Build failed in Jenkins: royale-asjs_MXTests #23

[apacheroyaleci]

See 


--
[...truncated 1.96 MB...]
[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

[mxmlc] using source file: 

Re: NextGenAS tutorials on Royale website

[Josh Tynjala]

dts2as hasn't been updated with any syntax changes since TypeScript 2.4.1. 
There were still edge cases where it failed too. It probably won't be very 
compatible with modern type definitions.

Keeping dts2as up to date with changes to TypeScript was a major pain. It 
wouldn't have been so bad if type definitions were kept pretty simple, but I 
found that the TypeScript community very quickly embraced new language features 
in their type definitions, and this caused dts2as to break frequently.

In short, I wouldn't recommend anyone try to do use or update dts2as (unless 
you have a lot of time and patience to dedicate to the process).

On the other hand, I wouldn't expect to see the format of Closure Compiler's 
externs changing very much going forward. Embracing externc seems like a better 
way to go for making type definitions for third-party libraries in Royale. It 
doesn't have a major community-driven repository of common libraries, like 
DefinitelyTyped exists for TypeScript, but that doesn't meant that someone in 
the Royale community couldn't get the momentum started.

- Josh

On 2018/10/04 09:13:54, Carlos Rovira  wrote: 
> Hi Harbs,
> 
> El mié., 3 oct. 2018 a las 11:54, Harbs () escribió:
> 
> > I agree that it would be great to make it easier to use Node. There are
> > some sticking points currently. To answer your points:
> >
> > 1. Yes. It’s kind of similar in that it declares dependencies and
> > optionally build tasks. Although there’s not necessarily a build step at
> > all in Node.js. It’s worth pointing out that there are some new
> > developments on the Node dependency front (besides Yarn and Turbo).[1]
> >
> >
> thanks, that's what I expected
> 
> 
> 
> > 2a) This is a problem. You’d either used untyped variables or create
> > Typedefs. I’d really like to get support for Typescript definition files
> > directly some day. Most popular nom packages have Typescript definition
> > files.
> >
> 
> So maybe the solution would be the use dts2as:
> 
> https://joshblog.net/2015/introducing-nextgen-actionscript/
> 
> there's another tutorial from josh that we can publish.
> 
> A side from this, I still think that adding TS at the same level than AS3
> would be a great movement for Royale, Although the task will be not
> trivial, probably most of the needed pieces are out there to use (TS Antlr
> grammar, lever, and so on), we just need someone with the interest,
> time and dedication to make this real...
> 
> 
> >
> > 2b) Linking is done using require(). This “just works” as long as the node
> > modules are installed.
> 
> 
> ...and I assume if you used package.json to download the dependencies, the
> modules should be installed.
> 
> 
> > Some tooling for debugging  and releasing would probably be helpful though.
> >
> 
> Ok, but I expect Debugging to be the same we already have in Royale
> (although in parallel, I think debugging still has some issues to solve,
> but not related to this).
> 
> Thanks
> 
> Carlos
> 
> 
> 
> >
> > HTH,
> > Harbs
> >
> > [1]
> > https://blog.npmjs.org/post/178027064160/next-generation-package-management
> >
> > > On Oct 3, 2018, at 12:06 PM, Carlos Rovira 
> > wrote:
> > >
> > > Ok! publishing! :)
> > >
> > > While I publish I want to do share some thoughts I had today with friends
> > > that use Node. They say that we need to make it as easy as possible to
> > work
> > > with Node since today is what front end devs use. So it's a must for us
> > to
> > > succeed in the current JS world. So tow things:
> > >
> > > 1.- package.json :  I assume that this is like build.xml or pom.xml. And
> > > people expect to add JS libraries and those ones could bring other
> > > dependencies as well. For me this is in the Node world and in Royale we
> > > don't need to do more.
> > >
> > > 2.- Once we have a node dependency downloaded in our local repo, we need
> > to
> > > be able to use easily. Two things comes to my mind here:
> > >   a) how we use the code directly in AS3 for a node library
> > >   b) how we link the library and it's dependencies quickly (until now
> > > we link via inject, from a CDN, but that's not what a node dev expect)
> > >
> > > hope you could share thoughts about this process to discover if we cover
> > it
> > > completely or we have black spots.
> > >
> > > thanks
> > >
> > >
> > >
> > >
> > > El mié., 3 oct. 2018 a las 10:45, Harbs ()
> > escribió:
> > >
> > >> I’m done. I think it’s good to go.
> > >>
> > >> Harbs
> > >>
> > >>> On Oct 3, 2018, at 11:37 AM, Harbs  wrote:
> > >>>
> > >>> I’m making a few small changes/corrections. I’ll be done in a few
> > >> minutes.
> > >>>
> >  On Oct 3, 2018, at 11:24 AM, Carlos Rovira 
> > >> wrote:
> > 
> >  Hi, I think a catch with all revisions. Please let me know if it's ok
> > >> to go:
> > 
> > 
> > >>
> > https://royale.codeoscopic.com/hello-node-how-to-transpile-actionscript-for-node-js/
> > 
> >  about Royale 0.8 to 0.9.4 change, I 'm not sure is what we want, 

