[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: problem with IE5

• From: Chris Lilley <chris@w...>
• To: David Brownell <db@e...>
• Date: Wed, 14 Apr 1999 19:28:47 +0200

David Brownell wrote:
> 
> Looks to me like:
> 
>   (a)   IE5 uses a nonvalidating XML 1.0 parser (modulo bugs)
>         for documents it tries to display;

Yes. Further, it is a parser which uses the full infoset - external DTD
subset and external entities are fetched.

>   (b)   IE5 however REQUIRES conformance to the namespace spec,
>         and thus rejects some well formed XML 1.0 documents,
>         such as Richard's original;

I don't actually have a problem with that; the XML namespace spec
extends the XML 1.0 spec, and the XML 1.0 spec gave fair warning about
colon-containing names and upcoming namespace work.

>   (c)   It also REQUIRES any "xmlns*" attributes found in a DTD
>         to be #FIXED (which is good style) and so rejects documents
>         which don't have #FIXED, yet conform to the namespace spec;

Good style is one thing, but if the XML ns spec does not require fixed
then their parser should not either (I am not online at the moment so
unable to check the spec).

>   (d)   It also REQUIRES a redundant declaration of such xmlns
>         attributes on elements, even in cases where the XML 1.0
>         specification requires the #FIXED default to be provided
>         from the processor (and the namespace spec requires it
>         to be used, effectively 'inherited');

Can you post an example, for checking with other parsers too?

>   (e)   It has some other conformance issue, where the namespace
>         declaration on just the "test" element doesn't work.  This
>         might be related to the issue (d) above.

An example would help here also.

> Chris -- is this basically accurate?

> > <!DOCTYPE test [
> > <!ELEMENT test (foo:bar)>
> > <!ELEMENT foo:bar ANY>
> > <!ATTLIST foo:bar xmlns:foo CDATA #FIXED "..." >
> > ]>
> > <test><foo:bar xmlns:foo="...">123</foo:bar></test>

I found that the redundant declaration in the instance was not required
by the XML processors that I had available.

There are some examples in the new (12 April) SVG spec [1] where the
namespace declaration is in the DTD and is not repeated in the instance;
when viewed in IE5, the source code is displayed (since there is no
stylesheet PI) and the namespace attribute turns up as do the assorted
XLink attributes which are also declared just once in the DTD.

[1] http://www.w3.org/TR/WD-SVG

--
Chris



xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@i...
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo@i... the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo@i... the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa@i...)

• References:
  • RE: problem with IE5
    • From: Chris Lovett <clovett@m...>
  • Re: problem with IE5
    • From: David Brownell <db@e...>

• Prev by Date: CD-ROM of list archives
• Next by Date: Re: ASN.1
• Previous by thread: Re: problem with IE5
• Next by thread: Re: problem with IE5
• Index(es):
  • Date
  • Thread