Re: Ant Build Problem

[Erik de Bruin]

I did several fresh pulls of the repos yesterday while testing as2wasm and
I can faithfully reproduce the issue(s). I can also reliably make it go
away if I do these things:

- comment out the ant task in the 'CoreTest' target in frameworks/build.xml
- remove the 'test' dependency from the 'main' target in
frameworks/projects/Basic/build.xml

HTH,

EdB



On Thu, Nov 16, 2017 at 7:43 AM, Yishay Weiss 
wrote:

> Thanks for the pointers Alex. Funnily I stopped getting it after playing
> around with some scripts, cleaning my changes and checking out from git the
> unit tests aren’t being run anymore and so I’m not getting it. I’ll come
> back to it when it starts itching again, but for now I can work.
>
>
>
> 
> From: Alex Harui 
> Sent: Wednesday, November 15, 2017 6:41:20 PM
> To: dev@royale.apache.org
> Subject: Re: Ant Build Problem
>
> My working copies are all torn up with the refactoring and I'd prefer not
> to have to stop and build up another working copy to debug this, assuming
> I can even reproduce it.
>
> The factors are:
> -The compiler Ant tasks use ROYALE_COMPILER_HOME to find the jsc.jar to run
> -Looks like there are bugs in the scripts where ROYALE_COMPILER_HOME is
> not set by each script but maybe only the main script
> -The FlexUnit compile should be using frameworks/royale-config.xml
> -The royale-config.xml should have:
>   COMPILE::SWFAUTO
>   COMPILE::JSAUTO
> -Recent (0.8.0 and later) compilers should convert AUTO to a valid value
> for the compile.
>
>
> HTH,
> -Alex
>
> On 11/15/17, 7:48 AM, "Erik de Bruin"  wrote:
>
> >I see the same issue! I am trying to write a quick tutorial for the new
> >wasm code, but I still can't successfully run "ant super-clean all" on
> >royale-asjs... Please help?
> >
> >compile:
> > [echo] Compiling FlexUnitRoyaleApplication.swf
> > [echo] ROYALE_HOME: /Users/erik/Desktop/tmp/royale-asjs
> > [echo] ROYALE_SWF_COMPILER_HOME: /Users/erik/Desktop/tmp/royale-asjs
> > [echo] FLEXUNIT_HOME: /Users/erik/Desktop/tmp/
> royale-asjs/in/flexunit
> > [echo] playerglobal.version: 11.1
> >[mxmlc] MXMLJSC
> >[mxmlc] -debug
> >[mxmlc] -compiler.targets=SWF
> >[mxmlc] +playerglobal.version=11.1
> >[mxmlc]
> >+env.PLAYERGLOBAL_HOME=/Users/erik/Documents/Royale/PlayerGlobal/player
> >[mxmlc]
> >-swf-library-path+=/Users/erik/Desktop/tmp/royale-asjs/frameworks/libs
> >[mxmlc]
> >-swf-library-path+=/Users/erik/Desktop/tmp/royale-asjs/
> in/flexunit/flexuni
> >t
> >[mxmlc] -define=CONFIG::dummy,false
> >[mxmlc] +royalelib=/Users/erik/Desktop/tmp/royale-asjs/frameworks/
> >[mxmlc]
> >-output=/Users/erik/Desktop/tmp/royale-asjs/frameworks/
> projects/Core/src/t
> >est/royale/FlexUnitRoyaleApplication.swf
> >[mxmlc] --
> >[mxmlc]
> >/Users/erik/Desktop/tmp/royale-asjs/frameworks/
> projects/Core/src/test/roya
> >le/FlexUnitRoyaleApplication.mxml
> >[mxmlc] Loading configuration:
> >/Users/erik/Desktop/tmp/royale-asjs/frameworks/royale-config.xml
> >[mxmlc]
> >[mxmlc] scanning for overrides: IFlexInfo
> >
> >
> >
> >[mxmlc] scanning for overrides: StringTrimmer
> >[mxmlc] 2.498695566 seconds
> >[mxmlc]
> >/Users/erik/Desktop/tmp/royale-asjs/frameworks/
> projects/Core/src/test/roya
> >le/flexUnitTests/KeyboardEventConverterTest.as(23):
> >col: 5 Error: Can not resolve config constant: 'SWF'
> >[mxmlc]
> >[mxmlc] COMPILE::SWF
> >[mxmlc] ^
> >[mxmlc]
> >[mxmlc]
> >/Users/erik/Desktop/tmp/royale-asjs/frameworks/
> projects/Core/src/test/roya
> >le/flexUnitTests/KeyboardEventConverterTest.as(56):
> >col: 13 Error: Can not resolve config constant: 'SWF'
> >[mxmlc]
> >[mxmlc] COMPILE::SWF{
> >[mxmlc] ^
> >[mxmlc]
> >[mxmlc]
> >/Users/erik/Desktop/tmp/royale-asjs/frameworks/
> projects/Core/src/test/roya
> >le/flexUnitTests/KeyboardEventConverterTest.as(68):
> >col: 13 Error: Can not resolve config constant: 'JS'
> >[mxmlc]
> >[mxmlc] COMPILE::JS{
> >[mxmlc] ^
> >[mxmlc]
> >[mxmlc]
> >/Users/erik/Desktop/tmp/royale-asjs/in/flexunit/
> flexunit/flexunit-4.2.0-20
> >140410-flex_4.12.0.swc
> >Warning: The definition mx.rpc.IResponder depended on by
> >org.flexunit.async.IAsyncTestResponder in the SWC
> >/Users/erik/Desktop/tmp/royale-asjs/in/flexunit/
> flexunit/flexunit-4.2.0-20
> >140410-flex_4.12.0.swc
> >could not be found
> >
> >...
> >
> >[mxmlc]
> >
> >test:
> >[mkdir] Created dir:
> >/Users/erik/Desktop/tmp/royale-asjs/frameworks/
> projects/Core/src/test/roya
> >le/out
> > [flexunit] Validating task attributes ...
> >
> >BUILD FAILED
> >/Users/erik/Desktop/tmp/royale-asjs/build.xml:569: The following error
> >occurred while executing this line:
> >/Users/erik/Desktop/tmp/royale-asjs/frameworks/build.xml:118: The
> >following
> >error occurred 

Re: FAQ page on website

[Piotr Zarzycki]

Carlos,

Here is my feedback about faq:

"Can I have write access to GIT?":
It is talking about providing patches by diff etc. I think we should talk
there about pull requesting to get used to people to such behavior.


"I've used Apache Flex before. How do I start using Apache Royale?"
It would be great to have it "Visual Studio Code" clickable and make it as
a link to Josh's website
I would like to ask you also add there "Moonshine IDE" and link to
http://moonshine-ide.com/

Good Job! Thanks! :)

2017-11-15 19:56 GMT+01:00 Piotr Zarzycki :

> Hi Carlos,
>
> One question which come up to each of so far created website - Does they
> need to be checked by someone who is native english speaker ?
>
> Thanks, Piotr
>
>
> 2017-11-15 19:52 GMT+01:00 Carlos Rovira :
>
>> Hi,
>>
>> I put on place a first draft of the FAQ page. You can review it here:
>>
>> http://royale.codeoscopic.com/faq/
>>
>> As always, please check it and read to discover :
>>
>> * if the content is what you expect
>> * If sections are ok, and if you want to add more sections.
>> * for the content you consider right if is well written (think that I'm
>> spanish and I try to do my best, but some phrases or words could have
>> typos
>> and other irregularities )
>> * If you think some content should be added, please, post it here, so I
>> can
>> copy and paste to the page, or request for a user in the wordpress
>> instance
>> so you can enter it yourself.
>>
>> Thanks for your help!
>>
>> --
>> Carlos Rovira
>> http://about.me/carlosrovira
>>
>
>
>
> --
>
> Piotr Zarzycki
>
> Patreon: *https://www.patreon.com/piotrzarzycki
> *
>



-- 

Piotr Zarzycki

Patreon: *https://www.patreon.com/piotrzarzycki
*

RE: Ant Build Problem

[Yishay Weiss]

Thanks for the pointers Alex. Funnily I stopped getting it after playing around 
with some scripts, cleaning my changes and checking out from git the unit tests 
aren’t being run anymore and so I’m not getting it. I’ll come back to it when 
it starts itching again, but for now I can work.




From: Alex Harui 
Sent: Wednesday, November 15, 2017 6:41:20 PM
To: dev@royale.apache.org
Subject: Re: Ant Build Problem

My working copies are all torn up with the refactoring and I'd prefer not
to have to stop and build up another working copy to debug this, assuming
I can even reproduce it.

The factors are:
-The compiler Ant tasks use ROYALE_COMPILER_HOME to find the jsc.jar to run
-Looks like there are bugs in the scripts where ROYALE_COMPILER_HOME is
not set by each script but maybe only the main script
-The FlexUnit compile should be using frameworks/royale-config.xml
-The royale-config.xml should have:
  COMPILE::SWFAUTO
  COMPILE::JSAUTO
-Recent (0.8.0 and later) compilers should convert AUTO to a valid value
for the compile.


HTH,
-Alex

On 11/15/17, 7:48 AM, "Erik de Bruin"  wrote:

>I see the same issue! I am trying to write a quick tutorial for the new
>wasm code, but I still can't successfully run "ant super-clean all" on
>royale-asjs... Please help?
>
>compile:
> [echo] Compiling FlexUnitRoyaleApplication.swf
> [echo] ROYALE_HOME: /Users/erik/Desktop/tmp/royale-asjs
> [echo] ROYALE_SWF_COMPILER_HOME: /Users/erik/Desktop/tmp/royale-asjs
> [echo] FLEXUNIT_HOME: /Users/erik/Desktop/tmp/royale-asjs/in/flexunit
> [echo] playerglobal.version: 11.1
>[mxmlc] MXMLJSC
>[mxmlc] -debug
>[mxmlc] -compiler.targets=SWF
>[mxmlc] +playerglobal.version=11.1
>[mxmlc]
>+env.PLAYERGLOBAL_HOME=/Users/erik/Documents/Royale/PlayerGlobal/player
>[mxmlc]
>-swf-library-path+=/Users/erik/Desktop/tmp/royale-asjs/frameworks/libs
>[mxmlc]
>-swf-library-path+=/Users/erik/Desktop/tmp/royale-asjs/in/flexunit/flexuni
>t
>[mxmlc] -define=CONFIG::dummy,false
>[mxmlc] +royalelib=/Users/erik/Desktop/tmp/royale-asjs/frameworks/
>[mxmlc]
>-output=/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/t
>est/royale/FlexUnitRoyaleApplication.swf
>[mxmlc] --
>[mxmlc]
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/test/roya
>le/FlexUnitRoyaleApplication.mxml
>[mxmlc] Loading configuration:
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/royale-config.xml
>[mxmlc]
>[mxmlc] scanning for overrides: IFlexInfo
>
>
>
>[mxmlc] scanning for overrides: StringTrimmer
>[mxmlc] 2.498695566 seconds
>[mxmlc]
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/test/roya
>le/flexUnitTests/KeyboardEventConverterTest.as(23):
>col: 5 Error: Can not resolve config constant: 'SWF'
>[mxmlc]
>[mxmlc] COMPILE::SWF
>[mxmlc] ^
>[mxmlc]
>[mxmlc]
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/test/roya
>le/flexUnitTests/KeyboardEventConverterTest.as(56):
>col: 13 Error: Can not resolve config constant: 'SWF'
>[mxmlc]
>[mxmlc] COMPILE::SWF{
>[mxmlc] ^
>[mxmlc]
>[mxmlc]
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/test/roya
>le/flexUnitTests/KeyboardEventConverterTest.as(68):
>col: 13 Error: Can not resolve config constant: 'JS'
>[mxmlc]
>[mxmlc] COMPILE::JS{
>[mxmlc] ^
>[mxmlc]
>[mxmlc]
>/Users/erik/Desktop/tmp/royale-asjs/in/flexunit/flexunit/flexunit-4.2.0-20
>140410-flex_4.12.0.swc
>Warning: The definition mx.rpc.IResponder depended on by
>org.flexunit.async.IAsyncTestResponder in the SWC
>/Users/erik/Desktop/tmp/royale-asjs/in/flexunit/flexunit/flexunit-4.2.0-20
>140410-flex_4.12.0.swc
>could not be found
>
>...
>
>[mxmlc]
>
>test:
>[mkdir] Created dir:
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/test/roya
>le/out
> [flexunit] Validating task attributes ...
>
>BUILD FAILED
>/Users/erik/Desktop/tmp/royale-asjs/build.xml:569: The following error
>occurred while executing this line:
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/build.xml:118: The
>following
>error occurred while executing this line:
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/build.xml:233: The
>following
>error occurred while executing this line:
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/build.xml:61:
>The following error occurred while executing this line:
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/test/roya
>le/build.xml:141:
>The provided 'swf' property value [] could not be found or is not a valid
>remote URL.
>
>On Wed, Nov 15, 2017 at 2:53 PM, Yishay Weiss 
>wrote:
>
>> I changed ROYALE_SWF_COMPILER_HOME and FLEXUNIT_HOME to match yours but
>> the result is the same.
>>
>> FWIW this is ROYALE_COMPILER_HOME which 

