• To: <xml-dev@l...>
• Subject: Re: genx - consensus items
• From: "Karl Waclawek" <karl@w...>
• Date: Fri, 23 Jan 2004 10:10:19 -0500
• References: <200401230932.i0N9Wkc6071702@a...> <40110F9A.28544717@n...>

----- Original Message ----- 
From: "Ed Davies" <edavies@n...>

> Perhaps StartStartTag or StartTagOpen or whatever can return an opaque
> element reference (could be a pointer to an internal structure or simply 
> a nesting stack level count) which is then passed to StartTagClose and EndTag
> to allow early detection of mismatches.  A bit more lightweight than
> passing the element name + namespace gubbins and maybe having to do
> string compares.

One could use the same approach for namespaces:

- genxDeclarePrefix takes prefix and uri and returns an opaque reference
  (call just before genxStartTag at desired level)
- whereever you need this prefix mapping, just pass the reference to the
  API (genxStartTag, genxAttribute) instead of separate uri/prefix pointers
- to undeclare a prefix, call genxDeclarePrefix with a NULL uri
  (legal for default namespaces in version 1.0, legal for all namespaces in 1.1)
- to have a prefix auto-generated, call genxDeclarePrefix with a NULL prefix
  (but a non-NULL uri)
- to have a default namespace declared, call genxDeclarePrefix
  with the prefix "xmlns" (which is not a legal prefix)


Karl

• References:
  • Re: genx - consensus items
    • From: David Tolpin <dvd@d...>
  • Re: genx - consensus items
    • From: Ed Davies <edavies@n...>

• Prev by Date: Re: genx - consensus items
• Next by Date: RE: XLink and mixed vocabulary design
• Previous by thread: Re: genx - consensus items
• Next by thread: implicit StartTagClose [Re: genx - consensus items]
• Index(es):
  • Date
  • Thread