[gst-devel] linking pads (was: wingo gstreamer: gstreamer/ gstreamer/gst/)

[Andy Wingo]

Sorry for the delay on this one,

On Wed, 2004-02-25 at 21:27, David Schleef wrote:
> On Wed, Feb 25, 2004 at 06:16:36AM -0800, Andy Wingo wrote:
> > 2004-02-25  Andy Wingo  <wingo at pobox.com>
> > 	* gst/gstpad.c (gst_pad_link_prepare): Re-add
> > 	gst_pad_link_prepare. Please email the list with specific reasons
> > 	for reverting.
> 
> Um, no.  That's not how we work around here.  It's _your_ duty
> to explain why the change is good.

After chatting with Company, and in a discussion with ds that was meant
to be on the list but accidentally wasn't, it seems the complaint is
about playing ping-pong with CVS. I confess and repent my sins! And now
on to the patch itself...

> Note that there is a reason for reverting in the ChangeLog.

It wasn't very verbose, so I didn't understand it. And it wasn't
accompanied by a list mail. My opinions on that are in another mail.

Here's the question: what must be true about the arguments of 
gst_pad_can_link_filtered and gst_pad_link_filtered? Note that presently
this API isn't specified really, although there is existing practice. So
any changes might have to wait for 0.9 at this point.

* gst_pad_can_link_filtered

I think we need to assert that the pads have parents; otherwise there is
surely a programming error somewhere. By assert, I mean use
g_return_val_if_fail.

Their directions are not specified, and they can have incompatible caps.
Perhaps we should assert that they not be connected already? This is
really a question for the autoplug guys, so do speak up...

* gst_pad_link_filtered

I think that we should assert that gst_pad_can_link_filtered. Anything
else would be a programming error. Note that this function can still
fail if calling the linkfuncs doesn't work out.

Any comments? Some of the code in gstpad.c is really nasty, and having
specifications about this things would allow some consolidation and
simplification of some tests.

Regards,
-- 
Andy Wingo <wingo at pobox.com>