Alex Milowski wrote: > Here's a simple document which fails to validate with 'xsdvalid'. The > additions > to the substitution group are in the second schema called > 'citrus-fruits.xsd'. It > complains that there isn't a non-abstract element in the substitution > group--which there > is in the second schema. > > It does validate if you specify citrus-fruits.xsd on the command line > using -s. > > The files are attached.
Thanks for reporting the bug. The mechanism used to compose the different schemas found in the "xsi:schemaLocation" attribute like for example: --- xsi:schemaLocation='urn:publicid:IDN+cde.berkeley.edu:schemas:examples:fruit-basket:200403 fruit-basket.xsd urn:publicid:IDN+cde.berkeley.edu:schemas:examples:citrus-fruits:200403 citrus-fruits.xsd' --- is clearly not well thought and needs to be completely rewritten. Unfortunately we are releasing XXE today and you'll not get the bug fix before next release (2.6). (Today's release, 2.5 Patch 3, contains of course fixes for all the bugs related to XML-Schemas you have reported.)
