This I don't agree with Jesper. The autors of the generated classes are the authors of the parser and not the authors of the underlying XSD. My view: Postpone to the major refactoring step of SHRINKWRAP-21
SHRINKDESC-124MetadataParser verbose flag doesn't take effect:
Is simple to implement so that the xslt process is aware of the verbose flag. It is not a must for this release. My view: Postpone to the major refactoring step of SHRINKWRAP-21
SHRINKDESC-125XML representation for toString():
Is simple to implement. XmlDomDescriptorExporterImpl already provides this but is not visible to the generated impl classes. My view: Yes, we should do it for the next release.
SHRINKDESC-126JavaDoc incorrect for parameters:
I expect lot of changes when we implement SHRINKDESC-21. My view: Postpone to the major refactoring step of SHRINKWRAP-21
SHRINKDESC-127Fixed values doesn't get applied as constant value:
I don't think that we should implement too much validations and constraints into the parser. But I believe that we need a validation function as we have shortly discussed in Boston.
This would solve all are not implemented things like xsd:choice and other things. I have already spent some effort in this but was frustrated about the long initialisation
time of the xml transformer. You had the idea of serializing the transformer object. My view: Depends on the time frame. For my, this is THE most important missing feature.