RE: [xml-dev] Ambiguities in section 4.3.2 of XML 1.0 SE

["Derek Denny-Brown" <derekdb@microsoft.com>]

Unfortunately (for Xml Parsers at least, and users somewhat) Xml does
clearly disallow ]]> in _lots_ of contexts, even when it is clearly not
a CDATA context.  I know I spent a few days trying to implement that
efficiently.  

<rant>
The fact that it requires a 2-character lookahead, and that ']' is way
the other end of the lower 128 characters from all the other special
characters makes it a royal pain.  (By that I mean 'switch' statements
tend to be compiled to much less efficient code, just because you added
a case for that specific character.  I actually tried it with 2
different compilers and the perf of the switch statement was completely
foobared in both compilers, just by adding case ']'...)
</rant>

-derek

> -----Original Message-----
> From: Perry A. Caro [mailto:caro@adobe.com]
> Sent: Monday, August 11, 2003 5:03 PM
> To: xml-dev@lists.xml.org
> Subject: [xml-dev] Ambiguities in section 4.3.2 of XML 1.0 SE
> 
> [I sent the following to xml-editor@w3.org. Am I completely crazy, or
are
> some clarfications called for in the spec? Would you think either of
the
> following examples were not well-formed?
> 
> Example: <foo bar="]]>"/>
> 
> Example: <!ENTITY cdend="]]>">
>          ...
>          <foo bar="&cdend;"/>
> ]
> 
> With respect to section 4.3.2 of the XML 1.0 Specification Second
Edition
> and by implication XML 1.1 CR, there appear to be several ambiguities
> engendered by the following statement:
> 
>   An internal general parsed entity is well-formed if its replacement
>   text matches the production labeled content.
> 
> ... when considered in the context of CDATA Sections and "]]>". For
> example,
> this would imply that the following declaration in an internal DTD
subset
> would result in an internal general parsed entity that is not
well-formed:
> 
> <!ENTITY cdend "]]>">
> 
> ... because the replacement text does not match the [43] content
> production.
> 
> If so ...
> 
> 1) This contradicts statements about Literals in section 2.3, namely:
> 
>   Literal data is any quoted string not containing the quotation mark
>   used as a delimiter for that string. Literals are used for
>   specifying the content of internal entities (EntityValue),
> 
> ... and production [9] EntityValue. Production [9] permits "]]>" as a
> replacement text.
> 
> Furthermore, [10] AttValue also permits "]]>". It would be nonsensical
for
> <foo bar="]]>"/> to be well-formed, but not <foo bar="&cdend;"/>,
using
> the
> entity declaration above.
> 
> 2) This contradicts the last paragraph of section 4.3.2:
> 
>   A consequence of well-formedness in entities is that the logical
>   and physical structures in an XML document are properly nested; no
>   start-tag, end-tag, empty-element tag, element, comment, processing
>   instruction, character reference, or entity reference can begin in
>   one entity and end in another.
> 
> The list appears to be intended to be exhaustive. The lack of "CDATA
> Section" in the list might be interpreted to mean that you can start a
> CDATA
> Section in one entity, and end it in another. Therefore, the
declaration
> of
> &cdend; above should be well-formed.
> 
> ===============================
> 
> Since the well-formedness of internal general parsed entities is
> completely
> defined by productions [71] GEDecl, [73] EntityDef, and [9]
EntityValue,
> what is the value of the statement in section 4.3.2? What does it
intend
> to
> clarify?
> 
> Perry A. Caro
> Adobe Systems Incorporated
> 
> -----------------------------------------------------------------
> The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> initiative of OASIS <http://www.oasis-open.org>
> 
> The list archives are at http://lists.xml.org/archives/xml-dev/
> 
> To subscribe or unsubscribe from this list use the subscription
> manager: <http://lists.xml.org/ob/adm.pl>
>