Open eclipse-ocl-bot opened 1 week ago
By Adolfo Sanchez-Barbudo Herrera on Oct 03, 2013 08:35
(In reply to Ed Willink from comment #0)\
CS genmodels/CS visitors regularize e.g. BaseCST.ecore wrt BaseCSVisitor.
Done in commits between:
commit 651860e8e2e80d21cf0a9903442feb11ec232676
and
commit 261d2f6bbf29b85961b4c868471920841711d09e
For the track, naming convention agreed:
In the interim (done in commits above):
Generated visitors (emf-gen)\
Manual visitors (src)\
Eventually (when autogen is completely functional for OCL):
Generated visitors (emf-gen)\
Manual visitors (src)\
This changes has been done before non-examples promotion because:
By Ed Willink on Oct 21, 2013 09:19
commit 261d2f6bbf29b85961b4c868471920841711d09e\ Author: Adolfo SBH adolfosbh@gmail.com 2013-09-26 16:26:47\ Committer: Ed.Willink ed@willink.me.uk 2013-10-02 11:39:15
------ examples/org.eclipse.ocl.examples.xtext.base/META-INF/MANIFEST.MF ------ ...\
Introduced a redundant plugin dependency without a lower bound breaking Kepler compatibility.
We need to keep an eye out for the JDT refactoring facilities that are doing this.
By Adolfo Sanchez-Barbudo Herrera on Oct 21, 2013 15:02
(In reply to Ed Willink from comment #2)
commit 261d2f6bbf29b85961b4c868471920841711d09e Author: Adolfo SBH adolfosbh@gmail.com 2013-09-26 16:26:47 Committer: Ed.Willink ed@willink.me.uk 2013-10-02 11:39:15
------ examples/org.eclipse.ocl.examples.xtext.base/META-INF/MANIFEST.MF
...
- org.eclipse.emf.ecore;visibility:=reexport,
Introduced a redundant plugin dependency without a lower bound breaking Kepler compatibility.
We need to keep an eye out for the JDT refactoring facilities that are doing this.
Hi Ed,
Apologies for the mistake. I don´t remember adding that line myself. Who knows if JDT refactoring or myself...
Regards,\ Adolfo.
By Ed Willink on Oct 22, 2013 13:20
MetaModel => Metamodel
ConstructorExp => InstanceLiteralExp
? TypeExp => ElementLiteralExp
By Ed Willink on Nov 02, 2014 05:59
Most of the Pivot/AS changes are now done.
Package paths such as
org.eclipse.ocl.examples.xtext.oclstdlib.oclstdlibcs.impl
seem a bit bloated.
Clearly org.eclipse.ocl.xtext.oclstdlib will be the Xtext/Plugin root.
Perhaps we can trim to org.eclipse.ocl.xtext.oclstdlibcs.impl since having both ...oclstdlib and ...oclstdlibcs in the ...oclstdlib plugin will not be too confusing. (Unfortunately oclstdlibcs as the package name is the one part that we cannot sensibly change.)
NsURIs are still e.g. http://www.eclipse.org/ocl/3.1.0/OCLstdlibCST. I think we should follow EMF, platform UI (and almost Xtext), rather than UML2, and go for e.g. http://www.eclipse.org/ocl/2015/OCLstdlibCS
The Xtext/CS plugins will be tightly coupled to the language making it just about impossible to provide any changes without major version changes. Do we add an ".internal" to all xtext packages, or do we just document that ".xtext" means internal? OCL extending applications such as QVT are mandated to a minor version match on all CS plugins.
| --- | --- | | Bugzilla Link | 416229 | | Status | NEW | | Importance | P3 normal | | Reported | Aug 30, 2013 06:53 EDT | | Modified | Nov 02, 2014 05:59 EDT | | Reporter | Ed Willink |
Description
This Bugzilla is intended to accumulate renames to be done when the 'examples' is promoted to non-examples; deferring many API breakages till necessary. Some changes may occur earlier as part of evolution.
Pivot in name changes to AS
pivot as model element changes to ast
pivot in plugin stays unchanged
pivot as file extension changes to oclas
pivot as protocol vanishes
CS genmodels/CS visitors regularize e.g. BaseCST.ecore wrt BaseCSVisitor.
DomainXXXX/XXX/XXXImpl; do we need all three?