Front page | perl.perl5.porters |
Postings from August 2000
On Sun, Aug 06, 2000 at 11:47:40PM -0700, Jeffrey Friedl wrote:
> |> Such as? And can you explain them using the better/worse language of
> |> "Combining pieces together", or can they only be understood if you use
> |> terms like "backtracking"?
[...]
> But that's okay, since I don't believe that \1, \2, etc. could be described
> in that respect, either, since exactly what they even mean (what substrings
> they'll match) can change over the course of the regular expression
> application.
*This* is your problem. And I would think that this not-understanding
of yours may be related to your remarks on math. Here is the
conjecture I got reading what you wrote:
one of the difficulties associated to math is that it progresses by
*unlearning* things. "Number is what can be written as
numerator/denominator", "No, one should allow also infinite decimal
fractions", "No, number may have an imaginary part too", "No, the
part before the dot may be infinite too" (these are p-adics, BTW)
etc. etc. etc. Note how things progress by removing restrictions
which turn out to be unnecessary/hindering.
You know too much. In this whole discussion you refuse to *unlearn*
your knowledge of NFA and backtracking. You still refer to the "match
as a process which progresses in time". I consider the discovery that
this is *not nessary* as one of my largest contributions to the Perl
REx arena [*]. [For an alternative - and magnificently failing -
POSIX-related approach see TCL's engine.]
You are still free to think in the terms of backtracking, if this is
more convenient to you. However, any change to the REx engine which
*requires* one to think in terms of backtracking is a significant step
back.
[*] One still needs backtracking to understand how (?{}) works, but I
hope that enhancements to (?{}) will remove this exception ASAP.
P.S. About \1 etc.: given a match, you know which parts of the string
are "encircles" by particular ()'s. In a correct match \1, \2
etc should match *the same string*. As easy as that.
Say, \1(a) should match "aa". If it does not work now, it is a
bug. [The "correct" behaviour would be to reject/warn-on this
at compile time.]
Ilya
Thread Previous
|
Thread Next