Re: Re: RE_CONCAT: question about RE concatenation and subpattern matching

[David Korn <yyy@xxxxxxxxxxxxxxxx>]

I think that the main problem stems from the definition of subexpression.

In section 9.3.6 on page 172 of the new standard, item 2 it
defines a subexpression as the characters between \( and \) for a BRE
( for ERE it would be (...) ).


For ((week|wee)(night|knights))(s*)

The subexpressions are
1.      ((week|wee)(night|knights))
2.      (week|wee)
3.      (night|knights)
4.      (s*)
in that order.  What the RE group decided was that the lefmost
longest rule would be applied first to the outer level
subexpressons, 1 and 4, so that for this case
        \1 would be weekknights
        \4 empty
Then, \2 and \3 would be computed by looking at all the ways
that
        ((week|wee)(night|knights))
could match weeknights and choosing the leftmost longest of (week|wee).
However, the only match is
        wee knights
so that \2 is wee and \3 is knights.

I hope that the will clarify the intent at least.

David Korn
research!dgk
yyy@xxxxxxxxxxxxxxxx