[Playlist] Announcement: JSPF Parser

[Sebastian Pipping]

Chris Anderson wrote:
> I've put the finishing touches on my Javascript XSPF Parser. It's been
> tested on Firefox, Safari, and IE. The license is LGPL 3.0.
> 
> http://jchris.mfdz.com/static/jspf_parser/xspf_parser_runner.html
> 
> Please take a look and send me any suggestions. The output has been
> updated to the JSPF format as drafted on the Xiph wiki page linked
> from the release page.

-------------------------------------------------------------
Hey Chris, nice, nice!

Do I understand right your parser is intended to eat invalid
XSPF as well? It read <for_version_1/fail/track-badint-tracknum.xspf>
for me without complaining.

One thing I noticed is <trackNum> and <duration> values
are put in parentheses:

  "duration":"1"

I guess that's a bug?
-------------------------------------------------------------



> I'd like to put this in the Xiph svn, so hopefully after we've put it
> through its paces I will.
> 
> Especially take a look at the whitespace handling (I think I got it
> right)

-------------------------------------------------------------
What I have seen of whitespace handling so far looked good.
-------------------------------------------------------------



> and any suggestions for how to handle XML namespaces would be
> appreciated.

-------------------------------------------------------------
Not sure what exactly you want to know. I'll just write down
a few notes and hope I'll hit some of your possible questions:

* When you write "<playlist>" in XSPF the actual name
  of that tag is "http://xspf.org/ns/0/" + separator + "playlist".
  "+" is concatenation, the separator can be anything but should
  be a char not legal in URIs like space or '§', as you like.
  Not sure what's the best choice though.

* Namespaces can be opened anywhere and their scope *includes*
  the tag that they reside in (which makes sense since you
  couldn't set a namespace for root otherwise).
  This code

     <one xmlns="http://example.com/one/">
       <two xmlns="http://example.com/two/">
         <three />
       </two>
     </one>

  references these tags:
  * http://example.com/one/ one
  * http://example.com/two/ two
  * http://example.com/two/ three

* What you have seen so far was setting and overriding the
  default namespace. You can also assign prefixes and use
  them to better show namespace affiliation when desired:

    <one xmlns="http://example.com/one/">
      <b:two xmlns:b="http://example.com/two/">
        <b:three />
      </b:two>
    </one>

  references the same tags as before.

* So in general namespaces try to avoid naming conflicts
  without making names too long. They first seem complex
  but are quite easy once you're going.

* Most documents will have namespace references just in the
  root element - that is what is recommended to increase
  interoperability - but in theory they can appear anywhere,
  especially in XSPF extension tags. And since that is both
  valid XML and XSPF a parser should ideally be able to
  handle that cases as well.

* Real life example of XML namespaces in XSPF can be seen here:
  http://wiki.xiph.org/index.php/List_of_known_XSPF_extensions#Project_Opus

I hope there was at least something new to you in here so it
made sense to write it down. Please let me know what else
you need to know.



Sebastian