Check is something is Class or instance in Royale

[Carlos Rovira]

Hi,

in js, check if something is a class or is an instance is like this [1]

typeof myFunc === 'function'

In Royale since we are using AS3, doing

if(something is Class)
do this
else
   do that

should work either for SWF and for JS, but seems is not the case.
I must make two codes COMPILE::SWF (with "is Class") and COMPILE::JS (with
typeof ... ==='function')

Could we have this abstracted by compiler?

Thanks

[1]
https://stackoverflow.com/questions/405164/best-method-of-testing-for-a-function-in-javascript

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

Jenkins build is back to normal : royale-typedefs #772

[apacheroyaleci]

See 

Re: [royale-asjs] branch develop updated: clean compile of 'full' Flex mx:RemoteObject

[Carlos Rovira]

Thanks for the explanation Alex, much more clear now :)

El vie., 5 oct. 2018 a las 0:32, Alex Harui ()
escribió:

> It does not work yet.  That's what I'm working on now.
>
> Most of the simpler, fundamental UI widgets have lots of flash
> dependencies so it was better to produce a new platform-independent
> implementation in Basic and re-use those beads in the migration set.  As we
> find out what APIs folks need that Basic doesn't have, we will be busy
> creating new implementation code in Basic and using it in emulations.  The
> emulations of these components copy in as little code from Flex as possible
> to avoid bringing in Flash-dependent code, then we fill in the
> implementations as needed.
>
> But some large chunks of code, like container layout actually had
> relatively few Flash dependencies so it and aren't very browser dependent
> either.  Container layout does not use CSS since percentWidth/Height in
> Flex did not conform to CSS width %.  So I was able to copy in lots of code
> from Flex, change a few places and get it to run in I think fewer days than
> starting from as little code as possible and working upward.
>
> My bet is that the same is true for RemoteObject.  In copying the code and
> making the required changes, I think it will really boil down to an
> emulation of NetConnection which will call into the existing Royale AMF
> code.  And if I'm right, lots of other RemoteObject features might just
> start to work.  Consumer, Producer, ChannelSets, Operations, AsyncToken and
> more were all copied over and now compile and don't really depend on Flash,
> just Result and Fault events from NetConnection.
>
> I think the same is true for ResourceManager, although I need the compiler
> to handle bundles in some way for JS.
>
> And I think the same is true for Charts.  There are Basic Charts, but not
> as many as in Flex and instead of working from little code upward for
> Charts, I want to try copying all of Flex Charts.  I think Charts will boil
> down to drawing on Graphics.
>
> Hope that makes it more clear,
> -Alex
>
> On 10/4/18, 3:19 PM, "Carlos Rovira"  wrote:
>
> Hi Alex,
>
> full compile of this classes means as well that works? This kind of
> code
> since is about 80% not flash dependent, could be working? Just curious
>
> thanks
>
> Carlos
>
>
>
> El jue., 4 oct. 2018 a las 20:51,  escribió:
>
