[gst-devel] Docs/RFC: GstData

[Benjamin Otte]

On Sat, 20 Apr 2002, Baker, Steve wrote:

> Instead of having a GstEventLength, the GstEventNewMedia could contain the
> length of the stream if known.
>
That's a possible optimization I have thought about, too but omitted for
the KISS principle.
The length event is needed anyway, because many streams don't know the
length when they start playing and need to supply it later (VBR mp3 comes
to mind).

> Also, it would be good if GstEventDiscontinuous contained the byte offset so
> that (for example) gst_bytestream_tell can be updated correctly.
>
GstEventDiscontinuous is a GstEventInstream and so has an offset property.
This offset should be set to the new offset.
So it's already included.


Benjamin