Re: [xsl] problem with whitespace in mixed content (reverse indentation)

["Wolfhart Totschnig wolfhart.totschnig@xxxxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>]

Dear Martin,

Thank you for the prompt reply! I didn't occur to me that I could 
formulate my conditions using the sibling axis. Following your 
suggestion, I found the following solution to my problem:

<xsl:template match="text()">
    <xsl:choose>
        <xsl:when test="matches(.,'^\s+$') and preceding-sibling::* and 
following-sibling::*">
            <xsl:text> </xsl:text>
        </xsl:when>
        <xsl:when test="matches(.,'^\s+$') and 
(not(preceding-sibling::*) or not(following-sibling::*))"/>
        <xsl:otherwise>
            <xsl:value-of select="."/>
        </xsl:otherwise>
    </xsl:choose>
</xsl:template>

Thanks again,
Wolfhart


On 02/27/2015 03:29 PM, Martin Honnen martin.honnen@xxxxxx wrote:
> Wolfhart Totschnig wolfhart.totschnig@xxxxxxxxxxx wrote:
> > Hello,
> >
> > I have a problem with whitespace in mixed content to which I cannot find
> > the solution. I am hoping that one of you can help me.
> >
> > Due to imprudent use of indent="yes" on <xsl:output>, the whitespace in
> > the mixed-content elements of my data got messed up. I'll best explain
> > the problem with an example.
> >
> > The following original data
> >
> > <text><i>Italicized</i> normal <i>italicized</i> <b>bold</b>
> > <i><b>italicized and bold</b></i>.</text>
> >
> > now looks like this:
> >
> > <text>
> >     <i>Italicized</i> normal <i>italicized</i>
> >     <b>bold</b>
> >     <i>
> >        <b>italicized and bold</b>
> >     </i>.</text>
> >
> > I have found out how to avoid the indentation in mixed-content elements
> > in the future, namely by using saxon:suppress-indentation. My question
> > is how I can return the modified data back to its original form, i.e.,
> > reverse the indentation. The task can be formulated thus: In <text>
> > elements, whitespace-only text nodes that are situated between two start
> > tags should be eliminated, while whitespace-only text nodes that are
> > situated between an end tag and a start tag should be replaced by a
> > single space. How can that be done?
> >
> > I know that I can select whitespace-only text nodes with
> > test="matches(.,'^\s+$')". But how can I test whether the preceding tag
> > is a start tag or an end tag?
>
>
>
> You are dealing with text nodes in a tree, contained in element nodes, 
> you are not dealing tags.
> So you need to translate the conditions to one based on the tree 
> model, I think the condition
>
> > In <text>
> > elements, whitespace-only text nodes that are situated between two 
> start
> > tags should be eliminated,
>
> can only occur if the text node is the first child of an element node 
> and followed by an element node e.g.
>
> <xsl:template match="text//text()[not(normalize-space())
>                                   and . is ../node()[1]
>                                   and 
> following-sibling::node()[1][self::*]]"/>