[antlr-interest] v4 Python target

[Terence Parr]

v4 is a complete rewrite, though it shares a lot of the simpler parts of the runtime like the token objects and some of the stream stuff.

The hard part in this version will be translating the very complicated ATN (augmented transition network) grammar analysis / simulation stuff. It does adaptive analysis on-the-fly. The algorithm easily explodes into exponential whirlpools of death unless all of the tricks used in the Java version are implemented.

When implemented, I wonder how the Python target will do in terms of speed. This version requires an insane amount of data structure walking and manipulation. Garbage collection speed is crucial.

Ter

On Nov 5, 2012, at 12:30 PM, Juancarlo Añez wrote:

> Terr,
> 
> On Mon, Nov 5, 2012 at 2:15 PM, Terence Parr <parrt at cs.usfca.edu> wrote:
> 
>> no one has started yet :)
>> 
> 
> Who did the work for the previous version?
> 
> Have things changed much regarding targets? Would be necessary to start
> from scratch?
> 
> -- 
> Juancarlo *Añez*
> 
> List: http://www.antlr.org/mailman/listinfo/antlr-interest
> Unsubscribe: http://www.antlr.org/mailman/options/antlr-interest/your-email-address