Re: [royale-asjs] branch develop updated: Add ‘glue’ utility script to allow WASM to be loaded in HTML

[Erik de Bruin]

Carlos,

I'm seeing lot of attention this days for WebAssembly due to Safari and
> Edge supporting it.
> The expectations are high on that front and I'm reading that WASM could be
> almost ready for Spring in 2018.
> So I think this is a must for us in addition to HTML/JS/CSS effort. I'm
> trying to see the picture and even seems that both
> HTML/JS/CSS and WASM could be used at the same time right?, so in that
> case, we should think how this could be handled in Royale.
>

To help me grok the challenge of AS -> WebAssembly, I have stripped down
the transpiler to the bare minimum. This means that any tooling or
framework integration is basically excluded. The first priority must be to
get 'vanilla' AS to produce valid WAST code. At best that'll be months of
work, but seeing as I still have to learn WASM in depth, I'm sure there are
going to be some bumps in the road ;-)

I think that once there is a fair level of fidelity in the transpilation,
we can start to think about building/integrating UI frameworks with this
transpiler. I have some ideas that got snowed under when we started work on
asjs I'd like to try out. Think GCC compatible ES6 Material Design, with a
glue libraries for communication between WASM and JavaScript code.
Something like that.

Thanks,

EdB



-- 
Ix Multimedia Software

Jan Luykenstraat 27
3521 VB Utrecht

T. 06-51952295
I. www.ixsoftware.nl

[AS2WASM] Proof of concept code available

[Erik de Bruin]

Hi,

As the subject suggests, I can't make up my mind about a working title for
the WebAssembly effort.

But there is also some good news: I have now assembled a usable toolchain
to create and publish 'as2wasm' projects. As with any proof of concept,
there are serious caveats with regard to functionality... Mostly, there is
no actual transpiling yet ;-) But, to be fair, a week is not a lot of time
to get that 'detail' done ;-)

So, what is there to see? I have created a new, completely separate
transpiler client (WASTC), and all the code needed to build, package and
distribute it. It has currently one output for whatever you put in: a
functional wasm module that exports a simple 'add' function. The client has
a publisher that creates a viable project as output, compiling the fake
transpiled file into a valid binary file, together with a 'glue' file to
allow the .wasm file to load in the companion HTML file.

The core of the project is a couple of NPM files. The first one is part of
the framework and allows you to not only build the new transpiler client,
but also to create a new project skeleton. The second one is added to each
project skeleton and allows you to build and preview your project.

If you're interested, I have created a wiki page with setup
instructions[1]. There are a few nasty dependencies, please be patient and
keep in mind that WebAssembly is a very young technology and the tools are
few and far between.

So: with a bit of patience you can build and "use" as2wasm as if it were
fully functional. As long as you don't expect it to handle anything but the
project skeleton and that there is not yet real transpilation, just a
simple placeholder output to make everything work.

If you give it a spin, keep me posted and please don't hesitate to be blunt
and give me some (constructive) criticism.

Thanks and have fun!

EdB

1: https://github.com/apache/royale-asjs/wiki/User-Manual-AS2WASM



-- 
Ix Multimedia Software

Jan Luykenstraat 27
3521 VB Utrecht

T. 06-51952295
I. www.ixsoftware.nl

Re: Publishing royale to npm

[OmPrakash Muppirala]

At some point, we could do things like create a pom.xml based on the
features they want.  For example, the Spring Boot project has this page:
https://start.spring.io/
where you could go select the functionalities you want and download the
pom.xml.  Then you simply run mvn install, and everything is setup for you
automatically.

With npm, we already have Josh's yeoman project which can be further
tweaked to make it an interactive functionality selection process, which
would in turn generate the appropriate workspace files.

We could also have a github repo with various starter packs, that support
various functionalities and IDEs.

Stuff volounteers can start working on today :-)

Thanks,
Om

On Wed, Nov 15, 2017 at 10:37 AM, Alex Harui 
wrote:

> Hi Carlos,
>
> IMO, it is a classic trade-off problem.  It makes sense that many people
> want to know exactly what libraries (and versions of those libraries) are
> being used in the recipe for their app.  And, similarly, some people want
> to know exactly what code is being used in their app and will start with
> the Basic set and add beads.
>
> Flex became popular because it improves developer productivity, and
> hopefully Royale will too, and just like many people have expressed a
> desire for the Express set, others just want to create a new project, type
> a few lines of MXML and AS and have it build.
>
> In the IDE packages, all SWCs are available without any configuring.  I
> think that's actually true in the Maven archetypes as well.
>
> Explicitness and granularity generally take more time and thus are often
> traded off against getting something up and running quickly.  I believe
> that with Royale, we are giving folks more choices in more places.  Our
> "getting started" content will probably use Express components and IDE
> packaging or an archetype with all or most SWCs in the POM.  But folks
> will be able to use Basic and explicit SWC dependencies in their POMs if
> they want to.
>
> My 2 cents,
> -Alex
>
> On 11/15/17, 9:53 AM, "carlos.rov...@gmail.com on behalf of Carlos Rovira"
>  wrote:
>
> >Hi,
> >
> >I must recognize that I'm not fully understand what's the problem.
> >
> >If you use a Royale class to load data, you know it should link
> >Network.swc, To get html controls you use HTML.swc.
> >you want binding? so you refer to Binding.swc.
> >
> >For me this is completely natural.
> >
> >People working with an API must learn the that API, the classes implied,
> >and how is structured.
> >
> >I feel that not doing this could bring to a lots of unused code and
> >libraries and maybe will not affect the final weight or load times, but
> >for
> >sure the size of the project and the management will be affected since it
> >could carry things that are never needed.
> >
> >Again, most of this things is not critical, and can be part of the way
> >each
> >one see development, but want to explain this so people could fully
> >understand the goods of maven.
> >
> >
> >
> >
> >
> >2017-11-15 17:27 GMT+01:00 Alex Harui :
> >
> >> Harbs,
> >>
> >> While technically, it is true that for Royale, the pom.xml files need to
> >> have the right SWCs specified, I think if we maintain the archetypes
> >>then
> >> when you create a new Maven project for an app you will get every SWC
> >> specified in your POM, and I don't think there is an issue for having
> >>more
> >> SWCs than you actually need in your POM.
> >>
> >> Getting the right SWCs in the POM is more of an issue for the framework
> >> build since we do care about what SWCs depend on other SWCs and the
> >>order
> >> we build them.
> >>
> >> What I still dream about is ways of automating these things.  So that if
> >> someone adds a new SWC in frameworks/projects, they don't also have to
> >> remember to update the archetypes.  And similarly, the duplication of
> >> settings in the POMs for the SWCs and the -config.xml files for the
> >>SWCs.
> >>
> >> Of course, I could be wrong...
> >> -Alex
> >>
> >> On 11/15/17, 3:16 AM, "Harbs"  wrote:
> >>
> >> >> why too error-prone?
> >> >
> >> >Because it requires users to know which classes are in which swc. I
> >>don’t
> >> >see a reason for that. But, I’m not a Maven user, so take my thoughts
> >> >with a grain of salt.
> >> >
> >> >Harbs
> >> >
> >> >> On Nov 15, 2017, at 12:54 PM, Carlos Rovira  >
> >> >>wrote:
> >> >>
> >> >> Hi Harbs,
> >> >>
> >> >> why too error-prone? If you are developing a Royale app with Maven
> >>you
> >> >> create a pom, and you add libraries on demand. So your build is
> >>totally
> >> >> automated and you only need your project pom in the future and not
> >> >>manually
> >> >> check if libraries are the required ones or not. I see just the
> >>opposite
> >> >> and less error prone.
> >> >>
> >> >> I know this is lastly a question on how each looks to ways to do
> 

Re: FAQ page on website

[Piotr Zarzycki]

Hi Carlos,

One question which come up to each of so far created website - Does they
need to be checked by someone who is native english speaker ?

Thanks, Piotr


2017-11-15 19:52 GMT+01:00 Carlos Rovira :

> Hi,
>
> I put on place a first draft of the FAQ page. You can review it here:
>
> http://royale.codeoscopic.com/faq/
>
> As always, please check it and read to discover :
>
> * if the content is what you expect
> * If sections are ok, and if you want to add more sections.
> * for the content you consider right if is well written (think that I'm
> spanish and I try to do my best, but some phrases or words could have typos
> and other irregularities )
> * If you think some content should be added, please, post it here, so I can
> copy and paste to the page, or request for a user in the wordpress instance
> so you can enter it yourself.
>
> Thanks for your help!
>
> --
> Carlos Rovira
> http://about.me/carlosrovira
>



-- 

Piotr Zarzycki

Patreon: *https://www.patreon.com/piotrzarzycki
*

FAQ page on website

[Carlos Rovira]

Hi,

I put on place a first draft of the FAQ page. You can review it here:

http://royale.codeoscopic.com/faq/

As always, please check it and read to discover :

* if the content is what you expect
* If sections are ok, and if you want to add more sections.
* for the content you consider right if is well written (think that I'm
spanish and I try to do my best, but some phrases or words could have typos
and other irregularities )
* If you think some content should be added, please, post it here, so I can
copy and paste to the page, or request for a user in the wordpress instance
so you can enter it yourself.

Thanks for your help!

-- 
Carlos Rovira
http://about.me/carlosrovira

[Maven|Royale] Maven snapshot artifacts available in the Apache central repository

[Piotr Zarzycki]

Hi Guys,

About one week ago I have discovered that Maven artifacts for Royale
framework are not available in the central Apache repository. Anyone who
tried build Hello World with cleaned ".m2" maven folder failed.

I just fixed that issue and you can now remove everything from .m2 and
build your application without worry that something is missing - Maven
download everything from the Apache central repository. :)

Thanks,
-- 

Piotr Zarzycki

Patreon: *https://www.patreon.com/piotrzarzycki
*

Re: Publishing royale to npm

[Alex Harui]

Hi Carlos,

IMO, it is a classic trade-off problem.  It makes sense that many people
want to know exactly what libraries (and versions of those libraries) are
being used in the recipe for their app.  And, similarly, some people want
to know exactly what code is being used in their app and will start with
the Basic set and add beads.

Flex became popular because it improves developer productivity, and
hopefully Royale will too, and just like many people have expressed a
desire for the Express set, others just want to create a new project, type
a few lines of MXML and AS and have it build.

In the IDE packages, all SWCs are available without any configuring.  I
think that's actually true in the Maven archetypes as well.

Explicitness and granularity generally take more time and thus are often
traded off against getting something up and running quickly.  I believe
that with Royale, we are giving folks more choices in more places.  Our
"getting started" content will probably use Express components and IDE
packaging or an archetype with all or most SWCs in the POM.  But folks
will be able to use Basic and explicit SWC dependencies in their POMs if
they want to.

My 2 cents,
-Alex

On 11/15/17, 9:53 AM, "carlos.rov...@gmail.com on behalf of Carlos Rovira"
 wrote:

>Hi,
>
>I must recognize that I'm not fully understand what's the problem.
>
>If you use a Royale class to load data, you know it should link
>Network.swc, To get html controls you use HTML.swc.
>you want binding? so you refer to Binding.swc.
>
>For me this is completely natural.
>
>People working with an API must learn the that API, the classes implied,
>and how is structured.
>
>I feel that not doing this could bring to a lots of unused code and
>libraries and maybe will not affect the final weight or load times, but
>for
>sure the size of the project and the management will be affected since it
>could carry things that are never needed.
>
>Again, most of this things is not critical, and can be part of the way
>each
>one see development, but want to explain this so people could fully
>understand the goods of maven.
>
>
>
>
>
>2017-11-15 17:27 GMT+01:00 Alex Harui :
>
>> Harbs,
>>
>> While technically, it is true that for Royale, the pom.xml files need to
>> have the right SWCs specified, I think if we maintain the archetypes
>>then
>> when you create a new Maven project for an app you will get every SWC
>> specified in your POM, and I don't think there is an issue for having
>>more
>> SWCs than you actually need in your POM.
>>
>> Getting the right SWCs in the POM is more of an issue for the framework
>> build since we do care about what SWCs depend on other SWCs and the
>>order
>> we build them.
>>
>> What I still dream about is ways of automating these things.  So that if
>> someone adds a new SWC in frameworks/projects, they don't also have to
>> remember to update the archetypes.  And similarly, the duplication of
>> settings in the POMs for the SWCs and the -config.xml files for the
>>SWCs.
>>
>> Of course, I could be wrong...
>> -Alex
>>
>> On 11/15/17, 3:16 AM, "Harbs"  wrote:
>>
>> >> why too error-prone?
>> >
>> >Because it requires users to know which classes are in which swc. I
>>don’t
>> >see a reason for that. But, I’m not a Maven user, so take my thoughts
>> >with a grain of salt.
>> >
>> >Harbs
>> >
>> >> On Nov 15, 2017, at 12:54 PM, Carlos Rovira 
>> >>wrote:
>> >>
>> >> Hi Harbs,
>> >>
>> >> why too error-prone? If you are developing a Royale app with Maven
>>you
>> >> create a pom, and you add libraries on demand. So your build is
>>totally
>> >> automated and you only need your project pom in the future and not
>> >>manually
>> >> check if libraries are the required ones or not. I see just the
>>opposite
>> >> and less error prone.
>> >>
>> >> I know this is lastly a question on how each looks to ways to do
>>things,
>> >> but I use to look at how others are doing things and I think most of
>>the
>> >> projects at Apache are using maven as their way to build their
>>projects
>> >>and
>> >> manage continuous integration, so it must be a safe way to do thing
>>or
>> >> maven will be less used today.
>> >>
>> >> In my experience, I'm using a macbookpro and setup the environment
>>is a
>> >> breeze. Using Hombrew just need to bring to a new system java, git,
>> >>maven
>> >> and from there I get all downloaded and built. maven downloads only
>> >>what is
>> >> needed and I don't need to worry about it. Even don't need to set up
>> >> environment variables, what's so cool :)
>> >>
>> >> just my 2ctnms
>> >>
>> >> C.
>> >>
>> >>
>> >> 2017-11-15 11:04 GMT+01:00 Harbs :
>> >>
>> >>> The original topic was npm, but since we’re talking about Maven… ;-)
>> >>>
>> >>> It seems to me that Royale should have a single Maven artifacts
>> >>>dependency
>> >>> that users should be able to 

Re: Publishing royale to npm

[Carlos Rovira]

Hi,

I must recognize that I'm not fully understand what's the problem.

If you use a Royale class to load data, you know it should link
Network.swc, To get html controls you use HTML.swc.
you want binding? so you refer to Binding.swc.

For me this is completely natural.

People working with an API must learn the that API, the classes implied,
and how is structured.

I feel that not doing this could bring to a lots of unused code and
libraries and maybe will not affect the final weight or load times, but for
sure the size of the project and the management will be affected since it
could carry things that are never needed.

Again, most of this things is not critical, and can be part of the way each
one see development, but want to explain this so people could fully
understand the goods of maven.





2017-11-15 17:27 GMT+01:00 Alex Harui :

> Harbs,
>
> While technically, it is true that for Royale, the pom.xml files need to
> have the right SWCs specified, I think if we maintain the archetypes then
> when you create a new Maven project for an app you will get every SWC
> specified in your POM, and I don't think there is an issue for having more
> SWCs than you actually need in your POM.
>
> Getting the right SWCs in the POM is more of an issue for the framework
> build since we do care about what SWCs depend on other SWCs and the order
> we build them.
>
> What I still dream about is ways of automating these things.  So that if
> someone adds a new SWC in frameworks/projects, they don't also have to
> remember to update the archetypes.  And similarly, the duplication of
> settings in the POMs for the SWCs and the -config.xml files for the SWCs.
>
> Of course, I could be wrong...
> -Alex
>
> On 11/15/17, 3:16 AM, "Harbs"  wrote:
>
> >> why too error-prone?
> >
> >Because it requires users to know which classes are in which swc. I don’t
> >see a reason for that. But, I’m not a Maven user, so take my thoughts
> >with a grain of salt.
> >
> >Harbs
> >
> >> On Nov 15, 2017, at 12:54 PM, Carlos Rovira 
> >>wrote:
> >>
> >> Hi Harbs,
> >>
> >> why too error-prone? If you are developing a Royale app with Maven you
> >> create a pom, and you add libraries on demand. So your build is totally
> >> automated and you only need your project pom in the future and not
> >>manually
> >> check if libraries are the required ones or not. I see just the opposite
> >> and less error prone.
> >>
> >> I know this is lastly a question on how each looks to ways to do things,
> >> but I use to look at how others are doing things and I think most of the
> >> projects at Apache are using maven as their way to build their projects
> >>and
> >> manage continuous integration, so it must be a safe way to do thing or
> >> maven will be less used today.
> >>
> >> In my experience, I'm using a macbookpro and setup the environment is a
> >> breeze. Using Hombrew just need to bring to a new system java, git,
> >>maven
> >> and from there I get all downloaded and built. maven downloads only
> >>what is
> >> needed and I don't need to worry about it. Even don't need to set up
> >> environment variables, what's so cool :)
> >>
> >> just my 2ctnms
> >>
> >> C.
> >>
> >>
> >> 2017-11-15 11:04 GMT+01:00 Harbs :
> >>
> >>> The original topic was npm, but since we’re talking about Maven… ;-)
> >>>
> >>> It seems to me that Royale should have a single Maven artifacts
> >>>dependency
> >>> that users should be able to specify which pulls in all the possibly
> >>>needed
> >>> swcs.
> >>>
> >>> Maybe I’m missing something, but from the maven examples I’ve seen, it
> >>> looks like you need to specify which swcs are needed for them to be
> >>>pulled
> >>> in. That seems too error-prone.
> >>>
>  On Nov 15, 2017, at 11:52 AM, Piotr Zarzycki
> 
> >>> wrote:
> 
>  In case of Maven artifacts the only needs of doing one build is for
>  convenient people who are using Maven build to develop SDK itself. If
> >>> I'm a
>  user whom would like to use Royale and build my own application by
> Maven
> >>> I
>  don't need download repository and build myself whole sources. I just
> >>> need
>  to create simple pom file and all artifacts will be downloaded from
> >>> Apache
>  Maven central - my application will build.
> 
>  Piotr
> 
> 
> 
>  2017-11-15 10:46 GMT+01:00 Harbs :
> 
> > +1
> >
> > I think this is the simplest way to handle it.
> >
> > The only downside is that folks who don’t need the framework will
> >>> download
> > more than they need. but hard-drive space is pretty cheap and like
> >>> Carlos
> > says, we can split if afterwards if there’s demand.
> >
> >> On Nov 15, 2017, at 11:38 AM, Carlos Rovira
> >>
> > wrote:
> >>
> >> Hi,
> >>
> >> I prefer :
> 

Re: Ant Build Problem

[Alex Harui]

My working copies are all torn up with the refactoring and I'd prefer not
to have to stop and build up another working copy to debug this, assuming
I can even reproduce it.

The factors are:
-The compiler Ant tasks use ROYALE_COMPILER_HOME to find the jsc.jar to run
-Looks like there are bugs in the scripts where ROYALE_COMPILER_HOME is
not set by each script but maybe only the main script
-The FlexUnit compile should be using frameworks/royale-config.xml
-The royale-config.xml should have:
  COMPILE::SWFAUTO
  COMPILE::JSAUTO
-Recent (0.8.0 and later) compilers should convert AUTO to a valid value
for the compile.


HTH,
-Alex

On 11/15/17, 7:48 AM, "Erik de Bruin"  wrote:

>I see the same issue! I am trying to write a quick tutorial for the new
>wasm code, but I still can't successfully run "ant super-clean all" on
>royale-asjs... Please help?
>
>compile:
> [echo] Compiling FlexUnitRoyaleApplication.swf
> [echo] ROYALE_HOME: /Users/erik/Desktop/tmp/royale-asjs
> [echo] ROYALE_SWF_COMPILER_HOME: /Users/erik/Desktop/tmp/royale-asjs
> [echo] FLEXUNIT_HOME: /Users/erik/Desktop/tmp/royale-asjs/in/flexunit
> [echo] playerglobal.version: 11.1
>[mxmlc] MXMLJSC
>[mxmlc] -debug
>[mxmlc] -compiler.targets=SWF
>[mxmlc] +playerglobal.version=11.1
>[mxmlc]
>+env.PLAYERGLOBAL_HOME=/Users/erik/Documents/Royale/PlayerGlobal/player
>[mxmlc]
>-swf-library-path+=/Users/erik/Desktop/tmp/royale-asjs/frameworks/libs
>[mxmlc]
>-swf-library-path+=/Users/erik/Desktop/tmp/royale-asjs/in/flexunit/flexuni
>t
>[mxmlc] -define=CONFIG::dummy,false
>[mxmlc] +royalelib=/Users/erik/Desktop/tmp/royale-asjs/frameworks/
>[mxmlc]
>-output=/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/t
>est/royale/FlexUnitRoyaleApplication.swf
>[mxmlc] --
>[mxmlc]
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/test/roya
>le/FlexUnitRoyaleApplication.mxml
>[mxmlc] Loading configuration:
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/royale-config.xml
>[mxmlc]
>[mxmlc] scanning for overrides: IFlexInfo
>
>
>
>[mxmlc] scanning for overrides: StringTrimmer
>[mxmlc] 2.498695566 seconds
>[mxmlc]
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/test/roya
>le/flexUnitTests/KeyboardEventConverterTest.as(23):
>col: 5 Error: Can not resolve config constant: 'SWF'
>[mxmlc]
>[mxmlc] COMPILE::SWF
>[mxmlc] ^
>[mxmlc]
>[mxmlc]
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/test/roya
>le/flexUnitTests/KeyboardEventConverterTest.as(56):
>col: 13 Error: Can not resolve config constant: 'SWF'
>[mxmlc]
>[mxmlc] COMPILE::SWF{
>[mxmlc] ^
>[mxmlc]
>[mxmlc]
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/test/roya
>le/flexUnitTests/KeyboardEventConverterTest.as(68):
>col: 13 Error: Can not resolve config constant: 'JS'
>[mxmlc]
>[mxmlc] COMPILE::JS{
>[mxmlc] ^
>[mxmlc]
>[mxmlc]
>/Users/erik/Desktop/tmp/royale-asjs/in/flexunit/flexunit/flexunit-4.2.0-20
>140410-flex_4.12.0.swc
>Warning: The definition mx.rpc.IResponder depended on by
>org.flexunit.async.IAsyncTestResponder in the SWC
>/Users/erik/Desktop/tmp/royale-asjs/in/flexunit/flexunit/flexunit-4.2.0-20
>140410-flex_4.12.0.swc
>could not be found
>
>...
>
>[mxmlc]
>
>test:
>[mkdir] Created dir:
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/test/roya
>le/out
> [flexunit] Validating task attributes ...
>
>BUILD FAILED
>/Users/erik/Desktop/tmp/royale-asjs/build.xml:569: The following error
>occurred while executing this line:
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/build.xml:118: The
>following
>error occurred while executing this line:
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/build.xml:233: The
>following
>error occurred while executing this line:
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/build.xml:61:
>The following error occurred while executing this line:
>/Users/erik/Desktop/tmp/royale-asjs/frameworks/projects/Core/src/test/roya
>le/build.xml:141:
>The provided 'swf' property value [] could not be found or is not a valid
>remote URL.
>
>On Wed, Nov 15, 2017 at 2:53 PM, Yishay Weiss 
>wrote:
>
>> I changed ROYALE_SWF_COMPILER_HOME and FLEXUNIT_HOME to match yours but
>> the result is the same.
>>
>> FWIW this is ROYALE_COMPILER_HOME which appears to not resolve
>>correctly:
>> C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale>echo
>> %ROYALE_COMPILER_HOME%
>> C:\dev\flexjs\royale-asjs/js
>> From: Harbs
>> Sent: Wednesday, November 15, 2017 12:01 PM
>> To: dev@royale.apache.org
>> Subject: Re: Ant Build Problem
>>
>> It works for me. It looks like your ROYALE_SWF_COMPILER_HOME is pointing
>> to a different location than mine.
>>
>> 

Re: Publishing royale to npm

[Alex Harui]

Harbs,

While technically, it is true that for Royale, the pom.xml files need to
have the right SWCs specified, I think if we maintain the archetypes then
when you create a new Maven project for an app you will get every SWC
specified in your POM, and I don't think there is an issue for having more
SWCs than you actually need in your POM.

Getting the right SWCs in the POM is more of an issue for the framework
build since we do care about what SWCs depend on other SWCs and the order
we build them.

What I still dream about is ways of automating these things.  So that if
someone adds a new SWC in frameworks/projects, they don't also have to
remember to update the archetypes.  And similarly, the duplication of
settings in the POMs for the SWCs and the -config.xml files for the SWCs.

Of course, I could be wrong...
-Alex

On 11/15/17, 3:16 AM, "Harbs"  wrote:

>> why too error-prone?
>
>Because it requires users to know which classes are in which swc. I don’t
>see a reason for that. But, I’m not a Maven user, so take my thoughts
>with a grain of salt.
>
>Harbs
>
>> On Nov 15, 2017, at 12:54 PM, Carlos Rovira 
>>wrote:
>> 
>> Hi Harbs,
>> 
>> why too error-prone? If you are developing a Royale app with Maven you
>> create a pom, and you add libraries on demand. So your build is totally
>> automated and you only need your project pom in the future and not
>>manually
>> check if libraries are the required ones or not. I see just the opposite
>> and less error prone.
>> 
>> I know this is lastly a question on how each looks to ways to do things,
>> but I use to look at how others are doing things and I think most of the
>> projects at Apache are using maven as their way to build their projects
>>and
>> manage continuous integration, so it must be a safe way to do thing or
>> maven will be less used today.
>> 
>> In my experience, I'm using a macbookpro and setup the environment is a
>> breeze. Using Hombrew just need to bring to a new system java, git,
>>maven
>> and from there I get all downloaded and built. maven downloads only
>>what is
>> needed and I don't need to worry about it. Even don't need to set up
>> environment variables, what's so cool :)
>> 
>> just my 2ctnms
>> 
>> C.
>> 
>> 
>> 2017-11-15 11:04 GMT+01:00 Harbs :
>> 
>>> The original topic was npm, but since we’re talking about Maven… ;-)
>>> 
>>> It seems to me that Royale should have a single Maven artifacts
>>>dependency
>>> that users should be able to specify which pulls in all the possibly
>>>needed
>>> swcs.
>>> 
>>> Maybe I’m missing something, but from the maven examples I’ve seen, it
>>> looks like you need to specify which swcs are needed for them to be
>>>pulled
>>> in. That seems too error-prone.
>>> 
 On Nov 15, 2017, at 11:52 AM, Piotr Zarzycki

>>> wrote:
 
 In case of Maven artifacts the only needs of doing one build is for
 convenient people who are using Maven build to develop SDK itself. If
>>> I'm a
 user whom would like to use Royale and build my own application by
Maven
>>> I
 don't need download repository and build myself whole sources. I just
>>> need
 to create simple pom file and all artifacts will be downloaded from
>>> Apache
 Maven central - my application will build.
 
 Piotr
 
 
 
 2017-11-15 10:46 GMT+01:00 Harbs :
 
> +1
> 
> I think this is the simplest way to handle it.
> 
> The only downside is that folks who don’t need the framework will
>>> download
> more than they need. but hard-drive space is pretty cheap and like
>>> Carlos
> says, we can split if afterwards if there’s demand.
> 
>> On Nov 15, 2017, at 11:38 AM, Carlos Rovira
>>
> wrote:
>> 
>> Hi,
>> 
>> I prefer :
>> 
>> * only one vote thread
>> * compiler bundled (no release separately) - if people demand it, we
> always
>> can do that
>> 
>> about maven, I remember there's 3 separate builds due to how maven
>>make
>> things, I'd like someone with maven skills could finaly join the
>>three
> into
>> one, that was something Chris was planning to do. The final step
>>would
>>> be
>> making only one "mvn clean install" and have compiler, typedefs and
>>> asjs
>> compiled and ready
>> 
>> thanks
>> 
>> 
>> 2017-11-15 9:09 GMT+01:00 Alex Harui :
>> 
>>> Hi Piotr,
>>> 
>>> That's fine, we'll see what others think, but we are also
>>>discussing
>>> whether the compiler is a separate release and vote thread or is
>>> bundled
>>> with the framework.
>>> 
>>> -Alex
>>> 
>>> On 11/15/17, 12:03 AM, "Piotr Zarzycki" 
> wrote:
>>> 
 Hi,
 
 Yep we didn't decide it yet how should be package 

Re: [royale-asjs] branch develop updated: Add ‘glue’ utility script to allow WASM to be loaded in HTML

[Carlos Rovira]

Hi,

I'm seeing lot of attention this days for WebAssembly due to Safari and
Edge supporting it.
The expectations are high on that front and I'm reading that WASM could be
almost ready for Spring in 2018.
So I think this is a must for us in addition to HTML/JS/CSS effort. I'm
trying to see the picture and even seems that both
HTML/JS/CSS and WASM could be used at the same time right?, so in that
case, we should think how this could be handled in Royale.




2017-11-15 7:47 GMT+01:00 Erik de Bruin :

> I have reverted my changes to develop@royale-asjs.
>
> EdB
>
>
>
> On Tue, Nov 14, 2017 at 11:57 PM, Piotr Zarzycki <
> piotrzarzyck...@gmail.com>
> wrote:
>
> > +1 for this. As I said it will help me to adjust maven build if it go to
> > separate branch.
> >
> > Thanks, Piotr
> >
> > On Tue, Nov 14, 2017, 23:44 Alex Harui  wrote:
> >
> > > It is great to see progress on this front.
> > >
> > > Logistical question:  We are hoping to cut a release of Royale "soon".
> > Is
> > > this code intended to ship in that release?  Are we ready to draw
> > > attention to the WebASM effort in the release?  Otherwise, can this
> stuff
> > > be in a feature branch so we don't have to adjust the release scripts
> to
> > > filter out this code?
> > >
> > > Thanks,
> > > -Alex
> > >
> > > On 11/14/17, 1:35 PM, "Harbs"  wrote:
> > >
> > > >I’m curious about the context of this script.
> > > >
> > > >Will all environments where this is being run understand ES6 arrow
> > > >functions?
> > > >
> > > >> On Nov 14, 2017, at 5:54 PM, erikdebr...@apache.org wrote:
> > > >>
> > > >> This is an automated email from the ASF dual-hosted git repository.
> > > >>
> > > >> erikdebruin pushed a commit to branch develop
> > > >> in repository
> > > >>
> > > https://na01.safelinks.protection.outlook.com/?url=
> > https%3A%2F%2Fgitbox.a
> > > >>pache.org
> > > %2Frepos%2Fasf%2Froyale-asjs.git=02%7C01%7C%7Cb270a11a588f4
> > > >>79744cf08d52ba794b0%7Cfa7b1b5a7b34438794aed2c178de
> > cee1%7C0%7C0%7C63646292
> > > >>1125471254=wgy8IDd9OnNwic3mKCVY05t5iF9WOk
> > Usxl%2FK6yg0YNU%3D
> > > >>d=0
> > > >>
> > > >>
> > > >> The following commit(s) were added to refs/heads/develop by this
> push:
> > > >> new 3dc37ce  Add ‘glue’ utility script to allow WASM to be
> loaded
> > > >>in HTML
> > > >> 3dc37ce is described below
> > > >>
> > > >> commit 3dc37ceb90c20be37639068061dba14b79e9461b
> > > >> Author: Erik de Bruin 
> > > >> AuthorDate: Tue Nov 14 16:54:13 2017 +0100
> > > >>
> > > >>Add ‘glue’ utility script to allow WASM to be loaded in HTML
> > > >>
> > > >>Signed-off-by: Erik de Bruin 
> > > >> ---
> > > >> .gitignore |  3 +++
> > > >> wast/resources/glue.js | 11 +++
> > > >> 2 files changed, 14 insertions(+)
> > > >>
> > > >> diff --git a/.gitignore b/.gitignore
> > > >> index a32c5ce..64c80a8 100644
> > > >> --- a/.gitignore
> > > >> +++ b/.gitignore
> > > >> @@ -141,3 +141,6 @@ vf2js/frameworks/js/
> > > >> #FlexJS generated files
> > > >> frameworks/js/FlexJS/generated-sources
> > > >> manualtests/FlexJSTest_SVG/bin
> > > >> +
> > > >> +#WAST generated files
> > > >> +/wast/examples/HelloWorld/bin
> > > >> diff --git a/wast/resources/glue.js b/wast/resources/glue.js
> > > >> new file mode 100644
> > > >> index 000..fe66877
> > > >> --- /dev/null
> > > >> +++ b/wast/resources/glue.js
> > > >> @@ -0,0 +1,11 @@
> > > >> +function fetchAndInstantiate(url, importObject) {
> > > >> +  return fetch(url).then(response =>
> > > >> +response.arrayBuffer()
> > > >> +  )
> > > >> +  .then(bytes =>
> > > >> +WebAssembly.instantiate(bytes, importObject)
> > > >> +  )
> > > >> +  .then(results =>
> > > >> +results.instance
> > > >> +  );
> > > >> +}
> > > >>
> > > >> --
> > > >> To stop receiving notification emails like this one, please contact
> > > >> ['"comm...@royale.apache.org" '].
> > > >
> > >
> > >
> >
>
>
>
> --
> Ix Multimedia Software
>
> Jan Luykenstraat 27
> 3521 VB Utrecht
>
> T. 06-51952295
> I. www.ixsoftware.nl
>



-- 
Carlos Rovira
http://about.me/carlosrovira

RE: Ant Build Problem

[Yishay Weiss]

I changed ROYALE_SWF_COMPILER_HOME and FLEXUNIT_HOME to match yours but the 
result is the same.

