Re: how do indirection in value-of or select?

["James Tauber" <jtauber@xxxxxxxxxxx>]

>i'm willing to change the syntax of the xml if that helps,
>but it does seem like xsl should have some way to do this.
>macros? constants? producing elements that are then processed
>again by the processor? or can xsl just not do a join?


I *think* this is similar to what I've wanted to do for a while and have
raised on this list before.

As far as I can tell, what would solve my problem (and I think yours) is if
an EqualityExpr could take not just a Literal but a select pattern.

In other words,

Say I have a document:

    <a>
        <blist>
            <b n="foo"><bname>Foo</bname></b>
            <b n="bar"><bname>Bar</bname>/b>
        </blist>
        <clist>
            <c>
                <cname>c1</cname>
                <d>foo</d>
            </c>
            <c>
                <cname>c2</cname>
                <d>bar</d>
            </c>
            <c>
                <cname>c3</cname>
                <d>foo</d>
                <d>bar</d>
            </c>
        </clist>
    </a>

I would like to be able to say:

    <xsl:template match="/a/blist/b">
        <xsl:apply-templates select="name"/>
        <xsl:apply-templates select='/a/clist/c[d=@n]'/>
    </xsl:template>

but you can't say d=@n, only d="literal".

At present I use:

    <xsl:template match="/a/blist/b">
        <xsl:apply-templates select="name"/>
        <xsl:choose>
            <xsl:when test='.[@n="foo"]'>
                <xsl:apply-templates select='/a/clist/c[d="foo"]'/>
            </xsl:when>
            <xsl:when test='.[@n="bar"]'>
                <xsl:apply-templates select='/a/clist/c[d="bar"]'/>
            </xsl:when>
        </xsl:choose>
    </xsl:template>

which involves knowing about "foo" and "bar" in advance. Fortunately, what
*is* possible is automatic generation of the above template from the
original document. It works but I don't think it should have to be done that
way.

I *think* a solution to my problem will be a solution to yours and I think
the best solution is if XSL is modified so that an EqualityExpr can take not
just a Literal but a select pattern.

Comments anyone?

James



 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list