Re: [xsl] Failing string() test fail

[Karl Stubsjoen <kstubs@xxxxxxxxx>]

Update

After additional research on the list, we discovered that the use of
string(node-set) evaluates to the first occurrance of the node in the
set.  So, the appropriate test is just:  node-set//text().
Thanks Wendell, found this:
http://www.stylusstudio.com/xsllist/200412/post30420.html

Karl..


On Fri, Jun 26, 2009 at 8:47 AM, Karl Stubsjoen<kstubs@xxxxxxxxx> wrote:
> OK, this is XSLT 1.0 (I should get in the habit of mentioning that)
> The general idea is to display a group heading with child heading and
> child detail below each child heading, with the following rules in
> place:
> 1)  only display a group heading if 1 or more child details items occur
> 2)  only display a child heading if 1 or more child details occur
> I'd say the above is typical, well, common anyhow.
> I have proven that my context is valid with xsl:comment(s) like <!--
> YES I AM HERE -->
> I have dummed down the xml (see below) and written similar xpath
> expressions, as a sanity check.  So now I am wondering what is it I
> could be doing wrong.  I can't freely post the xml and xsl source, as
> this is a solution for a client.
> I will continue to test and experiment but am afraid that I might have
> to resort to a workaround.  The demands on this report are pretty
> steep leading up to this point with left to right pivoted groupings
> requiring a few recursive template calls, and following::sibling types
> of selects.
> Is it possible that I have overworked the engine, found a bug?  It
> would be my 2nd in about 5 years if I had!
> Karl..
>
>
> On Fri, Jun 26, 2009 at 1:24 AM, Michael Kay <mike@xxxxxxxxxxxx> wrote:
>>
>> >
>> > What would cause a string() test to fail?
>>
>> Give us some clues:
>>
>> (a) XSLT 1.0 or 2.0?
>>
>> (b) By "fail", do you mean "throw an error", or "return false"?
>>
>> Regards,
>>
>> Michael Kay
>> http://www.saxonica.com/
>> http://twitter.com/michaelhkay