FWIW this is ROYALE_COMPILER_HOME which appears to not resolve correctly:
C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale>echo 
%ROYALE_COMPILER_HOME%
C:\dev\flexjs\royale-asjs/js
From: Harbs
Sent: Wednesday, November 15, 2017 12:01 PM
To: dev@royale.apache.org
Subject: Re: Ant Build Problem

It works for me. It looks like your ROYALE_SWF_COMPILER_HOME is pointing to a 
different location than mine.

Here’s the output I get:

CoreTest:

check-for-tests:

check-compiler-home:

check-transpiler-home:

check-compiler:

test:
[available] DEPRECATED -  used to override an existing property.
[available]   Build file should not reuse the same property name for different 
values.

clean:

compile:
 [echo] Compiling FlexUnitRoyaleApplication.swf
 [echo] ROYALE_HOME: /Users/harbs/Documents/ApacheRoyale/royale-asjs
 [echo] ROYALE_SWF_COMPILER_HOME: 
/Users/harbs/Documents/ApacheRoyale/royale-asjs
 [echo] FLEXUNIT_HOME: /Users/harbs/Documents/ApacheRoyale/flex-flexunit
 [echo] playerglobal.version: 11.1
[mxmlc] MXMLJSC
[mxmlc] -debug
[mxmlc] -compiler.targets=SWF
[mxmlc] +playerglobal.version=11.1
[mxmlc] 
+env.PLAYERGLOBAL_HOME=/Users/harbs/Documents/ApacheRoyale/frameworks/libs/player
[mxmlc] 
-swf-library-path+=/Users/harbs/Documents/ApacheRoyale/royale-asjs/frameworks/libs
[mxmlc] 
-swf-library-path+=/Users/harbs/Documents/ApacheRoyale/flex-flexunit/FlexUnit4/target/flexunit-4.3.0-20140410-as3_4.12.0.swc
[mxmlc] 
-swf-library-path+=/Users/harbs/Documents/ApacheRoyale/flex-flexunit/FlexUnit4CIListener/target
[mxmlc] 
+royalelib=/Users/harbs/Documents/ApacheRoyale/royale-asjs/frameworks/
[mxmlc] 
-output=/Users/harbs/Documents/ApacheRoyale/royale-asjs/frameworks/projects/Core/src/test/royale/FlexUnitRoyaleApplication.swf
[mxmlc] --
[mxmlc] 
/Users/harbs/Documents/ApacheRoyale/royale-asjs/frameworks/projects/Core/src/test/royale/FlexUnitRoyaleApplication.mxml
[mxmlc] Loading configuration: 
/Users/harbs/Documents/ApacheRoyale/royale-asjs/frameworks/royale-config.xml


> On Nov 15, 2017, at 11:44 AM, Yishay Weiss  wrote:
>
> I’m not sure which -config.xml is being used but the ROYALE_COMPILER_HOME is 
> pointing to  C:\dev\flexjs\royale-asjs/js
>
>
>
> The tests are run as part of
>
>
>
> C:\dev\flexjs\royale-asjs> ant clean all
>
>
>
> When I try to run the test separately I get
>
>
>
> C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale>ant
>
> Buildfile: 
> C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale\build.xml
>
>
>
> clean:
>
>
>
> compile:
>
> [echo] Compiling FlexUnitRoyaleApplication.swf
>
> [echo] ROYALE_HOME: C:\dev\flexjs\royale-asjs
>
> [echo] ROYALE_SWF_COMPILER_HOME: C:\dev\flexjs\royale-compiler\compiler
>
> [echo] FLEXUNIT_HOME: C:\dev\flexjs\royale-asjs/in/flexunit
>
> [echo] playerglobal.version: 11.1
>
>
>
> BUILD FAILED
>
> C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale\build.xml:98:
>  
> C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale\${ROYALE_COMPILER_HOME}\lib
>  does not exist.
>
>
>
> 
> From: Alex Harui 
> Sent: Tuesday, November 14, 2017 8:22:01 PM
> To: dev@royale.apache.org
> Subject: Re: Ant Build Problem
>
> The FlexUnit compile is trying to use the Ant tasks instead of directly
> calling the compiler, which is probably worth testing.  You can see that
> it picks up the Ant tasks from
>
>
>   includes="compiler-royaleTasks.jar"/>
>
>
>
> If ROYALE_COMPILER_HOME is pointing somewhere unexpected, then you might
> be loading a really old compiler that is picking up a really old
> -config.xml that doesn't have COMPILE::SWF in it.
>
> HTH,
> -Alex
>
> On 11/14/17, 10:01 AM, "Yishay Weiss"  wrote:
>
>> The failure is on the compilation of FlexUnitRoyaleApplication.mxml. The
>> rest of royale-asjs seems to compile ok. One thing I noticed is that the
>> compiler is invoked differently.
>>
>> >   file="${basedir}/FlexUnitRoyaleApplication.mxml"
>>   output="${basedir}/FlexUnitRoyaleApplication.swf">
>>   
>>   
>>   
>>   
>>   > />
>>   > value="-swf-library-path+=${ROYALE_HOME}/frameworks/libs" />
>>   
>>   
>>   
>>
>> Versus
>>
>> 
>>   
>>   
>>   
>>   
>>   
>>   
>>   
>>   
>>   
>>   > value="-load-config=${basedir}/src/main/config/compile-swf-config.xml" />
>>   > value="-js-load-config=${ROYALE_HOME}/frameworks/js-config.xml" />
>>   > value="-js-load-config+=${basedir}/../../js/projects/${ant.project.name}JS
>> /src/main/config/compile-js-config.xml" 

Re: Publishing royale to npm

[Harbs]

> why too error-prone?

Because it requires users to know which classes are in which swc. I don’t see a 
reason for that. But, I’m not a Maven user, so take my thoughts with a grain of 
salt.

Harbs

> On Nov 15, 2017, at 12:54 PM, Carlos Rovira  wrote:
> 
> Hi Harbs,
> 
> why too error-prone? If you are developing a Royale app with Maven you
> create a pom, and you add libraries on demand. So your build is totally
> automated and you only need your project pom in the future and not manually
> check if libraries are the required ones or not. I see just the opposite
> and less error prone.
> 
> I know this is lastly a question on how each looks to ways to do things,
> but I use to look at how others are doing things and I think most of the
> projects at Apache are using maven as their way to build their projects and
> manage continuous integration, so it must be a safe way to do thing or
> maven will be less used today.
> 
> In my experience, I'm using a macbookpro and setup the environment is a
> breeze. Using Hombrew just need to bring to a new system java, git, maven
> and from there I get all downloaded and built. maven downloads only what is
> needed and I don't need to worry about it. Even don't need to set up
> environment variables, what's so cool :)
> 
> just my 2ctnms
> 
> C.
> 
> 
> 2017-11-15 11:04 GMT+01:00 Harbs :
> 
>> The original topic was npm, but since we’re talking about Maven… ;-)
>> 
>> It seems to me that Royale should have a single Maven artifacts dependency
>> that users should be able to specify which pulls in all the possibly needed
>> swcs.
>> 
>> Maybe I’m missing something, but from the maven examples I’ve seen, it
>> looks like you need to specify which swcs are needed for them to be pulled
>> in. That seems too error-prone.
>> 
>>> On Nov 15, 2017, at 11:52 AM, Piotr Zarzycki 
>> wrote:
>>> 
>>> In case of Maven artifacts the only needs of doing one build is for
>>> convenient people who are using Maven build to develop SDK itself. If
>> I'm a
>>> user whom would like to use Royale and build my own application by Maven
>> I
>>> don't need download repository and build myself whole sources. I just
>> need
>>> to create simple pom file and all artifacts will be downloaded from
>> Apache
>>> Maven central - my application will build.
>>> 
>>> Piotr
>>> 
>>> 
>>> 
>>> 2017-11-15 10:46 GMT+01:00 Harbs :
>>> 
 +1
 
 I think this is the simplest way to handle it.
 
 The only downside is that folks who don’t need the framework will
>> download
 more than they need. but hard-drive space is pretty cheap and like
>> Carlos
 says, we can split if afterwards if there’s demand.
 
> On Nov 15, 2017, at 11:38 AM, Carlos Rovira 
 wrote:
> 
> Hi,
> 
> I prefer :
> 
> * only one vote thread
> * compiler bundled (no release separately) - if people demand it, we
 always
> can do that
> 
> about maven, I remember there's 3 separate builds due to how maven make
> things, I'd like someone with maven skills could finaly join the three
 into
> one, that was something Chris was planning to do. The final step would
>> be
> making only one "mvn clean install" and have compiler, typedefs and
>> asjs
> compiled and ready
> 
> thanks
> 
> 
> 2017-11-15 9:09 GMT+01:00 Alex Harui :
> 
>> Hi Piotr,
>> 
>> That's fine, we'll see what others think, but we are also discussing
>> whether the compiler is a separate release and vote thread or is
>> bundled
>> with the framework.
>> 
>> -Alex
>> 
>> On 11/15/17, 12:03 AM, "Piotr Zarzycki" 
 wrote:
>> 
>>> Hi,
>>> 
>>> Yep we didn't decide it yet how should be package release. In my
 opinion
>>> this should look like that:
>>> 
>>> 1) Package called royale-flexjs -0.9 Where it compiles to SWF and JS
>>> 2) Package called royale-0.9 where it compiles to JS only.
>>> 
>>> I like the idea of voting once where whole framework is in place, in
 case
>>> of Maven during release process three repositories will land as
>> staging
>>> artifacts and we can vote.
>>> 
>>> Thoughts ?
>>> Piotr
>>> 
>>> 
>>> 2017-11-15 8:09 GMT+01:00 Alex Harui :
>>> 
 Royale will be using artifacts from royale-compiler, not
>> flex-falcon.
 
 I'm not sure we've decided on how to package our releases.  The Ant
 scripts are currently set up for two artifacts (compiler and
 framework),
 Maven is set up for 1 or 3, depending how you count.
 
 I'm pretty sure we'll have to adjust scripts anyway to smooth out
>> how
 Maven and Ant work together to create all of the artifacts so making

Re: Publishing royale to npm

[Carlos Rovira]

Hi Harbs,

why too error-prone? If you are developing a Royale app with Maven you
create a pom, and you add libraries on demand. So your build is totally
automated and you only need your project pom in the future and not manually
check if libraries are the required ones or not. I see just the opposite
and less error prone.

