Sorry, that should be:

asjsc source/Example.as
On Oct 22, 2015 5:41 PM, "Josh Tynjala" <joshtynj...@gmail.com> wrote:

> I'm able to get -js-compiler-option to work. It's not ideal, but I can add
> it to my examples.
>
> I can't get ADVANCED_OPTIMIZATIONS to work with asjsc at all, though. Not
> even the simplest example that I can think of. No externs required.
>
> source/Example.as:
>
> package
> {
>     public class Example
>     {
>         public static function run():void
>         {
>             trace("hello world");
>         }
>     }
> }
>
> index.html:
>
> <!DOCTYPE html>
> <html>
> <body>
> <script src="bin/js-release/Example.js"></script>
> <script>
> Example.run();
> </script>
> </body>
> </html>
>
> Build command line:
> asjsc Example.as
>
> Any idea why?
>
> - Josh
>
> On Wed, Oct 21, 2015 at 12:26 PM, Alex Harui <aha...@adobe.com> wrote:
>
>>
>>
>> On 10/21/15, 11:27 AM, "Josh Tynjala" <joshtynj...@gmail.com> wrote:
>> >
>> >By the way, the SIMPLE_OPTIMIZATIONS argument for the closure compiler
>> >might be an acceptable middle ground for minification. When I tried it, I
>> >was able to load up my CreateJS demo and it actually rendered everything.
>> >However, I noticed that it wasn't responding correctly to touch events,
>> so
>> >that's why I fell back to WHITESPACE_ONLY. At the time, I couldn't
>> >investigate further, but maybe now I can figure out what's going on.
>> >SIMPLE_OPTIMIZATIONS should minify a lot more than WHITESPACE_ONLY, and
>> it
>> >seems to work without externs.
>>
>> FWIW, SIMPLE didn’t seem to make a difference vs WHITESPACE_ONLY on the
>> one example I tried.  ADVANCED makes a big difference, dropping the
>> example from 568K to 141K.
>>
>> I’m just pushed the -js-compiler-option changes and set the default back
>> to ADVANCED.  Or will it be a problem to have to add this option to your
>> examples?
>>
>> It should just be:
>>         -js-compiler-option=“—compilation_level WHITESPACE_ONLY”
>>
>> -Alex
>>
>> >
>> >Since Harbs mentioned Angular being minfied so well, it looks like they
>> >use
>> >closure compiler with SIMPLE_OPTIMIZATIONS for most of their codebase
>> (one
>> >particular file uses ADVANCED_OPTIMIZATIONS, for some reason):
>> >
>> >
>> https://github.com/angular/angular.js/blob/d077966ff1ac18262f4615ff1a533db
>> >24d4432a7/lib/grunt/utils.js#L188
>> >
>> >- Josh
>> >
>> >On Wed, Oct 21, 2015 at 10:45 AM, Alex Harui <aha...@adobe.com> wrote:
>> >
>> >>
>> >>
>> >> On 10/21/15, 10:16 AM, "Josh Tynjala" <joshtynj...@gmail.com> wrote:
>> >>
>> >> >The cross-compiler that generates ActionScript from TypeScript
>> >> >definitions?
>> >> >Yeah, I could probably do that. Generating both ActionScript and
>> >>externs
>> >> >files adds some complexity that I'd prefer to hide from ActionScript
>> >> >developers, though. Ideally, most developers wouldn't need to know
>> >>about
>> >> >the externs files.
>> >>
>> >> Actually, I was thinking that you could take the generated .as files
>> and
>> >> run them through FalconJX and package the JS as externs files.
>> >>
>> >> Today, any of the FlexJS SWCs like Core.swc have a build script that
>> >>runs
>> >> a couple of passes to cross-compile the AS to JS, then a final pass
>> that
>> >> compiles the AS into a SWC and packages the generated JS.  It looks
>> from
>> >> the code, that if you put the JS in an externs folder inside the SWC
>> and
>> >> folks reference these SWCs on the external-library-path, that the right
>> >> thing should happen.
>> >>
>> >> >
>> >> >I guess I could redesign dts2as to look for FlexJS and ask it to
>> >> >automatically build the final SWC file with both the generated
>> >> >ActionScript
>> >> >and the generated externs files. That would simplify my tutorials
>> >>either
>> >> >way, since developers won't need to run compc manually.
>> >> >
>> >> >What's the usual environment variable a developer might add for
>> >>FlexJS? Is
>> >> >it FLEXJS_HOME?
>> >>
>> >> FlexJS tries to look just like a regular Flex SDK, so folks should be
>> >>able
>> >> to use FLEX_HOME.
>> >>
>> >> -Alex
>> >>
>> >>
>>
>>
>

Reply via email to