RE: [xsl] Collect word count with xslt2.0 on saxon 8

["Karen McAdams" <kamcadam@xxxxxxxxx>]

Just a short note about this problem: I did not end up having the time
to follow your suggestion and as well I found out I did not need to -  I
used another xml source tree and the transform worked!

It turns out that the other source tree I was working with I had applied
pretty printing to and subsequently saved.

Once I used a source tree that did not contain the indenting, etc... The
transformation behaved as expected.

Karen

-----Original Message-----
From: George Cristian Bina [mailto:george@xxxxxxxxxxxxx]
Sent: Friday, May 19, 2006 1:19 AM
To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
Subject: Re: [xsl] Collect word count with xslt2.0 on saxon 8

Dear Karen,

I cannot see any reason why adding that empty template will cause the
stylesheet to loop, or at least the stylesheet I posted. Anyway such
loops are easily detected with a debugger. As you use Saxon 8 then you
can try either with oXygen or with stylus. oXygen has a cycle detection
option and by default it will notify you when the stack depth is greater

than 300. Or you can post a cut down sample (XML and XSL) that has this
problem.

Best Regards,
George
---------------------------------------------------------------------
George Cristian Bina
<oXygen/> XML Editor, Schema Editor and XSLT Editor/Debugger
http://www.oxygenxml.com


Karen McAdams wrote:
> HI George;
>
> Thanks for your help. Naturally I tried the draft-comment template
first
> before emailing however I got an error when I tried that -
>
> "Fatal Error! Too many nested apply-templates calls. The stylesheet is
> probably looping."
>
> All I did was add the following to your stylesheet
>
> <xsl:template match="draft-comment"  mode="getText"/>
>
> Any ideas why this is happening?
>
> Thanks.
>
> -----Original Message-----
> From: George Cristian Bina [mailto:george@xxxxxxxxxxxxx]
> Sent: Thursday, May 18, 2006 12:41 AM
> To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> Subject: Re: [xsl] Collect word count with xslt2.0 on saxon 8
>
> Hi Karen,
>
>  > It seems that the template rule would remove the node from the
source
>
> In XSLT the source is read-only, XSLT does not change the source in
any
> way.
>
> Have a look at the built-in template rules
> http://www.w3.org/TR/xslt#built-in-rule
>
> If you have an empty stylesheet that will actually use only built-in
> rules and will walk the source document and output all the text nodes.
>
> The sample I posted used two modes so that we can walk the source tree

> with two sets of rules. The first set of rules in the default mode
> identify the elements whose text content we want to count and start a
> walking of the source from each identified element children using the
> second set of rules that are placed in the getText mode.
> <xsl:apply-templates mode="getText" select="node()"/>
> If we do not define any template for this mode then the built-in
> template rules will get the all the text nodes, but as we want to
> exclude elements with the specific mark then  we just add a rule that
> matches them and does nothing. If you want also the comments to be
> excluded then you should add a rule for that:
>
> <xsl:template match="comment()"  mode="getText"/>
>
> Placing the call to apply-templates with the getText mode inside a
> variable redirects the output to that variable, then we use it to
> generate the required output on the template rules in the default
mode.
>
> Best regards,
> George
> ---------------------------------------------------------------------
> George Cristian Bina
> <oXygen/> XML Editor, Schema Editor and XSLT Editor/Debugger
> http://www.oxygenxml.com
>
>
> Karen McAdams wrote:
>> Hi George-
>>
>> Thanks for the code sample. I would like to better understand how
this
> apply-templates works-
>>    <xsl:template match="*[contains(@class, 'topic/topic')]"
>>      mode="getText"/>
>>        <xsl:apply-templates mode="getText" select="node()"/>
>>
>> It seems that the template rule would remove the node from the source
> at the processing context yet with the apply templates it actually
> retrieves it- could you elaborate on this approach - I also need to
> remove from the node comment elements  before I count the text
contents
> of the node -
>> If I add an template
>> <xsl:template match="draft-comment"  mode="getText"/>
>>
>> I get  -
>> BUILD FAILED
>> java.lang.StackOverflowError
>>
>> ( I am building with Ant but this is an xsl related error . )
>>
>> Karen
>>
>> ------------------------------
>>
>> Date: Tue, 16 May 2006 10:04:22 +0300
>> To:  xsl-list@xxxxxxxxxxxxxxxxxxxxxx
>> From: George Cristian Bina <george@xxxxxxxxxxxxx>
>> Subject: Re: [xsl] Collect word count with xslt2.0 on saxon 8
>> Message-ID: <44697976.1000108@xxxxxxxxxxxxx>
>>
>> Hello Karen,
>>
>> You can get the count of words more easily than that. First you can
> get
>> the text in a variable that belongs to an element with topic/topic
but
>
>> not to other elements inside it with the same mark and then just
count
>
>> the words in that.
>> For getting the text once we match on a topic/topic element we use a
> new
>> mode for apply-template on which we do nothing on elements with
>> topic/topic thus we exclude their text content.
>>
>> The following stylesheet shows that
>>
>> <xsl:stylesheet version="2.0"
>>    xmlns:xsl="http://www.w3.org/1999/XSL/Transform";>
>>    <xsl:output indent="yes"/>
>>    <xsl:template match="/">
>>      <counts>
>>        <xsl:apply-templates/>
>>      </counts>
>>    </xsl:template>
>>    <xsl:template match="text()"/>
>>    <xsl:template match="*[contains(@class, 'topic/topic')]">
>>      <xsl:variable name="text">
>>        <xsl:apply-templates mode="getText" select="node()"/>
>>      </xsl:variable>
>>      <record>
>>        <text>
>>          <xsl:value-of select="$text"/>
>>        </text>
>>        <count>
>>          <xsl:value-of
>>
>>
>
select="count(tokenize(lower-case($text),'(\s|[,.!:;]|[n][b][s][p][;])+'
> )[string(.)])"
>>          />
>>        </count>
>>      </record>
>>      <xsl:apply-templates/>
>>    </xsl:template>
>>
>>    <xsl:template match="*[contains(@class, 'topic/topic')]"
>>      mode="getText"/>
>> </xsl:stylesheet>