[ogg-dev] liboggplay: RGBA overlay video, rendering with libtiger

[ogg.k.ogg.k at googlemail.com]

With feedback incorporated.

> Should these be wrapped in HAVE_KATE? Same with the definition of the

I've left the public APIs in, but the entire body is now defined out
apart from the error return code (but this means that using a
wrong track id or type won't be caught, FWIW). Wouldn't be nice
to others (eg, Python bindings) to remove the call altogether, I'd
prefer to let them in with the int foo() { return ENOTIMPL;} kind
of stub, unless you really want the extra handful of bytes.

[unchecked memory allocation]

I've fixed mine and a couple others, but some of the remaining
ones are in code I'm not too familiar with, so I'll look into them
later as it's not obvious *what* should be done should they fail.

> In oggplay.c there are checks like the following that do 'track >=
> me->num_tracks' and at least one that does 'track > me->num_tracks'. I

The only culprit was oggplay_set_offset, fixed.

BTW, I'd already committed a couple 'easy' fixes to liboggplay,
in case you want to have a look too. That last oggplay_set_offset
could go in now too, though I don't know what it's used for.

Cheers
-------------- next part --------------
A non-text attachment was scrubbed...
Name: lop.diff
Type: application/octet-stream
Size: 42687 bytes
Desc: not available
Url : http://lists.xiph.org/pipermail/ogg-dev/attachments/20081216/70827965/attachment-0001.obj