[FFmpeg-devel] [PATCH] libavformat/mxfdec: read source timecode from pulldown component

[tomas.hardin at codemill.se]

On 2014-10-26 00:33, Mark Reid wrote:
> On Sat, Oct 25, 2014 at 1:42 PM, Tomas Härdin 
> <tomas.hardin at codemill.se>
> wrote:
> 
>> On Fri, 2014-10-24 at 17:31 -0700, Mark Reid wrote:
>> > ---
>> >  libavformat/mxf.h      |  1 +
>> >  libavformat/mxfdec.c   | 31 +++++++++++++++++++++++++++++--
>> >  tests/ref/lavf/mxf     |  6 +++---
>> >  tests/ref/lavf/mxf_d10 |  2 +-
>> >  4 files changed, 34 insertions(+), 6 deletions(-)
>> >
>> > diff --git a/libavformat/mxf.h b/libavformat/mxf.h
>> > index 036c15e..5b95efa 100644
>> 
>> Looks good. Simple enough that it shouldn't cause any problems. Do you
>> have a spec for it, or is it just an undocumented Avid extension?
>> 
> 
> I found the uuids via the MXFDump util and pulldown objects are 
> documented
> in the aaf sdk
> 
> http://aaf.cvs.sourceforge.net/viewvc/aaf/AAF/doc/aafobjectspec-v1.1.pdf
> (page 56)
> http://sourceforge.net/p/bmxlib/libmxf/ci/master/tree/tools/MXFDump/AAFMetaDictionary.h#l1158

Check.

>> Taking the opportunity here to note that the strong ref resolving 
>> stuff
>> is getting a bit hairy. Could use some refactoring in the future
>> perhaps?
>> 
>> 
> I was thinking the same thing, I'll try take a look into doing that and
> send a refactoring patch in the future.
> thanks for taking the time to review.

That'd be very cool :)

One huge improvement would be to change so that refs are stored in a 
map/hashmap. This would fix a potential problem with specially crafted 
files taking a long time to parse. If that's too much work for a simple 
refactoring run then I might add it as a suggested OPW task for next 
round.

/Tomas