ietf
[Top] [All Lists]

Re: XML related issues in metalink, was: Last Call: draft-bryan-metalink (The Metalink Download Description Format) to Proposed Standard

2009-12-14 13:24:15
On Sun, Dec 13, 2009 at 5:09 AM, Julian Reschke 
<julian(_dot_)reschke(_at_)gmx(_dot_)de> wrote:
The IESG wrote:

The IESG has received a request from an individual submitter to consider
the following document:

- 'The Metalink Download Description Format '
  <draft-bryan-metalink-24.txt> as a Proposed Standard
...

Hi,

I did a quick check on a few XML related issues, and found:

1) References to W3C specs

- the references currently use many different formats (for
seriesName/value), it would be great if those were consistent (I recommend
to use the format used for "REC-xml" for compactness)

- the up-to-date check yielded:

REC-xml-20060816: [FirstEdition] obsoleted by REC-xml-20081126
REC-xml-infoset-20040204: [REC] ok
REC-xml-names-20060816: [FirstEdition] obsoleted by REC-xml-names-20091208
REC-xmlbase-20010627: [FirstEdition] obsoleted by REC-xmlbase-20090128
REC-xmldsig-core-20080610: [REC] ok

(where the out-of-date xml-names reference is excused :-).

These references have been updated and made more consistent.

2) RNC

- was there an automated check that the collected RNC and the fragments are
in sync? For "metalinkFile" I see a difference in ordering which may
indicate that this didn't always happen (the difference appears to be
irrelevant, but who knows...)

- I found the RNC to miss a few characters (commas, closing braces), which
indicates it may not have been checked recently. I recommend to do that, and
also to validate the examples in the spec against the RNC.

Fixed.

3) XML vs whitespace

I'm not sure I understand the whitespace treatment.

One example has:

   <url location="de" priority="1">
      ftp://ftp.example.com/example.ext
   </url>

while the prose says in Section 3: "Note that there MUST NOT be any white
space in a Date construct or in any IRI."

I personally would prefer that whitespace is NOT ignorable, but in any case
this should be stated somewhere more clearly.

Very helpful, your sentence was added:

"All leading and trailing whitespace is part of the element content,
and MUST be ignored. Consequently, it is disallowed for elements where
the defined type does not allow whitespace, such as dates, integers,
or IRIs."

(Note I didn't review the spec, I just did a few XML related checks)

Thank you, Julian, for these needed checks! Your fresh eyes picked up
things unnoticed. "It takes a whole village to raise a spec" :)

For those interested, the txt, XML, and html updates in revision 25
are at http://metalinks.svn.sourceforge.net/viewvc/metalinks/internetdraft/

-- 
(( Anthony Bryan ... Metalink [ http://www.metalinker.org ]
  )) Easier, More Reliable, Self Healing Downloads
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf