On Thursday 18 of March 2010 18:55:07 Daniel Stenberg wrote:
> On Thu, 18 Mar 2010, Kamil Dudka wrote:
> >> + else if(!nread)
> >> + return CURLE_RECV_ERROR;
> >
> > Looks good to me. Thank you for looking into this!
>
> No, it isn't good enough it seems. Test case 301 fails for me with OpenSSL
> after this. I need to dig around a bit to understand it better as I think
> the fix looks correct...

As mentioned on IRC, we simply need to distinguish among successful read of
zero bytes and an error during read. An attempt of a better patch attached.
The patch for https://bugzilla.redhat.com/565972 , which is actually the
motivation for this one, will follow.

Kamil

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