Yes, it waits for the connections to finish. This is because it will open a new set of connections when they are done, or perhaps open a new connection as each one finishes.

But what is necessary when reusing a connection? Do I need to change the URL and then call perform again? Could I just call perform if I wanted the same URL?

...Daniel

There may be a problem with the python bindings

--- On Fri, 7/8/11, Daniel Stenberg <daniel_at_haxx.se> wrote:

> From: Daniel Stenberg <daniel_at_haxx.se>
> Subject: Re: Curl object reuse
> To: dhavey_at_yahoo.com, "curl with python" <curl-and-python_at_cool.haxx.se>
> Date: Friday, July 8, 2011, 5:36 AM
> On Thu, 7 Jul 2011, Daniel Havey
> wrote:
>
> [disclaimer: I'm no good with python, but I know libcurl so
> I'll respond anyway]
>
> > I would like persistent connections.  I'm trying
> to reuse the handles in my multi curl.
>
> When using the libcurl multi interface, the connection pool
> is kept within the multi handle so all new handles added to
> the multi handle can potentially re-use a connection that a
> previous handle used.
>
> > I have some code hacked from retriever-multi. 
> Basically I just do the initialisation then call a
> subroutine that runs perform and waits for the results.
>
> It seems to wait for all transfers to complete from what I
> could see.
>
>
> --
> / daniel.haxx.se
>

_______________________________________________
http://cool.haxx.se/cgi-bin/mailman/listinfo/curl-and-python
Received on 2011-07-08