Re: [xsl] convert XML to NFC

[Michael Kay <mike@xxxxxxxxxxxx>]

On 27/04/2011 17:14, David Carlisle wrote:
> well it's probably quicker to use a tool treating the file as plain 
> text (since the unicode normalisation should not break the xml)
> but since you asked on xsl-list (and I don't know off hand of a text 
> normaliser) then you can have an identity template and templates
>
> <xsl:template match="text()">
> <xsl:value-of select="normalize-unicode(.)"/>
> </xsl:template>
>
>
> <xsl:template match="@*">
> <xsl:attribute name="{name()}" select="normalize-unicode(.)"/>
> </xsl:template>
>
> David

That solution (and Martin Honnen's) will only normalize the content of 
attribute and text nodes. They won't normalize element and attribute 
names, or comments and processing instructions.

I would be inclined to do this at the serialization level:

<xsl:output method="xml" normalization-form="NFC"/>
<xsl:template match="/">
<xsl:copy-of select="."/>
</xsl:template>

Michael Kay
Saxonica