Summary: | '=' operation on attribute of EEnum type and enum literal works incorrectly | ||||||
---|---|---|---|---|---|---|---|
Product: | [Modeling] OCL | Reporter: | Radomil Dvorak <dvorak.radek> | ||||
Component: | Core | Assignee: | Christian Damus <give.a.damus> | ||||
Status: | CLOSED FIXED | QA Contact: | |||||
Severity: | major | ||||||
Priority: | P2 | Keywords: | greatbug | ||||
Version: | unspecified | ||||||
Target Milestone: | RC | ||||||
Hardware: | PC | ||||||
OS: | Windows XP | ||||||
Whiteboard: | |||||||
Attachments: |
|
Description
Radomil Dvorak
2006-04-19 13:52:30 EDT
Created attachment 38957 [details]
unit-test
Thanks again for making another bug so easy to fix by providing a JUnit test. I also took the opportunity to fix the un-registered faked OCL package that you pointed out. The reason why it had to be faked is that EMF does not generate an EPackage implementation for packages that contain no EClassifiers (the ocl package only has subpackages). Attempting to rectify this by initializing the packages from the ecore file resulted in too many ecore files copied into the JAR, as well as EPackages containing themselves at run-time, causing StackOverflowErrors. What I didn't anticipate was that anyone would write OCL expressions on the OCL AST but, Hey, why not? That's perfectly meta-circular as EMF is also :-) Oops, forgot to resolve the bug when I added comment #2. Move to verified as per bug 206558. Closing after over a year in verified state. Closing after over a year in verified state. |