[modeling-pmc] [CQ 2342] QVT Abstract syntax support

[emo-ip-team@xxxxxxxxxxx]

http://dev.eclipse.org/ipzilla/show_bug.cgi?id=2342





--- Comment #13 from Ed Willink <ed@xxxxxxxxxxxxx>  2008-07-11 14:07:45 ---
> 1.  One of the About Files refers to the ModelMoref site and 
> references
> Sreedhar Reddy and his team at TCS with respect to 
> ModelMorfExamples.  I do not
> see a Non EPL CQ for ModelMoref?  Can we receive some 
> feedback on whether this
> code was removed from CVS and if the About file simply was 
> missed.  If this
> code is contained in the contribution, a NON EPL CQ will be 
> required ASAP.

I have emailed Sreedhar Reddy for further permission to apply EPL to his
examples.

The examples are not a critical part of the contribution.

It is probably easiest to withdraw
org.eclipse.qvt.declarative.examples.qvtrelation
from this CQ and use a separate CQ for the TCS usage ASAP.

[org.eclipse.qvt.declarative.examples.qvtcore contains small cut and pastes
from QVT 1.0 only].

> 2.  The About Files contained do not appear to be our 
> standard templates. 
> Please arrange to provide new ones as a separate attachment 
> to this CQ Please
> refer to http://www.eclipse.org/legal/epl/about.php for the 
> standard templates.

I don't have time to do this at present. I hope Quentin can do it.

You haven't observed that about files are missing from auto-generated
and a couple of other plug-ins. Would it be quicker to just delete them all?

> 3.  Since the content realizes various specifications from 
> OMG such as MOF 2.0,
> OCL 2.0 and QVT 1.0 - we should actually have Non EPL CQs in 
> our systems to
> reflect this.  I know that we have not done so to date but 
> for completeness we
> should do so now.

The MOF 2.0 CQ is a problem for EMF not QVT Declarative.

The OCL 2.0 CQ is a problem for MDT OCL not QVT Declarative.

The QVT 1.0 CQ was resolved during CQ 2004, which I thought
resolved all use of OMG modelling specifications.

> 4.  There are many Gif images contained in the contribution.  
> Did these
> originate from Eclipse or were they authored by the committer, or?

Almost all of the GIFs are autogenerated. A couple are created
by the committer. A couple may be copied from Eclipse platform.

[We should try to find some way to share the approximately 1000 not
usefully distinct GIFs.]

> 5.  QVTTypeResolverImpl.java indicates the documentation was 
> copied from the
> inherited specification?  Can you confirm which OMG 
> specification this is
> referencing?

The source code line

{       // FIXME derive from org.eclipse.ocl.ecore.TypeResolverImpl and
eliminate TupleFactory once bugzilla 182994 resolved

indicates that QVTTypeResolverImpl.java is a copy of
org.eclipse.ocl.ecore.TypeResolverImpl.java
awaiting a bug fix. All comments are copied too.

> 6.  There are many files in the contribution that only 
> contain <copyright>
> which no copyright information or license header.  These may 
> be generated files
> but they should be amended to include the copyright and 
> license information. 
> Please amend or provide feedback?

Perhaps 889 files (including the imminent QVT OML CQ too).

These appear to be all due to missing copyright fields in the auto-generation
templates. Fixing these and regenerating may leave about 10 miscellaneous
properties files to be dealt with as well.

> 7.  RoseUMLImporterPlugin.java contains a comment "javadoc 
> copied from base
> class".  Is this base class EPL?

a) RoseUMLImporterPlugin.java is in the
org.eclipse.qvt.declarative.importer.roseuml
which the original submission indicated is a clone of
org.eclipse.emf.importer.rose.
The same comment appears in
org.eclipse.emf.importer.rose.RoseImporterPlugin.java.

b) the base class is org.eclipse.emf.common.EMFPlugin.

> 8.  QVTTextSourceViewerConfiguration.java contains a comment 
> "this class
> introduces the fix for Bug 195416 from 
> SpellingReconcileStrategy 1.12".  Please
> help us understand this comment.  Is the file from or does it 
> contain code that
> originated from SpellingReconcileStrategy 1.12?  And if so, 
> where does this
> code originate from and under which license?

a) Bug 195416 shows that
org.eclipse.ui.texteditor.spelling.SpellingReconcileStrategy is part of Eclipse
Platform.

b) The code copies code including the CVS delta from 1.11 to 1.12 to avoid
the bug appearing for Eclipse 3.3 users.

c) QVT Declarative will have Eclipse 3.4 as a prerequisite, so this code
copy can be eliminated at an early opportunity.

> 9.  QVT.mdl contains a comment "C++ Reverse Engineered"???

This is one of many gratuitous attributes supplied by the IBM Rational Rose
tool support.

> 10.  RoseUMLEcoreBuilder.java contains a comment "the derived 
> implementation". 
> Which implementation is this referencing?

See 7.

> 11.  AbstractOCLParser.html and OCLParser.html contain author 
> information of 
> Author=Christian Vogt.  Was this a contributor?  If so, how 
> large was the code
> contribution?

Oops. The org.eclipse.qvt.declarative.modelregistry.doc/html/javadoc folders
were autogenerated a long time ago. They should not have been part of the
submission. They can be regenerated later if required.

I think Christian Vogt was one of the committers/contributors to MDT OCL.


-- 
Configure CQmail: http://dev.eclipse.org/ipzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the CQ.