[gst-devel] Re: [gst-cvs] wingo gstreamer: gstreamer/ gstreamer/gst/ gstreamer/testsuite/ gstreamer/testsuite/ghostpads/

[in7y118 at public.uni-hamburg.de]

Quoting David Schleef <ds at schleef.org>:

> It doesn't seem necessary to deprecate gst_value_get_caps(), etc.
> We should also use G_SIGNAL_TYPE_STATIC_SCOPE whenever it's useful,
> which means essentially all the time with caps and structures.
> 
Well, it's completely redundant with (GstCaps *) g_value_get_boxed, so it 
seemed natural to get rid of it.

> Also, the boxed type registration can't take gst_caps_copy(), since
> gst_caps_copy() barfs on a NULL.
> 
According to glib source, a GBoxed may not be NULL. At least g_boxed_copy 
barfs on NULL, too. Which might have been one of the reasons why it didn't 
work with the old caps system.

> Please note that there is only one test in testsuite/ that checks
> correct behavior of a signal.  So relying on the testsuite as your
> only means of testing is probably not wise.
> 
I checked gst-launch -v and gst-typefind, too and both didn't complain.
I don't think there's more code that uses caps.

Benjamin