httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Manoj Kasichainula <ma...@io.com>
Subject Re: canonical list of i/o layering use cases
Date Fri, 31 Mar 2000 09:06:00 GMT
On Tue, Mar 28, 2000 at 11:37:10AM -0500, Ryan Bloom wrote:
> 
> > I suspect it was at about this point in my thinking that I become
> > happy that this was slipping outside the scope of 2.0 where it
> > could stew longer.
> 
> There is a part of me that wishes this would go away for 2.0 as well.  But
> then I see a bunch of other people that are really counting on this
> functionality soon.

Aiming for getting every single feature that people wanted into 2.0
was one of the things that delayed it for so long. Sometimes, people
have to wait; that's life.

> I would like to get something in and working, even if it isn't the most
> optimal solution.  I think we should get something working, and let real
> people play with it.  If it doesn't work, we can easily back it out and
> try again.  Doesn't that fit in with the whole release early release often
> mind set?

Only if we actually ever get around to a release.

After finally reading all this, I think I'm neutral on the changes
I've seen described, but only as long as we are ready to scrap it all
for the next version.

Real, well-designed I/O layering will end up interacting with
state-machine based HTTP processing, true file system abstraction, and
fundamentally affect most modules that want to take advantage of the
new architecture, so I don't think we can get a stable API for this
into 2.0 anyway.

So, the question is then whether we should even bother with it for
2.0. Not if you want to maintain binary compatibility.


Mime
View raw message