very well. At least part of the problem is due to the fact that
>>>they confluct with core classes, and I’m not sure how to best handle
>>>this. Here’s a link of the as code:
>>>https://www.dropbox.com/s/pziyrqj7k1ob9p7/ExtendScript.zip?dl=0
>>>
>>> I’m not sure how to best handle this. If anyone has good ideas, please
>>>let me know.
>>>
>>> On Apr 25, 2016, at 9:28 PM, Harbs <ha...@gmail.com> wrote:
>>>
>>>> I was guessing that the release would probably work. I am concerned
>>>>about debugging though.
>>>>
>>>> I will probably try this suggestion next week and see how far I can
>>>>get without further help. Chances are I’ll be back here before I’m
>>>>successful though… ;-)
>>>>
>>>> Thanks!
>>>> Harbs
>>>>
>>>> On Apr 25, 2016, at 6:27 PM, Alex Harui <ah...@adobe.com> wrote:
>>>>
>>>>>
>>>>>
>>>>> On 4/25/16, 8:16 AM, "Josh Tynjala" <jo...@gmail.com> wrote:
>>>>>
>>>>>> In the bin/js-release directory, all of the generated JavaScript is
>>>>>> concatenated into a single file, so it no longer uses
>>>>>>goog.require(). That
>>>>>> should work in environments that cannot load multiple scripts.
>>>>>
>>>>> I was about to suggest that as well. By default, the single-file
>>>>>output
>>>>> is minified so is hard to debug. You can add
>>>>> -js-compiler-option="--compilation_level WHITESPACE_ONLY"
>>>>>
>>>>> to the cross-compile and I think you'll still get a single file
>>>>>without
>>>>> goog.require but it will be debuggable.
>>>>>
>>>>> These options are handled by the compiler code in a Publisher.
>>>>> MXMLFlexJSPublisher has this default behavior. You can subclass it
>>>>>and
>>>>> create a different js-output-type get it to spit a single-file to the
>>>>> js-debug and a minified single-file to js-release. It will take a
>>>>>long
>>>>> time, though, as gathering in a single file is done by the Google
>>>>>Closure
>>>>> Compiler. But you don't to know much about compilers to make a
>>>>>custom
>>>>> Publisher. Everything is compiled at that point and you are
>>>>>basically
>>>>> dealing with files and configs for GCC.
>>>>>
>>>>> A harder task is to make the goog.require replaceable with some other
>>>>> pattern.
>>>>>
>>>>> -Alex
>>>>>
>>>>
>>>
>>
>
Re: [FalconJX] compiling non-browser javascript
Posted by Alex Harui <ah...@adobe.com>.
We've recently made changes to allow keywords in more places. I
personally don't enjoy doing that work, but maybe Josh will.
-Alex
On 5/1/16, 12:34 PM, "Harbs" <ha...@gmail.com> wrote:
>I get the same error for other reserved words:
>in,
>include,
>with,
>etc…
>
>On May 1, 2016, at 10:28 PM, Harbs <ha...@gmail.com> wrote:
>
>> I manually deleted most of the core classes to get it to compile.
>>
>> I’m now getting an error which I don’t know if it’s valid or a bug in
>>Falcon:
>>
>> public function findKeyStrings(for:String):String{return null;}
>> public function translateKeyString(for:String):String{return null;}
>>
>> When trying to compile a class which contains code like this, I get:
>> ERROR
>>/Users/harbs/Desktop/InDesign10.2/src/com/adobe/indesign/Application.as[6
>>6:33]:
>> 'for' is not allowed here
>>
>> Is “for” really not allowed as the name of a parameter, or i it a bug
>>that the compiler thinks it’s a for loop?
>>
>> Harbs
>>
>> On May 1, 2016, at 3:50 PM, Harbs <ha...@gmail.com> wrote:
>>
>>> Here’s my stab at producing ActionScript files from the OMV files:
>>>https://github.com/unhurdle/omv2as
>>>
>>> The output is actually pretty good. I get error-free output on
>>>InDesign files with the exception of File types because I don’t yet
>>>have the core types linked. Photoshop output is not as good, for the
>>>most part because the OMV files types are not all true types.
>>>
>>> I do have a question though (before I even got to the point where I’m
>>>trying to use this to cross-compile code): When I run the base classes
>>>through the app, I get a bunch of classes which do not compile into a
>>>SWC very well. At least part of the problem is due to the fact that
>>>they confluct with core classes, and I’m not sure how to best handle
>>>this. Here’s a link of the as code:
>>>https://www.dropbox.com/s/pziyrqj7k1ob9p7/ExtendScript.zip?dl=0
>>>
>>> I’m not sure how to best handle this. If anyone has good ideas, please
>>>let me know.
