> That's a pity I think... But yes, we should at least start with
> documenting
> this. Somewhere. Possibly another lib/README.* file in the same style
> we
> already have a bunch.
>
I can volunteer myself for this.

> We don't currently have a way to build libcurl without the multi
> interface so
> it'd take a little work to get that going. Of course it would further
> decrease
> the footprint...
>
A smaller footprint use would be awfully nice for us. Any idea if memory usage would be reduced as well? That would be even better. Unfortunately, I'd probably break lots of things if I tried to volunteer for this one.

> > Perhaps a happy medium is to have axtls.c notice when an FD is non-
> blocking.
>
> The socket is always set non-blocking within libcurl so that won't help
> us!
>
Really? (I suppose I could just look at the code) I'm surprised the easy interface worked then. If the axTLS code sees "read" return -1, it assumes the connection is lost, even if errno==EAGAIN. Maybe all of my tests were with very fast servers that responded very quickly.

Thanks,
Eric
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette: http://curl.haxx.se/mail/etiquette.html
Received on 2011-01-12