[NUT-devel] [RFC][PATCH] recursive metadata and language & contry for metadata

[Michael Niedermayer]

On Tue, Sep 23, 2008 at 12:53:25PM -0700, Baptiste Coudurier wrote:
> Hi,
> 
> Michael Niedermayer wrote:
> > Hi
> > 
> > Patch below would add $subj
> > 
> > comments and alternative suggestions are welcome!
> > 
> > Index: nut.txt
> > ===================================================================
> > --- nut.txt	(revision 660)
> > +++ nut.txt	(working copy)
> > @@ -970,6 +970,29 @@
> >          might be less pretty.
> >      Note: Nonstandard fields should be prefixed by "X-".
> >  
> > +    When there are multiple tags with the same name but different values,
> > +    then each except the first MUST have a sequential number starting at 2
> > +    appended, for example
> > +    "Author", "Author2", "Author3", ...
> > +    This is required to avoid ambiguity when info packets are repeated.
> 
> Why not overwriting "Author" tag, it will be easier for demuxers, if
> info packet is a repeated one, it would be simpler to mark it as
> "repeated", no ?

I dont fully understand what you suggest, please elaborate.
There can be multiple authors for one part or it may be that in
real time streams, metadata is updated when new becomes available,
packets can be lost and repeated ...

Simply overwriting when a repeat flag is set doesnt work:
"Author" "Michael Niedermayer" [this packet is lost, though even if not the
                                case is still ambiguous]
"Author" "Baptiste Coudurier"
...
"Author" "Michael Niedermayer" repeated
"Author" "Baptiste Coudurier"  repeated

With numbers the problem does not arise
"Author"  "Michael Niedermayer" 
"Author2" "Baptiste Coudurier"
...
"Author"  "Michael Niedermayer"
"Author2" "Baptiste Coudurier"

Also
"Author"            "Michael Niedermayer"
"Author2/X-EMAIL"   "foo at bar"
"Author2"           "Baptiste Coudurier"

Is another case where droping the numbers would not work.


> 
> > [...]
> >
> > +
> > +    names MUST not contain / - or end in numbers to avoid ambiguity.
> 
> Escaping mechanism is needed and would be more flexible IMHO.

IMHO escaping should not be needed, there should be no need to have arbitrary
characters in identifers.

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

There will always be a question for which you do not know the correct awnser.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/nut-devel/attachments/20080924/f3721930/attachment.pgp>