|
History
| Log In
|
|
|
|
Werner,
Back on version 1.0.2 the only problem was So I guess the direct answer is I have not been able to get this far. I think Thanks Okay, I'll have a look later on today. In general, it would be great if you could isolate a problem and provide me with a trimmed down XML schema, as debugging/analysing such big schemas never is easy ....
Let me just repeat that: the jxdm schema has way too many dependencies to be used in a meaningful way, and I don't want to be responsible to building an environment where I can run the source generator successfully. Can you please either trim down the XML schema(s) involved, or present me with a complete set of files where all dependencies are resolved ? Thanks in advance ...
Werner,
Attached is a complete set of xsd files. Also, when I created I have tried to trim it down to a better test case but ran into to may dependancy issues. Using the two attached xsd files with the case does not cause any dependancy issues that I could see, only the compile issues. These two files create by themselves, all the errors. Eclipse does not want to show me all the errors even though I have increased the max number for the project and the workspace but the number of errors match exactly. Thanks It's funny, Eclipse crashes on me however much memory I assign to it .. ;-(.
That class TestGen1 is not attached any more to
Okay, some random observations (from the limited set of generated classes, before my JVM dies):
Class: VictimOffenseChargeText.java extends: TextType.java extends: String.java (not java.util.String) Apparently, that's wrong, as both VictimOffenseChargeText.unmarshal() and TextType.unmarshal() should return xml.c1729.generated.String, and not xml.c1729.generated.TextType. Werner,
Attached is the TestGen1 class I have been using. This should answer your questions. Let me know if I should add the parameter you mentioned. Also, I was having the same problem with eclipse crashing, come to find out, I had some bad memory modules. I had to let the memory test run all night before it would identify a problem, of course by that time I had reformatted my drives and everything else. Thanks The problem with the wrong method signature is a result of generating the schemas individually.
That way the inheritance across multiple schemas is not generated correctly. This can be corrected by using the "-generateImportedSchemas" option and only calling the SourceGenerator for the top schema (in this case "jxdm/3.0.3/jxdm.xsd"). However currently this will result in some other problems:
I attached a patch that fixes both problems. This is a test case that uses the Source Generator in the way described above.
On my machine this succesfully generates compileable code. (Although there are 6 name conflicts). WARNING: If you execute this test case the Source Generator will create more than 10,000 files. When compile this with Eclipse (and probably with other IDEs as well) this is going to take very long. Steven, would it be possible for you to come up with a minimal test case that we could integrate into the CTF suite ?
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
Bug
Open
Major
Code Generator - compile errors - missing types and casting exceptions
CASTOR-1659, or have you never been able to get this far ?