Re: recdescent 2.0 and Perl 6

Aamer Akhter wrote:

> ok, this may be a really dumb question, it's not meant to be a political 
> one, just to help me plan.
> 
> We're starting a project that includes a parser that would convert a 
> huge swath of multi-device output (output of explicit commands) to an 
> XML tokenized form. More than likely, perl would run in an embedded from 
> from tcl (this is the reality we're dealing with-- but we're hoping 
> parrot can save us ;-))
> 
> Some of this output is potentially very large (although repeating). I've 
> written parse::recdescent grammars for some of this output in the past 
> but speed was the biggest issue. Has anybody done speed comparisons 
> between Perl6::Rules and prd?
> 
> I realize that perl6 is still evolving, but at least we can change our 
> copy of Perl6::Rules on our schedule, if the benefits outweigh the 
> shifting sands... 

Do NOT use Perl6::Rules for production code. The module is only at release 
0.03, which is *far* too early to rely on it.

Apart from the fact that the rule syntax and matching semantics for Perl 6 are 
still changing occasionally, the module still has a raft of known bugs and 
limitations. Many of these are in fact bugs and limitations in the underlying 
Perl 5 regex engine (which has apparently never been pushed quite as far as 
I'm pushing it now). Some of the most serious underlying bugs will probably 
not be fixed in the foreseeable future -- perhaps not before perl 5.10 is 
released, maybe not even then.

All of which means Perl6::Rules is NOT production-ready, and unlikely to be 
production-ready in any reasonable time-frame.

I am currently a little preoccupied with the upcoming conference season, and 
with working on Exegesis 12. At the same time I *am* reassessing the 
Parse::RecDescent and Perl6::Rules modules to determine how best to proceed.
Options include:

	* Working (with P5P) to resolve the major bugs in Perl6::Rules

	* Implementing an entirely new and separate regex matching engine
	  to sit under Perl6::Rules

	* Proceeding with my original plans to improve Parse::RecDescent
           performance

	* Proceeding with my later plans to create a new, faster
	  Parse::FastDescent module

	* Giving up and moving to Tibet ;-)

However, I will not be making any decisions on this until at least September,
when I return from my forthcoming speaking tour.

In the meantime my best advice remains the same: if you can't live with the 
performance of Parse::RecDescent, and can't tweak your grammar to improve that 
performance sufficiently, consider using Parse::Yapp or perl-byacc instead.

Sorry not to be of more help,

Damian