Yehuda Katz
Chief Technologist | Strobe
(ph) 718.877.1325


On Mon, Aug 1, 2011 at 5:08 PM, Eric Wong <normalperson / yhbt.net> wrote:

> Yehuda Katz <wycats / gmail.com> wrote:
> > On Mon, Aug 1, 2011 at 4:07 PM, Eric Wong <normalperson / yhbt.net> wrote:
> > > Yehuda Katz <wycats / gmail.com> wrote:
> > > > * :read_would_block
> > > > * :write_would_block
> > > > * :eof
> > >
> > > Why :eof instead of nil?  IO#read already returns nil on EOF
> >
> > Interesting. I like this and will update the patch.
>
> I'm also curious about :*_would_block vs the :wait_*able names used by
> kgio.  I picked :wait_*able for kgio since the IO::Wait*able name is
> already used by Ruby 1.9.2+ exceptions, so it's less of a jump.
>

I picked *_would_block based on the C EWOULDBLOCK errno, but I'd be happy to
change it to match the Ruby 1.9.x exception names. I personally am a bit
confused by those names (not sure what the WaitReadable adjective would
mean), but it *is* the name, so I'll gladly go with it.


> I feel less strongly about these than nil for EOF, though it might help
> with porting any kgio-using apps to newer code.
>

That's good enough for me.


>
> --
> Eric Wong
>
>