[Webkit-unassigned] [Bug 34317] [GStreamer] Add a WebKit HTTP/HTTPS source that uses WebKit's network infrastructure

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=34317





--- Comment #19 from Sebastian Dröge <slomo at circular-chaos.org>  2010-02-02 01:10:50 PST ---
(In reply to comment #18)
> (From update of attachment 47766 [details])
> 
> > +	-gstapp-0.10
> 
> Are you sure, you don't want -lgstapp-0.10 here?

Yes, that's what I wanted. I wonder why it worked nonetheless...

> > +    priv->appsrc = GST_APP_SRC(gst_element_factory_make("appsrc", 0));
> > +    if (!priv->appsrc) {
> > +        GST_ERROR_OBJECT(src, "Failed to create giostreamsrc");
> > +        return;
> > +    }
> 
> > +static void webkit_web_src_stop(WebkitWebSrc* src, gboolean resetRequestedOffset)
> > +{
> 
> > +    gst_app_src_set_caps(priv->appsrc, 0);
> 
> Does this mix well? Will it crash or just print a critical warning? Or will
> this never happen because the src_change_state will report a missing plugin and
> then stop is never called?

Good catch, actually _stop() is called from the instance initialization
function (webkit_web_src_init) and there appsrc could be NULL.

> The next question is can the stop method be called multiple times? Or is there
> any gurantee that GStreamer will not do this?

It can be called multiple times and probably will, all state changes from
PAUSED to READY will go through the stop method. That's why I check if the
things to be freed are not NULL before freeing and set them to NULL afterwards

Anyway, I'll attach a new patch shortly with the above issues fixed. Thanks for
reviewing :)

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.