[gst-devel] GDP updates and questions

[Thomas Vander Stichele]

On Tue, 2006-05-09 at 13:03 -0700, David Schleef wrote:
> On Tue, May 09, 2006 at 08:29:40PM +0200, Thomas Vander Stichele wrote:
> > Strictly speaking this would be an ABI change; however, nothing in
> > GStreamer actually serializes events at the moment.  The question is,
> > which of the following is preferred:
> > 1) changing it without revving the GDP major/minor version
> > 2) changing it, revving the major/minor, thus making it so older core
> > cannot use GDP with newer core processes
> > 3) changing it, revving, and implementing fallback code for the older
> > GDP version
> 
> I think what you actually mean is:
> 
> 1) changing it without revving the GDP major/minor version, thus making
> it so older core communicating with newer core *breaks in non-intuitive
> ways*.

IIRC the data protocol lib will just do GST_WARNING and drop packets it
doesn't understand, so that would be fine IMO.  I will double-check when
I have multifdsink properly using gdppay.

> > Personally I lean towards 1) because nobody is using the event
> > serialization code right now.
> 
> The same reasoning justifies 2).  If people have a need for backward
> compatibility, they can use the old libgstdataprotocol, as well.

Hm, how would they use the old one, apart from building an old version
of gstreamer and then manually putting the library in there ? It sounds
to me like if we want to allow people to use the old one, we need to put
in compatibility code in the library of the new gstreamer version.


I don't necessarily mind doing that, though it's some work and some
bloat, but I want to know first if people consider it a necessity.

Doing 2) will ensure that old and new refuse to work together, which to
me seems worse than just not deserializing events (and thus implement
the same behaviour as in old core)

Thomas