eclipse-ocl / org.eclipse.ocl

Eclipse Public License 2.0
0 stars 0 forks source link

[pivot] Inconsistent proxies #1694

Open eclipse-ocl-bot opened 1 month ago

eclipse-ocl-bot commented 1 month ago

| --- | --- | | Bugzilla Link | 492430 | | Status | NEW | | Importance | P3 normal | | Reported | Apr 26, 2016 05:30 EDT | | Modified | Oct 26, 2016 04:55 EDT | | Reporter | Ed Willink |

Description

Metamodel and Library seem to get slightly inconsistent proxies.

http://www.eclipse.org/ocl/2015/Pivot#T.pivot.Class\ http://www.eclipse.org/ocl/2015/Library.oclas#T.ocl.Class

http://www.eclipse.org/ocl/2015/Pivot#T.pivot.Class is a reference to the installed form of Pivot.ecore which has no corresponding xmi:ids.

It should be

http://www.eclipse.org/ocl/2015/Pivot.oclas#T.pivot.Class

It probably works because of some legacy workaround for when the conversion policy was less clear.

eclipse-ocl-bot commented 1 month ago

By Ed Willink on May 03, 2016 07:15

(In reply to Ed Willink from comment #0)

It should be

http://www.eclipse.org/ocl/2015/Pivot.oclas#T.pivot.Class

It probably works because of some legacy workaround for when the conversion policy was less clear.

Confirmed. It only works if some other *.oclas import pulls it in. Editing http://www.eclipse.org/ocl/2015/Pivot#T.pivot.Class to http://www.eclipse.org/ocl/2015/Pivot.oclas#T.pivot.Class manually improves the Sample Reflective Ecore Editor.

The construction style of OCLstdlib.java and OCLmetamodel.java is not consistent.

eclipse-ocl-bot commented 1 month ago

By Ed Willink on May 03, 2016 09:00

(In reply to Ed Willink from comment #1)

The construction style of OCLstdlib.java and OCLmetamodel.java is not consistent.

Revising OCLmetamodel seems mostly good (it ripples to a regenerate of QVTbaseMetamodel too). All OCL tests still pass. Sme QVTd tests now have better references. But some tests fail. It would appear that having an old-style Pivot URI installed, and new-style Pivot URI in the workspace encourages some metamodel schizophrenia.

eclipse-ocl-bot commented 1 month ago

By Ed Willink on Oct 26, 2016 04:55

(In reply to Ed Willink from comment #2)

Revising OCLmetamodel seems mostly good (it ripples to a regenerate of QVTbaseMetamodel too).

Revisiting. It seems that some of this has made it to master, but far from all.