On Wed, 12 Mar 2008, Daniel Stenberg wrote:

> #1 - There's a bug in that the timeout libcurl should use to wait for a 100
> continue before it ignores that and continues anyway, is always ignored.
> This will of course make libcurl continue too often with the sending.
>
> It would be easy to assume this is what you experience. (commit to fix this
> issue will be done soon)

The fix I had in mind turned out not so good nor easy so I'll have to work a
bit more on this before I commit anything. The entire current 100-continue
code is a bit confusing so I think I'll take this opportunity to clear it up.

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