On Thu, 26 Sep 2013, Paras S wrote:

> It looks like libcurl treats the various stages of NTLM challenge response
> as a redirect to the same URL.

It does, at least partly.

> I can think of the following solutions:

> 1. TIMER_STARTSINGLE be set at the end of CURLM_STATE_INIT, rather than the
> start of CURLM_STATE_CONNECT. Will doing so affect wild card transfers?

That would be totally wrong and completely change what the timer measures! The
documentation already says what it should return and we can't set it to start
there.

> 2. A new "CURLOPT_EASY_PERFORM_TIMEOUT" which looks at the transfer
> operation in a holistic view and compares against SessionHandle's
> progress.start for expiration.

A new option won't fix this problem for existing applications though!

I rather believe in:

3) Make sure we set a timer internally that isn't reset between each redirect
or new request and make sure we have the "total timeout" compared against that
timer and not the single-request one.

Are you up to helping us write that fix or a test case that can be used to
verify someone else's fix?

-- 
  / daniel.haxx.se
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette:  http://curl.haxx.se/mail/etiquette.html