On Thu, 25 Jan 2001, Johan Nilsson wrote:

> > Feel free to add a timeout that does what you thought it did. I'm sure more
> > than you would like such a one. I'd appriciate a patch!
>
> Maybe I'll see what I can do in the future. Right now I don't have a lot
> of time :-)

Who has? ;-)

> > Without the timeout specificly set, it doesn't use any timers at all apart
> > from those the OS itself will use on system/network calls.
>
> What error code is signalled by libcurl if those OS timeouts occur? Is it
> CURLE_COULDNT_CONNECT, CURLE_COULDNT_RESOLVE_HOST or?

It depends on what function that returns failure. Both the above return codes
are valid return codes.

> Are the same codes sent even though some of the file has already been
> transfered, i.e. the transfer lost its "connection" to the server.

No. The code CURLE_COULDNT_CONNECT is for failed connect() and
CURLE_COULDNT_RESOLVE_HOST is for when the host name didn't resolve. When the
connection to the server is lost it may return CURLE_PARTIAL_FILE. It isn't
always possible to detect the difference between a "lost connection" and
end-of-file and then if so it may return CURLE_OK.

-- 
  Daniel Stenberg -- curl project maintainer -- http://curl.haxx.se/
_______________________________________________
Curl-library mailing list
Curl-library_at_lists.sourceforge.net
http://lists.sourceforge.net/lists/listinfo/curl-library