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

Re: Non-validating parser, non-existing DTD,undefined entity a

• To: Murali Mani <mani@C...>
• Subject: Re: Non-validating parser, non-existing DTD,undefined entity and error reporting
• From: David Brownell <david-b@p...>
• Date: Fri, 25 Jan 2002 11:50:09 -0800
• Cc: Heikki Toivonen <heikki@n...>, xml-dev@l...
• References: <Pine.SOL.4.33.0201251128310.16382-100000@p...>

> Section 2.9: XML 1.0 -- "Standalone="yes" indicates there are no external
> markup declarations that affect the information passed from the XML
> processor to the application"
> 
> -- Can I ask you once more whether you think specifying an external subset
> is legal?? -- doesn't an external subset imply that there are external
> markup declarations??

They're legal if they don't _affect_ the information provided.
Where the information affected is a small subset of what
a parser normally provides; for example it doesn't include
attribute or element typing information.

For example, it would be illegal for the external subset to have
a declaration that caused an attribute value to change because
of normalization.  But it would be legal if the value were already
normalized, but the declaration just told you its type was "ID".


> > Not so -- see the conformance section of the XML rec, it
> > allows non-validating parsers which read external entities.
> > In fact as I noted, such parser modes are typical in Java.
> >
> > As I said, a SAX parser that tries to read that WILL get
> > an error, which is of course allowed.
> 
> Yes, they may process external entities, but am I right if I say that --
> "A non-validating parser *need not* issue an error if the URI pointing to
> an external entity is invalid"?? -- again this is what I kind of
> understood from Section 5: Conformance of XML 1.0..

In the scenario that I described -- a SAX parser is configured
to process those enitities -- the parser does _need_ to issue
an error.  "A non-validating parser" in general need not, like
a SAX parser that doesn't recognize that feature setting.

But a SAX parser that's explicitly told to process those entities
had darn well better process them!!

- Dave

• References:
  • Re: Non-validating parser, non-existing DTD, undefinedentity and error reporting
    • From: Murali Mani <mani@C...>

• Prev by Date: Re: There is a meaning, but it's not in the data alone
• Next by Date: Re: AF and namespaces, once again (was Re: There is a m eaning, but it's not in the data alone)
• Previous by thread: Re: Non-validating parser, non-existing DTD, undefinedentity and error reporting
• Next by thread: Re: [Sax-devel] RFE [ #496133 ] New resolver entry point
• Index(es):
  • Date
  • Thread