[tei-council] Possible bad bug in attribute classes?

[Martin Holmes]

The @key attribute gets deleted, in theory, in the ODD:

            <classSpec ident="att.canonical" module="tei" type="atts" mode="change">
               <attList>
                  <attDef ident="key" mode="delete"/>
               </attList>
            </classSpec>

and yet it shows up twice in the RNG:

         <ref name="att.personal.attribute.key"/>
         <ref name="att.naming.attribute.key"/>

in many places. Something else odd shows up with the atLeast attribute: it's defined twice:

<define name="att.dimensions.attribute.atLeast">
      <optional>
         <attribute name="atLeast">
            <a:documentation xmlns:a="http://relaxng.org/ns/compatibility/annotations/1.0"/>
            <text/>
         </attribute>
      </optional>
   </define>

   <define name="att.editLike.attribute.atLeast">
      <optional>
         <attribute name="atLeast">
            <a:documentation xmlns:a="http://relaxng.org/ns/compatibility/annotations/1.0"/>
            <text/>
         </attribute>
      </optional>
   </define>

Something done recently to oddbyexample seems to have borked its ability to handle nested attribute classes.

Cheers,
Martin


Martin Holmes
mholmes at uvic.ca
martin at mholmes.com
mholmes at halfbakedsoftware.com

________________________________________
From: tei-council-bounces at lists.village.Virginia.EDU [tei-council-bounces at lists.village.Virginia.EDU] on behalf of Lou Burnard [lou.burnard at retired.ox.ac.uk]
Sent: October 27, 2014 9:06 AM
To: tei-council at lists.village.Virginia.EDU
Subject: Re: [tei-council] Possible bad bug in attribute classes?

Aha, I see this is an odd generated by oddbyexample... in which we now
know (after the workshop) there is a mysterious bug

On 27/10/14 16:01, Martin Holmes wrote:
> I think I've misunderstood this -- att.personal is not explicitly a member of att.canonical -- so now I'm mystified as to why I'm getting this duplicate key error. I'm attaching a zip with the ODD file and generated schema. Can someone else confirm that they also get the error I'm seeing when building a schema from this ODD?
>
> Cheers,
> Martin
>
> Martin Holmes
> mholmes at uvic.ca
> martin at mholmes.com
> mholmes at halfbakedsoftware.com
>
> ________________________________________
> From: tei-council-bounces at lists.village.Virginia.EDU [tei-council-bounces at lists.village.Virginia.EDU] on behalf of Martin Holmes [mholmes at uvic.ca]
> Sent: October 27, 2014 8:56 AM
> To: TEI Council
> Subject: [tei-council] Possible bad bug in attribute classes?
>
> I'm looking at these two attribute classes:
>
> <http://teijenkins.hcmc.uvic.ca/job/TEIP5/lastSuccessfulBuild/artifact/release/doc/tei-p5-doc/en/html/ref-att.naming.html>
>
> <http://teijenkins.hcmc.uvic.ca/job/TEIP5/lastSuccessfulBuild/artifact/release/doc/tei-p5-doc/en/html/ref-att.personal.html>
>
> They're both members of att.canonical, so they both inherit (e.g.) @key from att.canonical. But att.personal is also a member of att.naming, meaning that e.g. @key is inherited twice; this results in broken schemas for elements which are members of att.personal (<name> and friends) because they have @key defined through two separate routes.
>
> Am I right in thinking that att.personal should _not_ be a member of att.canonical because it already inherits that membership through its membership of att.naming? If so, we have a fairly bad error in the current release, and we need to fix it.
>
> I have an ODD which generates the error if required.
>
> Cheers,
> Martin
>
> Martin Holmes
> mholmes at uvic.ca
> martin at mholmes.com
> mholmes at halfbakedsoftware.com
> --
> tei-council mailing list
> tei-council at lists.village.Virginia.EDU
> http://lists.village.Virginia.EDU/mailman/listinfo/tei-council
>
> PLEASE NOTE: postings to this list are publicly archived
>
>

--
tei-council mailing list
tei-council at lists.village.Virginia.EDU
http://lists.village.Virginia.EDU/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived