[tei-council] attRef/@class not valid?

[Martin Holmes]

The attRef spec contains this example:

<attRef class="att.global" name="xml:id"/>

but when I try to use that in my ODD file, Oxygen complains that:

Description
attribute "class" not allowed here; expected attribute "copyOf", 
"corresp", "exclude", "n", "name", "next", "prev", "rend", "rendition", 
"sameAs", "select", "style", "synch", "xml:base", "xml:id", "xml:lang" 
or "xml:space"

I'm using the TEI ODDs schema:

file:/home/mholmes/Oxygen%20XML%20Editor%2015/frameworks/tei/xml/tei/custom/schema/relaxng/tei_odds.rng

If I point at the latest build:

<http://teijenkins.hcmc.uvic.ca/job/TEIP5/lastSuccessfulBuild/artifact/release/xml/tei/custom/schema/relaxng/tei_odds.rng>

That error disappears, but I get different errors from the Schematron:

Engine name: ISO Schematron
Severity: error
Description: cvc-id.2: There are multiple occurrences of ID value 
'tei_odds-relatedItem-targetorcontent1-constraint'.
Start location: 4661:73
End location: 4661:73
URL: http://www.w3.org/TR/xmlschema-1/#cvc-id

Engine name: ISO Schematron
Severity: error
Description: cvc-attribute.3: The value 
'tei_odds-relatedItem-targetorcontent1-constraint' of attribute 'id' on 
element 'pattern' is not valid with respect to its type, 'ID'.
Start location: 4661:73
End location: 4661:73
URL: http://www.w3.org/TR/xmlschema-1/#cvc-attribute

I think this is due to the fact that there are two things, a report and 
a constraint, inside the constraint element in the relatedItem.xml:

   <constraintSpec ident="targetorcontent1" scheme="isoschematron">
     <constraint>
       <sch:report test="@target and count( child::* ) &gt; 0">
If the @target attribute on <sch:name/> is used, the
relatedItem element must be empty</sch:report>
       <sch:assert test="@target or child::*">A relatedItem element 
should have either a 'target' attribute
         or a child element to indicate the related bibliographic 
item</sch:assert>
     </constraint>
   </constraintSpec>

I'm going to test this by splitting them out into individual 
constraintSpecs. If you know better, please put me right. Meanwhile, we 
seem to have a breakage in the published schema - - any idea why?

Cheers,
Martin