Bug #61478 for IO-Async: Support for KQueue

Mon Sep 20 07:07:24 2010 squeek [...] cpan.org - Ticket created

Subject:

Support for KQueue

IO::Async doesn't support KQueue. Could this be supported please, even if it just does a wrap of IO::KQueue.

Mon Sep 20 07:35:54 2010 leonerd-cpan [...] leonerd.org.uk - Correspondence added

On Mon Sep 20 07:07:24 2010, SQUEEK wrote: Show quoted text

> IO::Async doesn't support KQueue. Could this be supported please, even > if it just does a wrap of IO::KQueue.

An excellent idea, and one hinted at by the TODO notes: http://search.cpan.org/~pevans/IO-Async-0.29/lib/IO/Async.pm#TODO I don't have easy access to a BSD box, short of going the usual qemu- style approach, so I'm not sure how easily I could write such. Having had a look over the IO::KQueue interface, and the kqueue(2) manpage, I'd say it looks the same sort of idea as Linux's epoll; I'd reckon therefore you could make some progress by taking IO::Async::Loop::Epoll and adjusting it to use IO::KQueue instead. That should give you a good start at least. Additionally it supports things like signals, child processes, and filesystem watches; so those could be added on additionally. -- Paul Evans

Mon Sep 20 07:35:55 2010 The RT System itself - Status changed from 'new' to 'open'

Mon Sep 20 09:15:07 2010 leonerd-cpan [...] leonerd.org.uk - Correspondence added

This may also depend in some way on a fix for: https://rt.cpan.org/Ticket/Display.html?id=61481 -- Paul Evans

Wed Jan 05 07:36:20 2011 leonerd-cpan [...] leonerd.org.uk - Correspondence added

An initial version of IO::Async::Loop::KQueue is now on CPAN. This can be closed, and any further issues raised there instead. -- Paul Evans

Wed Jan 05 07:36:21 2011 leonerd-cpan [...] leonerd.org.uk - Status changed from 'open' to 'resolved'