You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
| --- | --- |
| Bugzilla Link | 583269 |
| Status | NEW |
| Importance | P3 normal |
| Reported | May 23, 2024 05:33 EDT |
| Modified | May 23, 2024 05:34 EDT |
| See also | 582494 |
| Reporter | Ed Willink |
Description
Bug 582494 struggles with the longstanding issue that EMF validation cannot know about a Complete OCL enhancement.
Perhaps we have been avoiding an easy solution.
EMF validation also has no fundamental knowledge of OCLinEcore constraints. That lack is remedied by EAnnotation wiring; an EPackage registration, an EClass list of constraints, and EClass constraint bodies.
Why not have a further EPackage registration, an extended EClass list of constraints, and maybe EClass constraint added value?
Fortunately the EPackage.freeze() at the end of built-in installation does not inhibit mutation of EAnnotations.
So OCLinEcore validation is supported by edit-time enhancement of the *.ecore that is persistred to run-time by the synthesized createEAnnotations().
Conversely CompleteOCL validation can be supported by run-time enhancement of the EPackage tree. To avoid leakage, the validator should only respect the extra validations wuithin the ResourceSet for which the Complete OCL was loaded.
The text was updated successfully, but these errors were encountered:
| --- | --- |
| Bugzilla Link | 583269 |
| Status | NEW |
| Importance | P3 normal |
| Reported | May 23, 2024 05:33 EDT |
| Modified | May 23, 2024 05:34 EDT |
| See also | 582494 |
| Reporter | Ed Willink |
Description
Bug 582494 struggles with the longstanding issue that EMF validation cannot know about a Complete OCL enhancement.
Perhaps we have been avoiding an easy solution.
EMF validation also has no fundamental knowledge of OCLinEcore constraints. That lack is remedied by EAnnotation wiring; an EPackage registration, an EClass list of constraints, and EClass constraint bodies.
Why not have a further EPackage registration, an extended EClass list of constraints, and maybe EClass constraint added value?
Fortunately the EPackage.freeze() at the end of built-in installation does not inhibit mutation of EAnnotations.
So OCLinEcore validation is supported by edit-time enhancement of the *.ecore that is persistred to run-time by the synthesized createEAnnotations().
Conversely CompleteOCL validation can be supported by run-time enhancement of the EPackage tree. To avoid leakage, the validator should only respect the extra validations wuithin the ResourceSet for which the Complete OCL was loaded.
The text was updated successfully, but these errors were encountered: