[Matroska-devel] EBML specification component for review - Element Data Size

wm4 nfxjfg at googlemail.com
Fri May 1 14:58:07 CEST 2015


On Fri, 01 May 2015 14:13:32 +0200
"Sebastian G. <bastik>" <bastik.public.mailinglist at gmx.de> wrote:

> 01.05.2015, 12:15 Moritz Bunkus:
> > Hey,
> > 
> > [...]
> >> In the context of Matroska, I'd disallow unknown lengths in almost all
> >> contexts (i.e. disallow them by default, allow them only when
> >> explicitly specifified). Because they make demuxing a pain, and most
> >> Matroska demuxers probably support unknown length at a few places at
> >> best.
> > 
> > While this being true we have to be careful not to invalidate existing
> > files retroactively. Unknown sizes have been in the specs since the
> > beginning; changing them now is not something we should do on a whim (if
> > at all).
> > 
> 
> 
> I have no preference here. In the context of Matroska the only case that
> gets to my mind where elements with unknown sizes occur would be
> (live)streaming where the file is written and then read, before the file
> is completed. (It's the only use-case that gets to my mind, that does
> not mean it is the only one.)
> 
> I strongly prefer demuxers to handle unknown sizes in all cases, no
> matter where they occur, but reality seems to be that software has
> flaws, rather than being perfect. This would make me argue against
> having unknown lengths being expected to occur everywhere. Instead I
> would favor them to be specified where they are allowed to occur.
> 
> On the other hand why should it be limited if there is no negative impact?!

Actually, handling unknown sizes seems to make parsing significantly
harder. Tracking the boundary of the end of the current parent segment
is not enough. You also have to keep track which elements are
(theoretically) possible at all in a given context, so you can properly
switch back to the grandparent segment if an element is specified not to
be possible in the current segment.

> Ideally demuxers would work in safe-mode by default, where they work
> just like the specs tell them, while also having a debug-mode in which
> they try to handle anything that comes along.
> 
> Best regards,
> Sebastian



More information about the Matroska-devel mailing list