On Thu, 22 Jun 2006, Dan Fandrich wrote:

> The more I think about this, the more I realize this kind of scheduling
> could get pretty complicated. I'll bet there's research published out there
> on how best to implement HTTP pipelining done by people who've put a lot
> more thought into it than I have.

Well, the HTTP RFCs gives us a hint as it says we should never connect more
than twice to the same HTTP server as I believe they consider the case where
the first connection is slow and you use the other one to get smaller/faster
files.

That said, your mail did indeed show that this can become a rather complex
beast unless we take caution.

But - I believe in starting doing things simple, so I believe this kind of
advanced system with multiple pipes and priorities etc will be subject for
inclusion later. We should first make pipelining work with a more simplistic
plain one-pipe-per-server-until-full-then-start-a-new apprach.

-- 
  Commercial curl and libcurl Technical Support: http://haxx.se/curl.html