Hi David,

Looks like the schema [1] in the XML Catalogs 1.1 spec has the correct
escaping. I have no idea where the currently distributed catalog.xsd [2]
came from but indeed it isn't valid.

Thanks.

[1]
http://www.oasis-open.org/committees/download.php/14809/xml-catalogs.html#a.w3cxmlschema
[2]
http://svn.apache.org/viewvc/xml/commons/trunk/java/etc/catalog.xsd?view=log

Michael Glavassevich
XML Parser Development
IBM Toronto Lab
E-mail: mrglavas@ca.ibm.com
E-mail: mrglavas@apache.org

David M Williams <david_williams@us.ibm.com> wrote on 12/21/2007 03:31:06
AM:

> I've recently started using Xerces 2.9.0, which has
xml-commons-resolver-1_2
>
> It has a catalog.xsd file, in org.apache.xml.resolver.etc
>
> This schema doesn't validate, even by Xerces 2.9 itself!!??
>
> The schema has
>
>   <xs:simpleType name='pubIdChars'>
>     <!-- A string of the characters defined as pubIdChar in production 13

>          of the Second Edition of the XML 1.0 Recommendation. Does not
include
>          the whitespace characters because they're normalized by XML
> parsing. -->
>     <xs:restriction base='xs:string'>
>       <xs:pattern value="[a-zA-Z0-9-'()+,./:=?;!*#@$_%]*"/>
>     </xs:restriction>
>   </xs:simpleType>
>
>
> But the validator says the pattern value should be
>
>     <xs:pattern value="[a-zA-Z0-9\-'()+,./:=?;!*#@$_%]*"/>
>
> That is, have an escaped '-' character where it's used literally.
> Some googling around implied the parentheses '(' and ')' should also
> be escaped, giving
>
>       <xs:pattern value="[a-zA-Z0-9\-'\(\)+,./:=?;!*#@$_%]*"/>
>
>
> So ... I'm curious, is this a simple bug and the file that's being
> distributed should be fixed?
> Or, is there some complex special case here where is doesn't matter?
> Or that is should be escaped for
> one type of context or processing but not escaped for another?
>
> I don't know of any cases of the resolving functionality actually
> failing because of this ... but,
> seems it should shipped in a valid form?
>
> Thanks for any education/insight you can give me.