Re: CRLF in xslt in Firefox and IE

Hi David,

you hit the point exactly, thank you. So its essentially a fake and
sloppy programming in IE.
The only positive point is, we can recognise IE from inside the xslt
easily, just in case one needs it.

Manfred

2005/7/7, David Carlisle <davidc@xxxxxxxxx>:
>
> I can confirm that you get CRLF in IE and LF in mozila.
> It's clear that IE is wrong here but it _isn't_ wrong because it has
> left CRLF pairs in the input (that would be terribly broken and would
> have been reported here before now)
>
> If you add
>
> [<xsl:value-of select="string-length('&#xD;&#xA;')"/>]
>
> to your stylesheet you will see you get the value 2 from a conformant
> system and 1 from IE.
>
> msxml is normalizing the #10#13 pair even though it is in the
> stylesheet as character references. so the XPath test that is actually
> being done when you do
>       <xsl:when test="contains(string(doc),'&#xD;&#xA;')">
> in IE is actually
>       <xsl:when test="contains(string(doc),'&#xA;')">
> This is incorrect but doesn't hit you in real stylesheets as you don't
> normally test for #10#13 pairs as you "know" they are not there....
>
> David