[FFmpeg-devel] [PATCH] Clip-Wrapped MXF support (attempt #4)

[Reimar Döffinger]

Hello,

I am not a maintainer but

On Thu, Jan 13, 2011 at 02:14:41PM +0200, Maksym Veremeyenko wrote:
> This is another attempt to submit patches for clip-wrapped MXF files
> support (generated by Panasonic's P2 camera).
> 
> Patches set:
> 
> *0001-add-MXFContainerUL-struct-for-containers-UL-dict.patch* - start
> use MXFContainerUL instead of MXFCodecUL struct for essence containers list.
> 
> *0002-revert-container-wrapping-detection-from-r11567.patch* - revert
> almost all optimization of clip-wrapped detection from r11567.
> 
> *0003-make-key-output-in-RP-224.10-form.patch* - on output debug use
> RP224 presentation UID for containers and codec - dot separated hex values.
> 
> *0004-add-new-essence-container-uls.patch* - extends well-known (IMHO)
> uid for DV videos and PCM audios recorded by P2 camera.
> 
> *0005-extend-MXFIndexTableSegment.patch* - extends reading datas for
> MXFIndexTableSegment
> 
> *0006-clip-wrapped-support-added-for-single-track-file.patch* - actually
> a $subj...

Having a huge set of (for many cases) barely relevant/related patches before
the actual "beef" doesn't help review.
I would also help if you described how your clip-wrapped support works.
To me it looks like it splits the data according to index entries.
But unless the specification actually _requires_ a complete index (which
means it requires an index at all, and it requires that also each non-keyframe
has an index entry) that is a solution that only works with "good luck", not
properly.
Does this code actually fix any case other that DV video? Because for DV it could be
considered the real bug that the raw DV demuxer was not implemented using a parser
(if it was, clip-wrapped DV should already work I think).