State of IOCP work

Mark Hammond mhammond at skippinet.com.au
Mon Nov 29 13:02:36 CET 2010


On 29/11/2010 7:10 PM, Marc Lehmann wrote:
> On Mon, Nov 29, 2010 at 02:01:29PM +1100, Mark Hammond<mhammond at skippinet.com.au>  wrote:
>> * What is that state of the code wrapped around EV_USE_IOCP?  Is it
>> abandonded, experimental, progressing slowly etc?
>
> It's vestigates of an old experiment to see if it could be done in a
> different way.

Thanks for the clarification.

>> educational purposes?  It would be nice to see the direction it is
>> taking so informed decisions can be made about how node.js might
>> support async IO on Windows (and might even lead to contributions on
>> this effort ;)
>
> async I/O is not something libev does or ever planned to, nor does it make
> any sense with libev or with sockets in general, as libev doesn't do any
> I/O for you. the experiment existed solely to see if libev could use iocp
> as a backend, but iocps are not featureful enough to support I/O events,
> so can't be done.

Understood - but to be clear, I wasn't suggesting libev performs the IO. 
  While neither libev or IOCP actually perform IO, they do provide a key 
mechanism enabling async IO to work in practice - depending on your 
definition of "work", of course... :)

Thanks,

Mark



More information about the libev mailing list