[metaas-dev] No viable alternative error

[David Holroyd]

On Mon, Feb 19, 2007 at 01:17:38AM +0100, kiroukou wrote:
> Great thanks :)
> 
> But another file seems to be hard to parse, see attached :) (it is an  
> implementation of Singleton pattern in AS3 in case you wonder why  
> this synthax ;) )

Ahhh; precomputed trig-tables.  I remember doing that in AMOS Basic on
my old Amiga 1200.  Happy days :)

Anyhow, the underlying ActionScript grammar does support files
structured like the one you provided, but as you've probably found,
metaas provides no API to actually expose things like package-level
elements that aren't class/interface definitions, or classes+imports
outside of the file's package-block.

Adding these things in will require a bit of a rethink of some of the
high level metaas APIs (ASCompilationUnit will need lots of extensions),
so this will take a bit longer to fix than the other problems you've
discovered.

Also, my road-map for the next release was to involve work on improved
JavaDoc support, which I've already started implementing.

So, can you live without support for this file for a bit, or should I
re-prioritise?  Give me an idea of how critical handling for files
structured in this way is to your project.


> Thanks for your help, and I hope my problems help you to make metaas  
> more robust :)

You are making a huge difference to metaas!  Thank you for your help and
suggestions! :)


ta,
dave