httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "David Reid" <dr...@jetnet.co.uk>
Subject Re: Filter I/O take 3.
Date Fri, 16 Jun 2000 21:50:56 GMT
> Apache moving to an async model simply looks impossible, without a
**HUGE**
> overhaul in every facet of its operation. At that point, it just won't be
> Apache. I could see a case where a module can say "I can deal with async;
> here is my API to call for read/write availability." To support that,
> though, would still require a huge revamp on the Apache-side of the
equation.
>

And...

> Async I/O would be nice, but it just won't be happening in Apache any time
> soon or far. I really don't think we can start saying "well, if that
doesn't
> support async, then it is out." By that rule, we could hardly make any
> changes :-)

As Dean keeps pointing out, asynch should be a big win, but it will need, as
jeff pointed out, carefully planning and forethought.

Think back to the origins of the A2 tree.  Dean took some time and ripped
apart 1.3 to create the MPM's.  From there APR was added and off we went to
roughly where we are now.

I'd suggest that once we get A2 out the door that maybe someone should do
the same radical reshape for asynch.  We ran with a small group looking at
A2 while most people carried on with 1.3 so why not do it again?  I just
can't see any other way of doing such a large revamp, so I agree that
support for asynch can't be used as a valid reason for vetos at present.
Sorry Bill :(

I seem to remember this being discussed before and we had a "rough" roadmap
to this effect didn't we?

> I'd love to be educated on how to change Apache to an async model. I just
> don't see it right now.

That would be good for all of us.  Maybe someone should either write a piece
on how it could be done or point us all at a document that gives that
information?  But maybe we should also wait until we get A2 out the door?

david


Mime
View raw message