> > This is an automated email from the ASF dual-hosted git repository.
> >
> > aharui pushed a commit to branch develop
> > in repository
> https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgitbox.apache.org%2Frepos%2Fasf%2Froyale-asjs.gitdata=02%7C01%7Caharui%40adobe.com%7Cf808264cf16a47daae8708d62a476d95%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636742883608520259sdata=sHQ3FYHCJbXwLPCk6nPg3uq4HfYYyVBV89a2CRREMoE%3Dreserved=0
> >
> >
> > The following commit(s) were added to refs/heads/develop by this
> push:
> >  new 16e9336  clean compile of 'full' Flex mx:RemoteObject
> > 16e9336 is described below
> >
> > commit 16e9336a55cfaca4c042a9e33949d1eef32b309b
> > Author: Alex Harui 
> > AuthorDate: Thu Oct 4 11:51:17 2018 -0700
> >
> > clean compile of 'full' Flex mx:RemoteObject
> > ---
> >  .../src/main/resources/mx-royale-manifest.xml  |2 +
> >  .../MXRoyale/src/main/royale/MXRoyaleClasses.as|1 -
> >  .../LogEventLevel.as => core/IMXMLObject.as}   |   34 +-
> >  .../IllegalOperationError.as}  |   68 +-
> >  .../events/{IOErrorEvent.as => AsyncErrorEvent.as} |   12 +-
> >  .../mx/events/{IOErrorEvent.as => ErrorEvent.as}   |   16 +-
> >  .../src/main/royale/mx/events/IOErrorEvent.as  |2 +-
> >  .../events/{IOErrorEvent.as => NetStatusEvent.as}  |   14 +-
> >  .../main/royale/mx/events/SecurityErrorEvent.as|2 +-
> >  .../MXRoyale/src/main/royale/mx/logging/ILogger.as |  303 +-
> >  .../src/main/royale/mx/logging/ILoggingTarget.as   |  166 +
> >  .../MXRoyale/src/main/royale/mx/logging/Log.as |   96 +-
> >  .../src/main/royale/mx/logging/LogEvent.as |  230 ++
> >  .../src/main/royale/mx/logging/LogEventLevel.as|   64 +-
> >  .../src/main/royale/mx/logging/LogLogger.as|  258 ++
> >  .../InvalidCategoryError.as}   |   58 +-
> >  .../main/royale/mx/messaging/AbstractConsumer.as   |  997 +
> >  .../main/royale/mx/messaging/AbstractProducer.as   |  760 
> >  .../src/main/royale/mx/messaging/Channel.as| 1743 
> >  .../src/main/royale/mx/messaging/ChannelSet.as | 4312
> > ++--
> >  .../src/main/royale/mx/messaging/Consumer.as   |  273 ++
> >  .../mx/messaging/ConsumerMessageDispatcher.as  |  326 ++
> >  .../src/main/royale/mx/messaging/MessageAgent.as   | 1391 +++
> >  .../main/royale/mx/messaging/MessageResponder.as   |  386 ++
>  

Re: Migrating Enterprise Flex Application

[Alex Harui]

Sorry, I thought we had a -js-source-path option.  

I think package paths in SWCs did work in Flex.  I haven't tried it in Royale, 
so I'm wondering if you've seen it work and it isn't working in this one case.

-Alex

On 10/4/18, 10:13 PM, "chembali"  wrote:

I am trying the package paths for swc for the first time. I use it quite a
lot in the Flex world. A bit confused about the -js-source-path option. I
don't see that option in mxmlc. Am I missing something?



--
Sent from: 
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fapache-royale-development.20373.n8.nabble.com%2Fdata=02%7C01%7Caharui%40adobe.com%7Cae92c5cb03f84c861e5208d62a8155ec%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636743132325036935sdata=ZrRDePbFK5U05QC3cQxYMUUjBdsFEy5tSiqwSxcmDPg%3Dreserved=0

Jenkins build is back to normal : royale-asjs_MXTests #20

[apacheroyaleci]

See