I know this is lastly a question on how each looks to ways to do things,
but I use to look at how others are doing things and I think most of the
projects at Apache are using maven as their way to build their projects and
manage continuous integration, so it must be a safe way to do thing or
maven will be less used today.

In my experience, I'm using a macbookpro and setup the environment is a
breeze. Using Hombrew just need to bring to a new system java, git, maven
and from there I get all downloaded and built. maven downloads only what is
needed and I don't need to worry about it. Even don't need to set up
environment variables, what's so cool :)

just my 2ctnms

C.


2017-11-15 11:04 GMT+01:00 Harbs :

> The original topic was npm, but since we’re talking about Maven… ;-)
>
> It seems to me that Royale should have a single Maven artifacts dependency
> that users should be able to specify which pulls in all the possibly needed
> swcs.
>
> Maybe I’m missing something, but from the maven examples I’ve seen, it
> looks like you need to specify which swcs are needed for them to be pulled
> in. That seems too error-prone.
>
> > On Nov 15, 2017, at 11:52 AM, Piotr Zarzycki 
> wrote:
> >
> > In case of Maven artifacts the only needs of doing one build is for
> > convenient people who are using Maven build to develop SDK itself. If
> I'm a
> > user whom would like to use Royale and build my own application by Maven
> I
> > don't need download repository and build myself whole sources. I just
> need
> > to create simple pom file and all artifacts will be downloaded from
> Apache
> > Maven central - my application will build.
> >
> > Piotr
> >
> >
> >
> > 2017-11-15 10:46 GMT+01:00 Harbs :
> >
> >> +1
> >>
> >> I think this is the simplest way to handle it.
> >>
> >> The only downside is that folks who don’t need the framework will
> download
> >> more than they need. but hard-drive space is pretty cheap and like
> Carlos
> >> says, we can split if afterwards if there’s demand.
> >>
> >>> On Nov 15, 2017, at 11:38 AM, Carlos Rovira 
> >> wrote:
> >>>
> >>> Hi,
> >>>
> >>> I prefer :
> >>>
> >>> * only one vote thread
> >>> * compiler bundled (no release separately) - if people demand it, we
> >> always
> >>> can do that
> >>>
> >>> about maven, I remember there's 3 separate builds due to how maven make
> >>> things, I'd like someone with maven skills could finaly join the three
> >> into
> >>> one, that was something Chris was planning to do. The final step would
> be
> >>> making only one "mvn clean install" and have compiler, typedefs and
> asjs
> >>> compiled and ready
> >>>
> >>> thanks
> >>>
> >>>
> >>> 2017-11-15 9:09 GMT+01:00 Alex Harui :
> >>>
>  Hi Piotr,
> 
>  That's fine, we'll see what others think, but we are also discussing
>  whether the compiler is a separate release and vote thread or is
> bundled
>  with the framework.
> 
>  -Alex
> 
>  On 11/15/17, 12:03 AM, "Piotr Zarzycki" 
> >> wrote:
> 
> > Hi,
> >
> > Yep we didn't decide it yet how should be package release. In my
> >> opinion
> > this should look like that:
> >
> > 1) Package called royale-flexjs -0.9 Where it compiles to SWF and JS
> > 2) Package called royale-0.9 where it compiles to JS only.
> >
> > I like the idea of voting once where whole framework is in place, in
> >> case
> > of Maven during release process three repositories will land as
> staging
> > artifacts and we can vote.
> >
> > Thoughts ?
> > Piotr
> >
> >
> > 2017-11-15 8:09 GMT+01:00 Alex Harui :
> >
> >> Royale will be using artifacts from royale-compiler, not
> flex-falcon.
> >>
> >> I'm not sure we've decided on how to package our releases.  The Ant
> >> scripts are currently set up for two artifacts (compiler and
> >> framework),
> >> Maven is set up for 1 or 3, depending how you count.
> >>
> >> I'm pretty sure we'll have to adjust scripts anyway to smooth out
> how
> >> Maven and Ant work together to create all of the artifacts so making
> >> other
> >> adjustments for npm is an option too.  Maybe the first question is:
> >> how
> >> many vote threads do we want?  I believe eventually we rate of
> change
> >> in
> >> royale-compiler will slow compared to royale-asjs and changes to
> >> royale-asjs won't depend on changes in royale-compiler, but we could
> >> change our packaging and number of vote 

Re: Ant Build Problem

[Harbs]

It works for me. It looks like your ROYALE_SWF_COMPILER_HOME is pointing to a 
different location than mine.

Here’s the output I get:

CoreTest:

check-for-tests:

check-compiler-home:

check-transpiler-home:

check-compiler:

test:
[available] DEPRECATED -  used to override an existing property.
[available]   Build file should not reuse the same property name for different 
values.

clean:

compile:
 [echo] Compiling FlexUnitRoyaleApplication.swf
 [echo] ROYALE_HOME: /Users/harbs/Documents/ApacheRoyale/royale-asjs
 [echo] ROYALE_SWF_COMPILER_HOME: 
/Users/harbs/Documents/ApacheRoyale/royale-asjs
 [echo] FLEXUNIT_HOME: /Users/harbs/Documents/ApacheRoyale/flex-flexunit
 [echo] playerglobal.version: 11.1
[mxmlc] MXMLJSC
[mxmlc] -debug
[mxmlc] -compiler.targets=SWF
[mxmlc] +playerglobal.version=11.1
[mxmlc] 
+env.PLAYERGLOBAL_HOME=/Users/harbs/Documents/ApacheRoyale/frameworks/libs/player
[mxmlc] 
-swf-library-path+=/Users/harbs/Documents/ApacheRoyale/royale-asjs/frameworks/libs
[mxmlc] 
-swf-library-path+=/Users/harbs/Documents/ApacheRoyale/flex-flexunit/FlexUnit4/target/flexunit-4.3.0-20140410-as3_4.12.0.swc
[mxmlc] 
-swf-library-path+=/Users/harbs/Documents/ApacheRoyale/flex-flexunit/FlexUnit4CIListener/target
[mxmlc] 
+royalelib=/Users/harbs/Documents/ApacheRoyale/royale-asjs/frameworks/
[mxmlc] 
-output=/Users/harbs/Documents/ApacheRoyale/royale-asjs/frameworks/projects/Core/src/test/royale/FlexUnitRoyaleApplication.swf
[mxmlc] --
[mxmlc] 
/Users/harbs/Documents/ApacheRoyale/royale-asjs/frameworks/projects/Core/src/test/royale/FlexUnitRoyaleApplication.mxml
[mxmlc] Loading configuration: 
/Users/harbs/Documents/ApacheRoyale/royale-asjs/frameworks/royale-config.xml


> On Nov 15, 2017, at 11:44 AM, Yishay Weiss  wrote:
> 
> I’m not sure which -config.xml is being used but the ROYALE_COMPILER_HOME is 
> pointing to  C:\dev\flexjs\royale-asjs/js
> 
> 
> 
> The tests are run as part of
> 
> 
> 
> C:\dev\flexjs\royale-asjs> ant clean all
> 
> 
> 
> When I try to run the test separately I get
> 
> 
> 
> C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale>ant
> 
> Buildfile: 
> C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale\build.xml
> 
> 
> 
> clean:
> 
> 
> 
> compile:
> 
> [echo] Compiling FlexUnitRoyaleApplication.swf
> 
> [echo] ROYALE_HOME: C:\dev\flexjs\royale-asjs
> 
> [echo] ROYALE_SWF_COMPILER_HOME: C:\dev\flexjs\royale-compiler\compiler
> 
> [echo] FLEXUNIT_HOME: C:\dev\flexjs\royale-asjs/in/flexunit
> 
> [echo] playerglobal.version: 11.1
> 
> 
> 
> BUILD FAILED
> 
> C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale\build.xml:98:
>  
> C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale\${ROYALE_COMPILER_HOME}\lib
>  does not exist.
> 
> 
> 
> 
> From: Alex Harui 
> Sent: Tuesday, November 14, 2017 8:22:01 PM
> To: dev@royale.apache.org
> Subject: Re: Ant Build Problem
> 
> The FlexUnit compile is trying to use the Ant tasks instead of directly
> calling the compiler, which is probably worth testing.  You can see that
> it picks up the Ant tasks from
> 
>
>   includes="compiler-royaleTasks.jar"/>
>
> 
> 
> If ROYALE_COMPILER_HOME is pointing somewhere unexpected, then you might
> be loading a really old compiler that is picking up a really old
> -config.xml that doesn't have COMPILE::SWF in it.
> 
> HTH,
> -Alex
> 
> On 11/14/17, 10:01 AM, "Yishay Weiss"  wrote:
> 
>> The failure is on the compilation of FlexUnitRoyaleApplication.mxml. The
>> rest of royale-asjs seems to compile ok. One thing I noticed is that the
>> compiler is invoked differently.
>> 
>> >   file="${basedir}/FlexUnitRoyaleApplication.mxml"
>>   output="${basedir}/FlexUnitRoyaleApplication.swf">
>>   
>>   
>>   
>>   
>>   > />
>>   > value="-swf-library-path+=${ROYALE_HOME}/frameworks/libs" />
>>   
>>   
>>   
>> 
>> Versus
>> 
>> 
>>   
>>   
>>   
>>   
>>   
>>   
>>   
>>   
>>   
>>   > value="-load-config=${basedir}/src/main/config/compile-swf-config.xml" />
>>   > value="-js-load-config=${ROYALE_HOME}/frameworks/js-config.xml" />
>>   > value="-js-load-config+=${basedir}/../../js/projects/${ant.project.name}JS
>> /src/main/config/compile-js-config.xml" />
>>   
>> 
>> From: Alex Harui
>> Sent: Tuesday, November 14, 2017 7:35 PM
>> To: dev@royale.apache.org
>> Subject: Re: Ant Build Problem
>> 
>> Even if you did, it is still pretty easy for an environment variable or
>> Ant property to get you to use a compiler in another folder, so keep an
>> eye open for that.  The Ant scripts usually report what compiler 

Re: Publishing royale to npm

[Piotr Zarzycki]

In case of Maven artifacts the only needs of doing one build is for
convenient people who are using Maven build to develop SDK itself. If I'm a
user whom would like to use Royale and build my own application by Maven I
don't need download repository and build myself whole sources. I just need
to create simple pom file and all artifacts will be downloaded from Apache
Maven central - my application will build.

