[MPlayer-dev-eng] [PATCH] Use high-level QuickTime decoding API

[Steinar H. Gunderson]

On Sat, Mar 06, 2010 at 03:06:01PM +0100, Reimar Döffinger wrote:
> I can understand it, but I am very much in disfavour of completely throwing
> out code that has been tested and proven working for years.

There are basically three options here:

 - Leave this decoder in its current, arguably broken state.
 - Implement this change, test as thoroughly as possible (I've already
   checked sample files for all QuickTime codecs I could get to work),
   and watch out for regressions.
 - Make a new, separate QuickTime decoder using the new API, and maintain
   both.

I can only reiterate the advantages: Support for two new codecs (AIC and
ProRes), fixing behavior of at least one (3ivx crash on exit), and less code
to maintain.

> What implications does this have? Will this still fallback to other available
> codecs if the QuickTime one is not available?

It never really has; if something failed in ImageCodecInitialize you'd have
the exact same problem.

/* Steinar */
-- 
Homepage: http://www.sesse.net/