RE: [search-ws-comment] RE: Can't Parse Schema

["Ray Denenberg, Library of Congress" <rden@loc.gov>]

I have (tentatively) fixed these problems, and the new versions are in
http://www.loc.gov/standards/sru/oasis/newSchemas/ 

I haven't checked it out carefully (as Ralph say's "time to go home")  but I
think it's all ok.  Once we have time to check this out we can arrange to
make the changes official.   Note that these corrections do not introduce
any substantive changes whatsoever.

Here are the changes:

> Both searchResultAnalysis.xsd and facetedResults.xsd define a complexType
> named datasourceDefinition.
 
> Both scan.xsd and facetedResults define a complexType named
termDefinition.
 
> Both scan.xsd and facetedResults define a complexType named
termsDefinition.

I gave them separate names for each case.

 
> Both sruResponse.xsd and scan.xsd define a complexType named
> diagnosticsDefinition.

In this case the definitions are identical so I removed the one from scan
and pointed to the definition in SRU.


> There are more examples of this and I?ll report them tomorrow, but it?s
> time to go home now.

I don't think there are any more, because it all validates ok now. 


 
> Ray, the schemaLocation attributes have whitespace embedded in the middle
> of the .xsd filenames and my jaxb parser is failing on them.
> 
> For instance:
> <xs:import namespace="http://docs.oasis-open.org/ns/search-
> ws/diagnostic" sc hemaLocation="http://docs.oasis-open.org/search-ws/
> searchRetrieve/v1.0/csd01/schemas/diagnostic.xsd"/>

I have changed all of the schema locations within import statements to
relative, thus for example

This:

 <xs:import 
namespace="http://docs.oasis-open.org/ns/search-ws/diagnostic"; 
schemaLocation="http://docs.oasis-open.org/search-ws/
 searchRetrieve/v1.0/csd01/schemas/diagnostic.xsd"/>

Becomes:

<xs:import 
namespace="http://docs.oasis-open.org/ns/search-ws/diagnostic"; 
schemaLocation="diagnostic.xsd"/>

And this does eliminate the whitespace problem (because they were all in the
absolute part of the path); but that's isn't why I did this.  These should
have been relative paths all along, so that one can move the entire set from
one directory to another (as in for testing, like I'm doing now) without
having to change all the pathnames.

Thanks for spotting this Ralph.

--Ray