Community
Participate
Working Groups
If this is not the case, XMLSchemaValidator uses the user directory as the base to normalize hints. In ARTOP tests, the XSD schema is searched in <testplugin>/working-dir if converters are used.
Sounds reasonable, but I need to better understand how exactly you want this to be done. First of all, with XMLInputSource, do you mean org.apache.xerces.xni.parser.XMLInputSource or org.xml.sax.InputSource? I imagine that the latter is the case given that org.jdom.input.SAXBuilder#build() methods accept only InputSource but not XMLInputSource. Second, from where and how should the publicId and systemId parameters be initialized? Should they be calculated in any way or should application be enabled to initialized them via setter methods?
(In reply to comment #1) > Sounds reasonable, but I need to better understand how exactly you want this to > be done. > > First of all, with XMLInputSource, do you mean > org.apache.xerces.xni.parser.XMLInputSource or org.xml.sax.InputSource? I > imagine that the latter is the case given that > org.jdom.input.SAXBuilder#build() methods accept only InputSource but not > XMLInputSource. Yes, sorry, org.xml.sax.InputSource > > Second, from where and how should the publicId and systemId parameters be > initialized? Should they be calculated in any way or should application be > enabled to initialized them via setter methods? if (resource.getURI() != null) { String resourceURI = resource.getURI().toString(); inputSource.setPublicId(resourceURI); inputSource.setSystemId(resourceURI); ... } See ExtendedXMLLoadImpl line 122. In XMLSchemaValidator line 2522, the fXSDDescription.setBaseSystemId(fLocator.getExpandedSystemId()) must set the base system to a value that makes it possible to find the XSD.
Fixed in AbstractModelConverter#doConvertLoad() as proposed in comment #2. Feel free to reopen if not appropriate.
Mass-closing Resolved tickets
