Re: yet another pattern-matching library for Lua

[roberto@... (Roberto Ierusalimschy)]

> - I wondered how to emulate {n,m} syntax of REs. It seems that patt^n is 
> close of that. Perhaps for newbies like me, you should provide more 
> examples, something with both bounds, if possible. Classical example in 
> REs is [a-z]{2,6} for TLDs (perhaps more on the upper bound).
> I understand that, per the doc, we have:
> ^n  -> {n,}
> ^-n -> {0,n}
> ^0  -> {0,}  -> *
> ^1  -> {1,}  -> +
> ^-1 -> {0,1} -> ?
> so, to have patt{n,m}, can we write
>   patt^n * patt^(n - m)

This is not correct. Remember that patt^n is actually "at least n". So,
patt^n will always consume as many patt as possible, leaving no one to
patt^(n - m) (which will always fail).


> ? Is it possible to express this without writing the pattern twice? (or 
> is it a non-issue?)

I am not sure whether {n,m} is that common to deserve primitive support.
But it should be easy to write a function to build that kind of patterns.
Something like this (untested!):

function rep (p, n, m)
  local r = lpeg.P""
  for i = 1, n do r = r * p end    -- at least n
  if not m then m = 0
  else m = n - m     -- assuming that n > m
  end
  return r * p^m
end

That's the point of composability. (We could have a library of these
auxiliary functions.)


> - Looking at the CSV example, is (lpeg.P(1) - '"') the same as the (more 
> used elsewhere) (1 - lpeg.P"'")? I suppose the rule here is to have at 
> least one pattern in the expression to trigger metamethods.

Sure. This has nothing to do with Lpeg. It is how Lua chooses metamethods.


> I suppose your library (parser) uses the Lua license, like most others. 

Yes.


-- Roberto