Piotr



2017-11-15 10:46 GMT+01:00 Harbs :

> +1
>
> I think this is the simplest way to handle it.
>
> The only downside is that folks who don’t need the framework will download
> more than they need. but hard-drive space is pretty cheap and like Carlos
> says, we can split if afterwards if there’s demand.
>
> > On Nov 15, 2017, at 11:38 AM, Carlos Rovira 
> wrote:
> >
> > Hi,
> >
> > I prefer :
> >
> > * only one vote thread
> > * compiler bundled (no release separately) - if people demand it, we
> always
> > can do that
> >
> > about maven, I remember there's 3 separate builds due to how maven make
> > things, I'd like someone with maven skills could finaly join the three
> into
> > one, that was something Chris was planning to do. The final step would be
> > making only one "mvn clean install" and have compiler, typedefs and asjs
> > compiled and ready
> >
> > thanks
> >
> >
> > 2017-11-15 9:09 GMT+01:00 Alex Harui :
> >
> >> Hi Piotr,
> >>
> >> That's fine, we'll see what others think, but we are also discussing
> >> whether the compiler is a separate release and vote thread or is bundled
> >> with the framework.
> >>
> >> -Alex
> >>
> >> On 11/15/17, 12:03 AM, "Piotr Zarzycki" 
> wrote:
> >>
> >>> Hi,
> >>>
> >>> Yep we didn't decide it yet how should be package release. In my
> opinion
> >>> this should look like that:
> >>>
> >>> 1) Package called royale-flexjs -0.9 Where it compiles to SWF and JS
> >>> 2) Package called royale-0.9 where it compiles to JS only.
> >>>
> >>> I like the idea of voting once where whole framework is in place, in
> case
> >>> of Maven during release process three repositories will land as staging
> >>> artifacts and we can vote.
> >>>
> >>> Thoughts ?
> >>> Piotr
> >>>
> >>>
> >>> 2017-11-15 8:09 GMT+01:00 Alex Harui :
> >>>
>  Royale will be using artifacts from royale-compiler, not flex-falcon.
> 
>  I'm not sure we've decided on how to package our releases.  The Ant
>  scripts are currently set up for two artifacts (compiler and
> framework),
>  Maven is set up for 1 or 3, depending how you count.
> 
>  I'm pretty sure we'll have to adjust scripts anyway to smooth out how
>  Maven and Ant work together to create all of the artifacts so making
>  other
>  adjustments for npm is an option too.  Maybe the first question is:
> how
>  many vote threads do we want?  I believe eventually we rate of change
> in
>  royale-compiler will slow compared to royale-asjs and changes to
>  royale-asjs won't depend on changes in royale-compiler, but we could
>  change our packaging and number of vote threads later.
> 
>  Thoughts?
>  -Alex
> 
>  On 11/14/17, 10:44 PM, "omup...@gmail.com on behalf of OmPrakash
>  Muppirala" 
> wrote:
> 
> > I'm in the middle of updating the npm scripts.  Are we planning to
> > continue
> > to download flex-falcon from the current release?  Or are we planning
>  on
> > pushing out a new release of falcon/royale-compiler?
> >
> > Thanks,
> > Om
> >
> > On Mon, Nov 13, 2017 at 12:57 AM, Alex Harui
>  >>>
> > wrote:
> >
> >> Yes, I was going to copy it someday so go ahead and do it now if you
> >> have
> >> time.
> >>
> >> -Alex
> >>
> >> On 11/13/17, 12:50 AM, "Harbs"  wrote:
> >>
> >>> Sounds good.
> >>>
>  On Nov 13, 2017, at 10:41 AM, OmPrakash Muppirala
>   wrote:
> 
>  BTW, I just realized that we don't have a royale-utilities git
>  repo.
> >> I
>  am
>  tempted to simply put the npm related code into royale-asjs/npm
>  directory
>  and add it as an exclude in the build.xml.  Any objections?
> 
>  Thanks,
>  Om
> 
>  On Sun, Nov 12, 2017 at 1:58 PM, Harbs 
>  wrote:
> 
> > If you call it 0.10.0, I think it’s pretty clear.
> >
> >> On Nov 12, 2017, at 11:46 PM, Idylog - Nicolas Granon <
> > ngra...@idylog.com> wrote:
> >>
> >> In developer's eyes, "0.10" is "lower" than "0.9".
> >>
> >> You'd better number it as "0.91".
> >>
> >> Nicolas Granon
> >>
> >>
> >>
> >>> 

RE: Ant Build Problem

[Yishay Weiss]

I’m not sure which -config.xml is being used but the ROYALE_COMPILER_HOME is 
pointing to  C:\dev\flexjs\royale-asjs/js



The tests are run as part of



C:\dev\flexjs\royale-asjs> ant clean all



When I try to run the test separately I get



C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale>ant

Buildfile: 
C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale\build.xml



clean:



compile:

 [echo] Compiling FlexUnitRoyaleApplication.swf

 [echo] ROYALE_HOME: C:\dev\flexjs\royale-asjs

 [echo] ROYALE_SWF_COMPILER_HOME: C:\dev\flexjs\royale-compiler\compiler

 [echo] FLEXUNIT_HOME: C:\dev\flexjs\royale-asjs/in/flexunit

 [echo] playerglobal.version: 11.1



BUILD FAILED

C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale\build.xml:98:
 
C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale\${ROYALE_COMPILER_HOME}\lib
 does not exist.




From: Alex Harui 
Sent: Tuesday, November 14, 2017 8:22:01 PM
To: dev@royale.apache.org
Subject: Re: Ant Build Problem

The FlexUnit compile is trying to use the Ant tasks instead of directly
calling the compiler, which is probably worth testing.  You can see that
it picks up the Ant tasks from


  



If ROYALE_COMPILER_HOME is pointing somewhere unexpected, then you might
be loading a really old compiler that is picking up a really old
-config.xml that doesn't have COMPILE::SWF in it.

HTH,
-Alex

On 11/14/17, 10:01 AM, "Yishay Weiss"  wrote:

>The failure is on the compilation of FlexUnitRoyaleApplication.mxml. The
>rest of royale-asjs seems to compile ok. One thing I noticed is that the
>compiler is invoked differently.
>
>file="${basedir}/FlexUnitRoyaleApplication.mxml"
>output="${basedir}/FlexUnitRoyaleApplication.swf">
>
>
>
>
>/>
>value="-swf-library-path+=${ROYALE_HOME}/frameworks/libs" />
>
>
>
>
>Versus
>
>
>
>
>
>
>
>
>
>
>
>value="-load-config=${basedir}/src/main/config/compile-swf-config.xml" />
>value="-js-load-config=${ROYALE_HOME}/frameworks/js-config.xml" />
>value="-js-load-config+=${basedir}/../../js/projects/${ant.project.name}JS
>/src/main/config/compile-js-config.xml" />
>
>
>From: Alex Harui
>Sent: Tuesday, November 14, 2017 7:35 PM
>To: dev@royale.apache.org
>Subject: Re: Ant Build Problem
>
>Even if you did, it is still pretty easy for an environment variable or
>Ant property to get you to use a compiler in another folder, so keep an
>eye open for that.  The Ant scripts usually report what compiler folder
>they are using.
>
>-Alex
>
>On 11/14/17, 9:32 AM, "Yishay Weiss"  wrote:
>
>>I updated compiler and typedefs as well. Will have a look tomorrow.
>>
>>
>>
>>
>>From: Alex Harui 
>>Sent: Tuesday, November 14, 2017 6:25:55 PM
>>To: dev@royale.apache.org
>>Subject: Re: Ant Build Problem
>>
>>I haven't seen that.  You are working with relatively new test code,
>>IIRC.
>> Could be you are using very old compiler?
>>
>>HTH,
>>-Alex
>>
>>On 11/14/17, 3:26 AM, "Yishay Weiss"  wrote:
>>
>>>[mxmlc]
>>>C:\dev\flexjs\royale-asjs\frameworks\projects\Core\src\test\royale\flexU
>>>n
>>>i
>>>tTests\KeyboardEventConverterTest.
>>>as(23): col: 5 Error: Can not resolve config constant: 'SWF'
>>>[mxmlc]
>>>[mxmlc] COMPILE::SWF
>>>[mxmlc] ^
>>>[mxmlc]
>>>
>>>Is anyone else getting it?
>>>
>>
>

Re: Publishing royale to npm

[Carlos Rovira]

Hi,

I prefer :

* only one vote thread
* compiler bundled (no release separately) - if people demand it, we always
can do that

about maven, I remember there's 3 separate builds due to how maven make
things, I'd like someone with maven skills could finaly join the three into
one, that was something Chris was planning to do. The final step would be
making only one "mvn clean install" and have compiler, typedefs and asjs
compiled and ready

thanks


2017-11-15 9:09 GMT+01:00 Alex Harui :

> Hi Piotr,
>
> That's fine, we'll see what others think, but we are also discussing
> whether the compiler is a separate release and vote thread or is bundled
> with the framework.
>
> -Alex
>
> On 11/15/17, 12:03 AM, "Piotr Zarzycki"  wrote:
>
> >Hi,
> >
> >Yep we didn't decide it yet how should be package release. In my opinion
> >this should look like that:
> >
> >1) Package called royale-flexjs -0.9 Where it compiles to SWF and JS
> >2) Package called royale-0.9 where it compiles to JS only.
> >
> >I like the idea of voting once where whole framework is in place, in case
> >of Maven during release process three repositories will land as staging
> >artifacts and we can vote.
> >
> >Thoughts ?
> >Piotr
> >
> >
> >2017-11-15 8:09 GMT+01:00 Alex Harui :
> >
> >> Royale will be using artifacts from royale-compiler, not flex-falcon.
> >>
> >> I'm not sure we've decided on how to package our releases.  The Ant
> >> scripts are currently set up for two artifacts (compiler and framework),
> >> Maven is set up for 1 or 3, depending how you count.
> >>
> >> I'm pretty sure we'll have to adjust scripts anyway to smooth out how
> >> Maven and Ant work together to create all of the artifacts so making
> >>other
> >> adjustments for npm is an option too.  Maybe the first question is:  how
> >> many vote threads do we want?  I believe eventually we rate of change in
> >> royale-compiler will slow compared to royale-asjs and changes to
> >> royale-asjs won't depend on changes in royale-compiler, but we could
> >> change our packaging and number of vote threads later.
> >>
> >> Thoughts?
> >> -Alex
> >>
> >> On 11/14/17, 10:44 PM, "omup...@gmail.com on behalf of OmPrakash
> >> Muppirala"  wrote:
> >>
> >> >I'm in the middle of updating the npm scripts.  Are we planning to
> >> >continue
> >> >to download flex-falcon from the current release?  Or are we planning
> >>on
> >> >pushing out a new release of falcon/royale-compiler?
> >> >
> >> >Thanks,
> >> >Om
> >> >
> >> >On Mon, Nov 13, 2017 at 12:57 AM, Alex Harui  >
> >> >wrote:
> >> >
> >> >> Yes, I was going to copy it someday so go ahead and do it now if you
> >> >>have
> >> >> time.
> >> >>
> >> >> -Alex
> >> >>
> >> >> On 11/13/17, 12:50 AM, "Harbs"  wrote:
> >> >>
> >> >> >Sounds good.
> >> >> >
> >> >> >> On Nov 13, 2017, at 10:41 AM, OmPrakash Muppirala
> >> >> >> wrote:
> >> >> >>
> >> >> >> BTW, I just realized that we don't have a royale-utilities git
> >>repo.
> >> >> I
> >> >> >>am
> >> >> >> tempted to simply put the npm related code into royale-asjs/npm
> >> >> >>directory
> >> >> >> and add it as an exclude in the build.xml.  Any objections?
> >> >> >>
> >> >> >> Thanks,
> >> >> >> Om
> >> >> >>
> >> >> >> On Sun, Nov 12, 2017 at 1:58 PM, Harbs 
> >> wrote:
> >> >> >>
> >> >> >>> If you call it 0.10.0, I think it’s pretty clear.
> >> >> >>>
> >> >>  On Nov 12, 2017, at 11:46 PM, Idylog - Nicolas Granon <
> >> >> >>> ngra...@idylog.com> wrote:
> >> >> 
> >> >>  In developer's eyes, "0.10" is "lower" than "0.9".
> >> >> 
> >> >>  You'd better number it as "0.91".
> >> >> 
> >> >>  Nicolas Granon
> >> >> 
> >> >> 
> >> >> 
> >> >> > -Message d'origine-
> >> >> > De : Harbs [mailto:harbs.li...@gmail.com]
> >> >> > Envoyé : dimanche 12 novembre 2017 10:15
> >> >> > À : dev@royale.apache.org
> >> >> > Objet : Re: Publishing royale to npm
> >> >> >
> >> >> > We’ve fixed a lot of things since 0.8.
> >> >> >
> >> >> > 0.9 does not need to jump to 1.0. We can have 0.10 (and 0.11…)
> >> >> >
> >> >> > I’d really like to streamline the release process so it’s
> >>painless
> >> >> > enough to release every couple/few weeks.
> >> >> >
> >> >> > Harbs
> >> >> >
> >> >> >> On Nov 12, 2017, at 10:35 AM, Carlos Rovira
> >> >> >>
> >> >> > wrote:
> >> >> >>
> >> >> >> Hi,
> >> >> >>
> >> >> >> about version number, I'll feel more happy with 0.8 than 0.9
> >>My
> >> >> >>point
> >> >> >> is that we are getting a first release of Royale and is the
> >>same
> >> >> >>that
> >> >> >> FlexJS 0.8 but with some new fixes and little things.
> >> >> >> As well I can see many things to do to reach 

Re: Publishing royale to npm

[Alex Harui]

Hi Piotr,

That's fine, we'll see what others think, but we are also discussing
whether the compiler is a separate release and vote thread or is bundled
with the framework.