>>>
>>> On Apr 25, 2016, at 9:28 PM, Harbs <ha...@gmail.com> wrote:
>>>
>>>> I was guessing that the release would probably work. I am concerned
>>>>about debugging though.
>>>>
>>>> I will probably try this suggestion next week and see how far I can
>>>>get without further help. Chances are I’ll be back here before I’m
>>>>successful though… ;-)
>>>>
>>>> Thanks!
>>>> Harbs
>>>>
>>>> On Apr 25, 2016, at 6:27 PM, Alex Harui <ah...@adobe.com> wrote:
>>>>
>>>>>
>>>>>
>>>>> On 4/25/16, 8:16 AM, "Josh Tynjala" <jo...@gmail.com> wrote:
>>>>>
>>>>>> In the bin/js-release directory, all of the generated JavaScript is
>>>>>> concatenated into a single file, so it no longer uses
>>>>>>goog.require(). That
>>>>>> should work in environments that cannot load multiple scripts.
>>>>>
>>>>> I was about to suggest that as well. By default, the single-file
>>>>>output
>>>>> is minified so is hard to debug. You can add
>>>>> -js-compiler-option="--compilation_level WHITESPACE_ONLY"
>>>>>
>>>>> to the cross-compile and I think you'll still get a single file
>>>>>without
>>>>> goog.require but it will be debuggable.
>>>>>
>>>>> These options are handled by the compiler code in a Publisher.
>>>>> MXMLFlexJSPublisher has this default behavior. You can subclass it
>>>>>and
>>>>> create a different js-output-type get it to spit a single-file to the
>>>>> js-debug and a minified single-file to js-release. It will take a
>>>>>long
>>>>> time, though, as gathering in a single file is done by the Google
>>>>>Closure
>>>>> Compiler. But you don't to know much about compilers to make a
>>>>>custom
>>>>> Publisher. Everything is compiled at that point and you are
>>>>>basically
>>>>> dealing with files and configs for GCC.
>>>>>
>>>>> A harder task is to make the goog.require replaceable with some other
>>>>> pattern.
>>>>>
>>>>> -Alex
>>>>>
>>>>
>>>
>>
>
Re: [FalconJX] compiling non-browser javascript
Posted by Harbs <ha...@gmail.com>.
I get the same error for other reserved words:
in,
include,
with,
etc…
On May 1, 2016, at 10:28 PM, Harbs <ha...@gmail.com> wrote:
> I manually deleted most of the core classes to get it to compile.
>
> I’m now getting an error which I don’t know if it’s valid or a bug in Falcon:
>
> public function findKeyStrings(for:String):String{return null;}
> public function translateKeyString(for:String):String{return null;}
>
> When trying to compile a class which contains code like this, I get:
> ERROR /Users/harbs/Desktop/InDesign10.2/src/com/adobe/indesign/Application.as[66:33]:
> 'for' is not allowed here
>
> Is “for” really not allowed as the name of a parameter, or i it a bug that the compiler thinks it’s a for loop?
>
> Harbs
>
> On May 1, 2016, at 3:50 PM, Harbs <ha...@gmail.com> wrote:
>
>> Here’s my stab at producing ActionScript files from the OMV files: https://github.com/unhurdle/omv2as
>>
>> The output is actually pretty good. I get error-free output on InDesign files with the exception of File types because I don’t yet have the core types linked. Photoshop output is not as good, for the most part because the OMV files types are not all true types.
>>
>> I do have a question though (before I even got to the point where I’m trying to use this to cross-compile code): When I run the base classes through the app, I get a bunch of classes which do not compile into a SWC very well. At least part of the problem is due to the fact that they confluct with core classes, and I’m not sure how to best handle this. Here’s a link of the as code: https://www.dropbox.com/s/pziyrqj7k1ob9p7/ExtendScript.zip?dl=0
>>
>> I’m not sure how to best handle this. If anyone has good ideas, please let me know.
>>
>> On Apr 25, 2016, at 9:28 PM, Harbs <ha...@gmail.com> wrote:
>>
>>> I was guessing that the release would probably work. I am concerned about debugging though.
>>>
>>> I will probably try this suggestion next week and see how far I can get without further help. Chances are I’ll be back here before I’m successful though… ;-)
>>>
>>> Thanks!
