Re: DOCBOOK-APPS: 1.57.0 - Numbereditem starts at 2?

[Bob Stayton <bobs at caldera dot com>]

On Fri, Nov 22, 2002 at 01:03:15PM -0500, Andy Jewell wrote:
> Hi, all. . .  I just started running my documents through the XSL-1.57
> style sheets and really like the new updates.  I'm seeing one new
> anomaly that I didn't get in 1.53.  I generate PDFs with FOP-.24 and
> with this XML:
> 
> <?xml version="1.0" encoding="iso-8859-1"?>
> <sect1>
>   <title>Sample title</title>
>   <para>Sample paragraph below the sample title.</para>
> <orderedlist>
>  <listitem><para>Item 1</para></listitem>
>  <listitem><para>Item 2</para></listitem>
>  </orderedlist>
> </sect1>
> 
> I get this output:
> Sample title
> Sample paragraph below the sample title.
> 2. Item 1
> 3. Item 2
> 
> Does anyone else get the same?  Or possibly I'm doing something wrong
> that 1.53 didn't care about but 1.57 does?  Any help would be
> appreciated!

This was a very interesting problem.
The 1.57.0 stylesheets have more sophisticated
numbering options, so you hit new code.
When I tested your sample with xsltproc, the
listitems were numbered 1 and 2.  When I
tested your sample with Saxon, they were numbered 
2 and 3.  Strange.

If you add a proper DOCTYPE declaration to your sample file
to reference the DocBook DTD,, then Saxon also produces 1 and 2.

The issue seems to be how the following
line at the beginning of the template named
'orderedlist-starting-number' in common/common.xsl:

<xsl:when test="$list/@continuation != 'continues'">1</xsl:when>

In the DTD, the 'continuation' attribute for orderedlist
has a default value of 'restarts'.  That's why it
works with the DTD, because this test is satisfied.
That is, 'restarts' != 'continuation' is true.

Without the DTD, and without a 'continuation' attribute
on orderedlist, is this test true or false?
I believe that "does not exist" is not equal to
'continues', so the test must be true.  
Why does saxon think it is false and use the xsl:otherwise
clause that starts the list at 2?  I tried this with
both the Alfred parser and the Xerces parser in Saxon,
with the same result.

Michael Kay's XSLT Reference has a table of what != means
for different operands.  For a nodeset (the attribute node)
and a string, he says:

"True if the node-set contains a node whose string-value
is not equal to the string operand".

But we have an empty node-set.  So that makes the
expression false?
Curious.

-- 

Bob Stayton                                 400 Encinal Street
Publications Architect                      Santa Cruz, CA  95060
Technical Publications                      voice: (831) 427-7796
The SCO Group                               fax:   (831) 429-1887
                                            email: bobs@sco.com