I wonder what chance there is of getting a "capture to named vars" contruct added?

I mentioned it once on p5p, but to a deafenaning silence. It would be lovely though wouldn't it? Although I can see how there might be serious questions about how it should work. What should happen if there are identical named sections? Where should the results be stored? Possibly %+ or something? Also I can see some of the p5p saying "Perl is not going to use the dotNet syntax."

:-)

I think my vision of the way it would work is that a named capture would get stored exactly the same as if

(...)(?{ $var = $^N })
[download]

That is to say, if a lexical named $var was in scope, it would get the captured string, else a global of that name would. If two blocks named the same var, the second would override the first just as with normal assignment.

I've no idea what the dotNet syntax is (or even that it had such), but Enlil and I had a discusion about it somewhere a few months ago. Unfortunately, it was tucked down in teh bowels of a thread with an unrelated name, so I can't find it right now.

Off the top of my head, I think that

(?$var:...)
[download]

would work. I don't think it would conflict with anything else?

The one distinction I would make is that if an array was given rather than a scalar,

(?@array:...){1,10}
[download]

then the captured string would be pushed onto the array. This would allow for captures with repeatition specifiers to do something sensible.

Perhaps harder to think through is what happens when a regex backtracks through a capture. undef the $var and pop the @array maybe?

---

Examine what is said, not who speaks.
"Efficiency is intelligent laziness." -David Dunham
"Think for yourself!" - Abigail
Hooray!

The dotNet syntax is provided (with problems associated) by Regex::Fields. It looks like

(?<NAME>PATTERN)
[download]

In dotNet the results are available through the regex object itself. In Perl I would assume that a special hash would be created for the results (In R::F its %{&}). I also suspect that a leftmost-outermost wins rule would be a reasonably useful rule if only one possibility was allowed, otherwise perhaps a hash of arrays would be cool. R::F supports also binding to implicit lexicals.

Incidentally I understand from another post here that this module causes problems in that it makes global changes to how the regexes are handles. I cant say if this is true however.



---
demerphq



First they ignore you, then they laugh at you, then they fight you, then you win.
-- Gandhi