Re: [FWP] innermost first parsing

[abigail@foad.org]

On Wed, Aug 02, 2000 at 12:27:58PM -0400, Ronald J Kimball wrote:
> On Wed, Aug 02, 2000 at 08:54:55AM -0700, Randal L. Schwartz wrote:
> > >>>>> "Jeff" == Jeff Pinyan <jeffp@crusoe.net> writes:
> > 
> > Jeff>   ($first) = $text =~ m{
> > Jeff>     BEGIN
> > Jeff>     (
> > Jeff>       [^BE]*  # 'B' and 'E' are first chars of tags
> > Jeff>       (?:
> > Jeff>         (?:
> > Jeff>           B+ (?! EGIN )  # match /B+/ if NOT 'BEGIN'
> 
> > I think yours fails on BEGIN ... BBBBBEGIN ... END ... END
> > as well, since you grab the multiple B's but only check the first
> > B for BEGIN'ness. :)
> 
> I don't think that's right.  B+(?!EGIN) checks each B for BEGIN-ness.
> In BBBBEGIN, it will match the first three Bs and reject the fourth, which
> is followed by EGIN.

No, it checks to see if the last B is followed by EGIN. It will backtrack
to a previous B only if it is.

But the solution doesn't generalize. It relies on the fact that the second
characters of the tokens are different from the first.  There would be a
problem if the you tried it with 'BBEGIN' and 'EEND' instead of 'BEGIN'
and 'END'.



Abigail

==== Want to unsubscribe from Fun With Perl?  Well, if you insist...
==== Send email to <fwp-request@technofile.org> with message _body_
====   unsubscribe