[XviD-devel] [BUG?] intra meaning

[Christoph Lampert]

On Tue, 21 Jan 2003, Dirk Knop wrote:

> Hi Folks,
> 
> Gom, a little childish we are today, aren't we?
> 
> Edouard Gomez wrote:
> 
> >It seems one of my patches is breaking vfw because i used a free slot of
> >the intra value of pFrame.
> 
> Ok, for the slow brains once more:

Now who's the childish one? 

> we decided to report some frame types - in the stats they look like this now
> #define NNSTATS_KEYFRAME    (1<<31)
> #define NNSTATS_BFRAME        (1<<30)
> #define NNSTATS_SKIPFRAME    (1<<29)
> #define NNSTATS_PADFRAME    (1<<28)
> #define NNSTATS_DELAYFRAME    (1<<27)

Sorry, but those at constants from VfW. The core should not depend on
this and it doesn't. Core works very well without VfW, so VfW should get
its constants from the core, not vice versa. 
If VfW needs the codec to return any special values, there should be a
comment explaining this, so people without VfW, like GomGom and me don't
mess it up. I told GomGom to use S_VOP==3, because that's what the core
does internally. 

Btw, as you already mentioned, it's not possible to have two frametypes
at the same time, so using a 32bit bitmask for them does not seem a very
logical choice. 


Anyway, let's choose a different value, and things are resolved.
If millions of users are stupid enough to download an "unstable
development branch" and overwrite their previous install without testing,
then whatever... and maybe the guys who create binaries can test them,
too, the next time, before putting them to download? 

gruel