>>> Harbs
>>>
>>> On Apr 25, 2016, at 6:27 PM, Alex Harui <ah...@adobe.com> wrote:
>>>
>>>>
>>>>
>>>> On 4/25/16, 8:16 AM, "Josh Tynjala" <jo...@gmail.com> wrote:
>>>>
>>>>> In the bin/js-release directory, all of the generated JavaScript is
>>>>> concatenated into a single file, so it no longer uses goog.require(). That
>>>>> should work in environments that cannot load multiple scripts.
>>>>
>>>> I was about to suggest that as well. By default, the single-file output
>>>> is minified so is hard to debug. You can add
>>>> -js-compiler-option="--compilation_level WHITESPACE_ONLY"
>>>>
>>>> to the cross-compile and I think you'll still get a single file without
>>>> goog.require but it will be debuggable.
>>>>
>>>> These options are handled by the compiler code in a Publisher.
>>>> MXMLFlexJSPublisher has this default behavior. You can subclass it and
>>>> create a different js-output-type get it to spit a single-file to the
>>>> js-debug and a minified single-file to js-release. It will take a long
>>>> time, though, as gathering in a single file is done by the Google Closure
>>>> Compiler. But you don't to know much about compilers to make a custom
>>>> Publisher. Everything is compiled at that point and you are basically
>>>> dealing with files and configs for GCC.
>>>>
>>>> A harder task is to make the goog.require replaceable with some other
>>>> pattern.
>>>>
>>>> -Alex
>>>>
>>>
>>
>
Re: [FalconJX] compiling non-browser javascript
Posted by Harbs <ha...@gmail.com>.
I manually deleted most of the core classes to get it to compile.
I’m now getting an error which I don’t know if it’s valid or a bug in Falcon:
public function findKeyStrings(for:String):String{return null;}
public function translateKeyString(for:String):String{return null;}
When trying to compile a class which contains code like this, I get:
ERROR /Users/harbs/Desktop/InDesign10.2/src/com/adobe/indesign/Application.as[66:33]:
'for' is not allowed here
Is “for” really not allowed as the name of a parameter, or i it a bug that the compiler thinks it’s a for loop?
Harbs
On May 1, 2016, at 3:50 PM, Harbs <ha...@gmail.com> wrote:
> Here’s my stab at producing ActionScript files from the OMV files: https://github.com/unhurdle/omv2as
>
> The output is actually pretty good. I get error-free output on InDesign files with the exception of File types because I don’t yet have the core types linked. Photoshop output is not as good, for the most part because the OMV files types are not all true types.
>
> I do have a question though (before I even got to the point where I’m trying to use this to cross-compile code): When I run the base classes through the app, I get a bunch of classes which do not compile into a SWC very well. At least part of the problem is due to the fact that they confluct with core classes, and I’m not sure how to best handle this. Here’s a link of the as code: https://www.dropbox.com/s/pziyrqj7k1ob9p7/ExtendScript.zip?dl=0
>
> I’m not sure how to best handle this. If anyone has good ideas, please let me know.
>
> On Apr 25, 2016, at 9:28 PM, Harbs <ha...@gmail.com> wrote:
>
>> I was guessing that the release would probably work. I am concerned about debugging though.
>>
>> I will probably try this suggestion next week and see how far I can get without further help. Chances are I’ll be back here before I’m successful though… ;-)
>>
>> Thanks!
>> Harbs
>>
>> On Apr 25, 2016, at 6:27 PM, Alex Harui <ah...@adobe.com> wrote:
>>
>>>
>>>
>>> On 4/25/16, 8:16 AM, "Josh Tynjala" <jo...@gmail.com> wrote:
>>>
>>>> In the bin/js-release directory, all of the generated JavaScript is
>>>> concatenated into a single file, so it no longer uses goog.require(). That
>>>> should work in environments that cannot load multiple scripts.
>>>
>>> I was about to suggest that as well. By default, the single-file output
>>> is minified so is hard to debug. You can add
>>> -js-compiler-option="--compilation_level WHITESPACE_ONLY"
>>>
>>> to the cross-compile and I think you'll still get a single file without
>>> goog.require but it will be debuggable.
>>>
>>> These options are handled by the compiler code in a Publisher.
>>> MXMLFlexJSPublisher has this default behavior. You can subclass it and
>>> create a different js-output-type get it to spit a single-file to the
>>> js-debug and a minified single-file to js-release. It will take a long
>>> time, though, as gathering in a single file is done by the Google Closure
>>> Compiler. But you don't to know much about compilers to make a custom
>>> Publisher. Everything is compiled at that point and you are basically
>>> dealing with files and configs for GCC.
>>>
>>> A harder task is to make the goog.require replaceable with some other
>>> pattern.
>>>
>>> -Alex
>>>
>>
>