W3C home > Mailing lists > Public > public-xml-core-wg@w3.org > June 2010

Re: tricky rxp/xmlproc question

From: Richard Tobin <richard@inf.ed.ac.uk>
Date: Tue, 1 Jun 2010 17:15:16 +0100 (BST)
To: ht@inf.ed.ac.uk (Henry S. Thompson), Richard Tobin <richard@inf.ed.ac.uk>, Daniel Veillard <veillard@redhat.com>
Cc: public-xml-core-wg@w3.org
Message-Id: <20100601161516.BF179841764@macpro.inf.ed.ac.uk>
>  1) In the spirit of trying to provide a deterministic answer to the
>     "What infoset do you get" question wrt our profiles, what should
>     we say wrt [element content whitespace], i.e. must be absent, must
>     be present and accurate?

I haven't looked at this in a long time.  As I remember it, whether
text is element content whitespace is a fact about the document
independent of the processor, but non-validating processors are not
required to determine it.  A processor that does not determine it
should set the character's [element content whitespace] property to
"unknown", and set the [all declarations processed] property of the
document to false.

>  2) Does RXP/xmlproc provide e-c-w information when it's not
>     validating but there is a doctype?

RXP doesn't.  It sets pcdata_ignorable_whitespace to false if
the Validate flag is not set.

>  4) Or, do you think the above analysis is wrong and it's actually an
>     error for a processor which isn't validating to supply e-c-w
>     information?

I don't think there's any spec that says it's an error for a processor
to return unnecessary information!

-- Richard

-- 
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
Received on Tuesday, 1 June 2010 16:16:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 1 June 2010 16:16:04 GMT