Re: yet another pattern-matching library for Lua

[Brian Hagerty <Brian.Hagerty@...>]

Mike,

you've posed a good LPEG quiz about parsing Lua long strings and 
long comments. -- I'll work on it.

However, my first thought is that since PEGs (or grammars in 
general) are supersets of RegExps, it should not be a problem 
(in principle) to represent RegExps in a PEG.  The question is 
whether Lua's LPEG version 0.2 is fully up to PEG & RegExp 
functionality yet.

In particular, whereas CFGs are expressed (typically) in Bakus 
Naur Form (BNF), PEGs extend the BNF-like semantics of grammar 
definition languages by *adding* regular expressions as 
first-class pattern constructs directly to the BNF rules 
themselves.  Thus PEGs, by definition, include RegExp 
facilities, as well as capture facilities (which provide 
additional/local context and virtual backtracking not found in 
conventional CFGs, though are often added as auxiliary functions 
in CFG parsers).

Whether Lua's LPEG has arrived "there" yet in terms of fully 
incorporating RegExps into its Grammar/Rule syntax is another 
matter.  LPEG's lpeg.P(pattern), R(range), S(set), T(table), 
C(capture), and I(index) constructs are the starting point for 
RegExp construction.  The other semantics provided by LPEG then 
extend the RegExps and provide the rich PEG grammar behavior.

I think I'll wait a little while longer, however, until LPEG is 
closer to 0.9 than 0.2 to test the limits of its PEG & RegExp 
expressivity!

Cheers,

// Brian


Mike Pall wrote:
> Maybe it's also interesting to show what cannot be expressed with
> PEGs. Apart from the textbook examples (see the Wikipedia links),
> I think I found one with practical relevance:
>
> I found no closed PEG for properly parsing Lua 5.1 long strings
> and long comments. This is easy for any finite subset:
>
>   "[==[" * (P(1)-"]==]")^0 * "]==]"