[vlc-devel] [RFC] demux/mp4: fix 17584: signal error if ReadBox_default fails

Filip Roséen filip at atch.se
Fri Nov 4 12:07:15 CET 2016


Hi,

Thank you for trying to explain the issue, but I must be a little slow
today (or there's something that I am missing).

On 2016-11-04 11:51, Francois Cartegnie wrote:

> Le 04/11/2016 à 11:25, Filip Roséen a écrit :
> >> Fix would hide from tree complete but unknown boxes and bring regressions.
> > 
> > Thank you for your reply, but I am not sure I completely understand
> > what you are referring to. Could you elaborate the explanation?
> 
> > From what I can tell the unknown box would simply be discarded if it
> > could not be parsed, or are you saying that the changes would cause
> > parsing to stop completely?
> 
> Unknown, non stsd, are -always- present in tree.

Sure, there might be unknown boxes which we cannot handle - but I
still fail to see how the patch would lead to a regression related to
how those boxes are handled.

>From my understanding other boxes can signal errors in the same way as
`MP4_ReadBox_default` would if the patch is applied, so I must be
missing something.

>  |   |   + udta size 83 offset 600
>  |   |   |   + creq size 36 offset 608
>  |   |   |   + cenc size 39 offset 644 (????)

Could you point me in the direction of a failure related to the patch
so that I properly understand why such thing should not be done (not
now, nor in the future)?

I am still trying to wrap my head around where the patch could go
wrong. I am well aware that you have spent a lot more time with
`demux/mp4` than I have, but I cannot find the point of failure you
are talking about.

Best Regards,\
Filip
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/vlc-devel/attachments/20161104/f6804643/attachment.html>


More information about the vlc-devel mailing list