-Alex

On 11/15/17, 12:03 AM, "Piotr Zarzycki"  wrote:

>Hi,
>
>Yep we didn't decide it yet how should be package release. In my opinion
>this should look like that:
>
>1) Package called royale-flexjs -0.9 Where it compiles to SWF and JS
>2) Package called royale-0.9 where it compiles to JS only.
>
>I like the idea of voting once where whole framework is in place, in case
>of Maven during release process three repositories will land as staging
>artifacts and we can vote.
>
>Thoughts ?
>Piotr
>
>
>2017-11-15 8:09 GMT+01:00 Alex Harui :
>
>> Royale will be using artifacts from royale-compiler, not flex-falcon.
>>
>> I'm not sure we've decided on how to package our releases.  The Ant
>> scripts are currently set up for two artifacts (compiler and framework),
>> Maven is set up for 1 or 3, depending how you count.
>>
>> I'm pretty sure we'll have to adjust scripts anyway to smooth out how
>> Maven and Ant work together to create all of the artifacts so making
>>other
>> adjustments for npm is an option too.  Maybe the first question is:  how
>> many vote threads do we want?  I believe eventually we rate of change in
>> royale-compiler will slow compared to royale-asjs and changes to
>> royale-asjs won't depend on changes in royale-compiler, but we could
>> change our packaging and number of vote threads later.
>>
>> Thoughts?
>> -Alex
>>
>> On 11/14/17, 10:44 PM, "omup...@gmail.com on behalf of OmPrakash
>> Muppirala"  wrote:
>>
>> >I'm in the middle of updating the npm scripts.  Are we planning to
>> >continue
>> >to download flex-falcon from the current release?  Or are we planning
>>on
>> >pushing out a new release of falcon/royale-compiler?
>> >
>> >Thanks,
>> >Om
>> >
>> >On Mon, Nov 13, 2017 at 12:57 AM, Alex Harui 
>> >wrote:
>> >
>> >> Yes, I was going to copy it someday so go ahead and do it now if you
>> >>have
>> >> time.
>> >>
>> >> -Alex
>> >>
>> >> On 11/13/17, 12:50 AM, "Harbs"  wrote:
>> >>
>> >> >Sounds good.
>> >> >
>> >> >> On Nov 13, 2017, at 10:41 AM, OmPrakash Muppirala
>> >> >> wrote:
>> >> >>
>> >> >> BTW, I just realized that we don't have a royale-utilities git
>>repo.
>> >> I
>> >> >>am
>> >> >> tempted to simply put the npm related code into royale-asjs/npm
>> >> >>directory
>> >> >> and add it as an exclude in the build.xml.  Any objections?
>> >> >>
>> >> >> Thanks,
>> >> >> Om
>> >> >>
>> >> >> On Sun, Nov 12, 2017 at 1:58 PM, Harbs 
>> wrote:
>> >> >>
>> >> >>> If you call it 0.10.0, I think it’s pretty clear.
>> >> >>>
>> >>  On Nov 12, 2017, at 11:46 PM, Idylog - Nicolas Granon <
>> >> >>> ngra...@idylog.com> wrote:
>> >> 
>> >>  In developer's eyes, "0.10" is "lower" than "0.9".
>> >> 
>> >>  You'd better number it as "0.91".
>> >> 
>> >>  Nicolas Granon
>> >> 
>> >> 
>> >> 
>> >> > -Message d'origine-
>> >> > De : Harbs [mailto:harbs.li...@gmail.com]
>> >> > Envoyé : dimanche 12 novembre 2017 10:15
>> >> > À : dev@royale.apache.org
>> >> > Objet : Re: Publishing royale to npm
>> >> >
>> >> > We’ve fixed a lot of things since 0.8.
>> >> >
>> >> > 0.9 does not need to jump to 1.0. We can have 0.10 (and 0.11…)
>> >> >
>> >> > I’d really like to streamline the release process so it’s
>>painless
>> >> > enough to release every couple/few weeks.
>> >> >
>> >> > Harbs
>> >> >
>> >> >> On Nov 12, 2017, at 10:35 AM, Carlos Rovira
>> >> >>
>> >> > wrote:
>> >> >>
>> >> >> Hi,
>> >> >>
>> >> >> about version number, I'll feel more happy with 0.8 than 0.9
>>My
>> >> >>point
>> >> >> is that we are getting a first release of Royale and is the
>>same
>> >> >>that
>> >> >> FlexJS 0.8 but with some new fixes and little things.
>> >> >> As well I can see many things to do to reach 1.0, and would be
>> >>great
>> >> >> to have still an intermediate release 0.9
>> >> >>
>> >> >>
>> >> >>
>> >> >> 2017-11-11 22:10 GMT+01:00 OmPrakash Muppirala
>> >> > :
>> >> >>
>> >> >>> On Thu, Nov 9, 2017 at 6:41 PM, Alex Harui
>> >> > 
>> >> >>> wrote:
>> >> >>>
>> >>  That might work.  One question:  we want the same bits that
>> >>were
>> >> >>> published
>> >>  as 0.9.0-rc1 to become the final bits where you would do:
>> >> 
>> >>  npm install -g apache-royale
>> >> 
>> >>  We aren't supposed to rebuild anything.  For Maven the same
>> >>bits
>> >>  from staging get copied to Maven central, for Ant/IDE the

Re: Publishing royale to npm

[Piotr Zarzycki]

Hi,

Yep we didn't decide it yet how should be package release. In my opinion
this should look like that:

1) Package called royale-flexjs -0.9 Where it compiles to SWF and JS
2) Package called royale-0.9 where it compiles to JS only.

I like the idea of voting once where whole framework is in place, in case
of Maven during release process three repositories will land as staging
artifacts and we can vote.

Thoughts ?
Piotr


2017-11-15 8:09 GMT+01:00 Alex Harui :

> Royale will be using artifacts from royale-compiler, not flex-falcon.
>
> I'm not sure we've decided on how to package our releases.  The Ant
> scripts are currently set up for two artifacts (compiler and framework),
> Maven is set up for 1 or 3, depending how you count.
>
> I'm pretty sure we'll have to adjust scripts anyway to smooth out how
> Maven and Ant work together to create all of the artifacts so making other
> adjustments for npm is an option too.  Maybe the first question is:  how
> many vote threads do we want?  I believe eventually we rate of change in
> royale-compiler will slow compared to royale-asjs and changes to
> royale-asjs won't depend on changes in royale-compiler, but we could
> change our packaging and number of vote threads later.
>
> Thoughts?
> -Alex
>
> On 11/14/17, 10:44 PM, "omup...@gmail.com on behalf of OmPrakash
> Muppirala"  wrote:
>
> >I'm in the middle of updating the npm scripts.  Are we planning to
> >continue
> >to download flex-falcon from the current release?  Or are we planning on
> >pushing out a new release of falcon/royale-compiler?
> >
> >Thanks,
> >Om
> >
> >On Mon, Nov 13, 2017 at 12:57 AM, Alex Harui 
> >wrote:
> >
> >> Yes, I was going to copy it someday so go ahead and do it now if you
> >>have
> >> time.
> >>
> >> -Alex
> >>
> >> On 11/13/17, 12:50 AM, "Harbs"  wrote:
> >>
> >> >Sounds good.
> >> >
> >> >> On Nov 13, 2017, at 10:41 AM, OmPrakash Muppirala
> >> >> wrote:
> >> >>
> >> >> BTW, I just realized that we don't have a royale-utilities git repo.
> >> I
> >> >>am
> >> >> tempted to simply put the npm related code into royale-asjs/npm
> >> >>directory
> >> >> and add it as an exclude in the build.xml.  Any objections?
> >> >>
> >> >> Thanks,
> >> >> Om
> >> >>
> >> >> On Sun, Nov 12, 2017 at 1:58 PM, Harbs 
> wrote:
> >> >>
> >> >>> If you call it 0.10.0, I think it’s pretty clear.
> >> >>>
> >>  On Nov 12, 2017, at 11:46 PM, Idylog - Nicolas Granon <
> >> >>> ngra...@idylog.com> wrote:
> >> 
> >>  In developer's eyes, "0.10" is "lower" than "0.9".
> >> 
> >>  You'd better number it as "0.91".
> >> 
> >>  Nicolas Granon
> >> 
> >> 
> >> 
> >> > -Message d'origine-
> >> > De : Harbs [mailto:harbs.li...@gmail.com]
> >> > Envoyé : dimanche 12 novembre 2017 10:15
> >> > À : dev@royale.apache.org
> >> > Objet : Re: Publishing royale to npm
> >> >
> >> > We’ve fixed a lot of things since 0.8.
> >> >
> >> > 0.9 does not need to jump to 1.0. We can have 0.10 (and 0.11…)
> >> >
> >> > I’d really like to streamline the release process so it’s painless
> >> > enough to release every couple/few weeks.
> >> >
> >> > Harbs
> >> >
> >> >> On Nov 12, 2017, at 10:35 AM, Carlos Rovira
> >> >>
> >> > wrote:
> >> >>
> >> >> Hi,
> >> >>
> >> >> about version number, I'll feel more happy with 0.8 than 0.9 My
> >> >>point
> >> >> is that we are getting a first release of Royale and is the same
> >> >>that
> >> >> FlexJS 0.8 but with some new fixes and little things.
> >> >> As well I can see many things to do to reach 1.0, and would be
> >>great
> >> >> to have still an intermediate release 0.9
> >> >>
> >> >>
> >> >>
> >> >> 2017-11-11 22:10 GMT+01:00 OmPrakash Muppirala
> >> > :
> >> >>
> >> >>> On Thu, Nov 9, 2017 at 6:41 PM, Alex Harui
> >> > 
> >> >>> wrote:
> >> >>>
> >>  That might work.  One question:  we want the same bits that
> >>were
> >> >>> published
> >>  as 0.9.0-rc1 to become the final bits where you would do:
> >> 
> >>  npm install -g apache-royale
> >> 
> >>  We aren't supposed to rebuild anything.  For Maven the same
> >>bits
> >>  from staging get copied to Maven central, for Ant/IDE the same
> >> bits
> >>  are moved (not copied) from dist/dev to dist/release.  So is it
> >> > true
> >>  that the RM
> >> >>> can
> >>  publish the final bits by taking the same bits that were once
> >>  published via
> >> >>>
> >> >>>
> >> 
> >>  npm publish --tag 0.9.0-rc1
> >> 
> >>  as the final release?  What does