I've learned that NQP relies on Capture not providing "array" to distinguish between multiple results and single results from a regex match. For this reason, Capture should continue to not provide "array" for now, at least.
On Tue, Jun 22, 2010 at 6:43 PM, Tyler Curtis <[email protected]> wrote: > The Capture PMC consists of an array and a hash, to which it delegates > array-related and hash-related vtables. Is there any reason why it > should not "provide" the "array" and "hash" roles? If not, I'm going > to making such a change in my GSoC branch(in order to facilitate some > changes that will help generalize the PAST::* libraries I've been > developing so that bacek's PIR compiler can use them). There are > other ways I could implement it, but the most straightforward would > require Capture to provide both "array" and "hash". > > -- > Tyler Curtis > -- Tyler Curtis _______________________________________________ http://lists.parrot.org/mailman/listinfo/parrot-dev
