flex-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Josh Tynjala <joshtynj...@gmail.com>
Subject Re: [FlexJS] Framework using externs (was: Setup Error)
Date Fri, 18 Sep 2015 16:02:18 GMT
I don't think that there's a currently a way to explicitly specify the
global namespace in a fully-qualified class name. I suspect it rarely came
up in the Flash runtime because it has so few global classes. This would be
very useful with the very large number of classes in the JavaScript global
namespace, though. I hope you can figure out how to add it!

- Josh

On Fri, Sep 18, 2015 at 8:04 AM, Alex Harui <aharui@adobe.com> wrote:

> On 9/18/15, 6:56 AM, "Frédéric THOMAS" <webdoublefx@hotmail.com> wrote:
> >> I’ve been trying to get your GCL.swc to build from Ant without requiring
> >> an application.home systemProperty.  I ran into an issue with missing
> >> builtins like Object.  It could be that IJ injects these builtins no
> >> matter what.  From Ant I need to have GCL depend on js.swc, but then I
> >>get
> >> other conflicts that I am trying to sort through in the compiler.
> >
> >That's my conf., basicaly, I removed all the libraries from the SDK to
> >build it and don't depend on js.swc either.
> >
> >http://pastebin.com/q7X3K38E
> Well, some SWC has to provide the builtins like Object, Array, etc.  My
> guess is that your config is depending on a playerglobal or airglobal, and
> that does work for me, but I was thinking that long-term we want GCL to
> depend on js.swc and use its Object and Event definitions in there because
> that is the true set of classes available to GCL developer at runtime.
> For example, in one of API right now I saw use of the Class builtin, which
> doesn’t exist in JS.
> Using js.swc does cause other errors.  Both GCL and js have Event and
> EventTarget definitions.  I’ve made a tweak to the compiler to not emit an
> AmbiguousDefinition error if it finds a conflict between        global
> namespace
> definitions and a definition in the same package as the class being
> compiled.  It will always use the one in the same package.  I think we’ll
> probably add more logic to ignore conflicts between global definitions and
> a definition that has been explicitly imported.  Right now, though, I’m on
> a little side-trip to see if there is an already existing way to specify
> the global namespace or if there is an easy way to add one.
> -Alex

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message