Re: Fwds Compat/Fallback Conformance Puzzles

Andrew Kimball wrote:
> 
> What is the expected behavior of a conformant processor when given each of
> the following:
> 
> 1.
> <y xsl:version="1.1">
>         <xsl:for-each select="//author">
>                 <xsl:sort-1.1 select="first-name">
>                         <xsl:fallback>
>                                 <xsl:sort select="last-name"/>
>                         </xsl:fallback>
>                 </xsl:sort-1.1>
>         </xsl:for-each>
> </y>

An XSLT processor must report an error.  The contents of xsl:fallback
must be a template.

> 2.
> <y xsl:version="1.1">
>         <xsl:text>
>                 Some text
>                 <xsl:nbsp>
>                         <xsl:fallback>
>                                 Fallback text
>                         </xsl:fallback>
>                 </xsl:nbsp>
>                 More text
>         </xsl:text>
> </y>

An XSLT processor must report an error.  The contents of xsl:text is not
a template, and there are no special forwards compatible processing
rules for elements other than for elements at the template level and
elements at the top-level.

> 3.
> <y xsl:version="1.1">
>         <xsl:apply-templates>
>                 <xsl:positional-param-1.1 select="foo">
>                         <xsl:fallback>
>                                 <xsl:with-param name="foo" select="foo"/>
>                         </xsl:fallback>
>                 </xsl:positional-param-1.1>
>         </xsl:apply-templates>
> </y>

An XSLT processor must report an error for the same reason as 2 since
the contents of xsl:apply-templates is not a template.

> 4.
> <y xsl:version="1.1">
>         <xsl:apply-imports>
>                 <xsl:with-param name="foo" select="foo">
>                         <xsl:fallback/>
>                 </xsl:with-param>
>         </xsl:apply-imports>
> </y>

An XSLT processor must report an error since xsl:with-param has a select
attribute but the content is not empty (first bullet of 11.2).

James

Received on Monday, 15 May 2000 23:28:14 UTC