On Tue, 18 Aug 2009, koettermar...@gmx.de wrote:
So ...for now... I decided to be more generic, fix CURLOPT_TIMEOUT, as it will work for most cases.
I agree with you that the timer situation could be improved. I'm not sure I share the exact same pesimistic view but that doesn't really matter.
Patch is attached.
It's a bit too simplistic though. It needs to deduct the already spent time when resetting the timer, as otherwise you'll just push the "deadline" ahead each time.
-- / daniel.haxx.se
