On Thu, 5 Oct 2017 16:39:25 +0200
Moritz Barsnick <barsn...@gmx.net> wrote:
> On Thu, Oct 05, 2017 at 13:41:07 +0200, wm4 wrote:
> > I don't really see where this patch checks whether the host changes?
> > The minimal requirements for reusing the connection would be using
> > the same protocol and the same host.
>
> I *believe* it just tries to reuse the connection and falls back to
> re-opening on failure. On any failure - that could be a cookie issue, a
> timeout, or whatever:
Couldn't find anything that would make ff_http_do_new_request fail if
the host changed, but maybe I didn't look close enough.
> > + ret = ff_http_do_new_request(uc, url);
> > + if (ret < 0) {
> > + ff_format_io_close(c->ctx, pb);
> > + ret = s->io_open(s, pb, url, AVIO_FLAG_READ, &tmp);
> > + }
>
> I wonder what the patch as a whole does if the server doesn't support
> this? The same as above - just fail and reconnect?
I'd think so.
> Does the hls demuxer support HTTP 1.0? This would need a "Connection:
> keep-alive" header in the request (and such a response from the
> server to confirm).
That's definitely added. Not sure what happens if the server doesn't
support it.
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
