[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [gmf-dev] Fw: [modeling-pmc] Re: concerned at the state of CVS with the recent changes to OCL

GMF Tooling needs OCL to produce a diagram editor (model validation, transformation and codegeneration). In runtime, one needs emf.validation.ocl (and MDT OCL then) only if uses OCL audits/constraints (i.e. use of Java audits doesn't impose emf.validation.ocl use).

Artem

On Tue, Dec 8, 2009 at 6:48 PM, Antoine Toulme <antoine@xxxxxxxxxxxxxxx> wrote:
Hi Anthony,

I have a rather stupid question. In the past, for the BPMN modeler, we never needed OCL. OCL only came as an optional part of the validation framework at the time, and I'm pretty sure we never used in the BPMN modeler runtime.

Would it be possible to not depend on OCL at all ? Where is it used now ?

Thanks,

Antoine

On Tue, Dec 8, 2009 at 07:16, Anthony Hunter <anthonyh@xxxxxxxxxx> wrote:

Forwarding to gmf-dev so we all are aware that GMF in Helios M4 will require MDT OCL 3.0.0.

Cheers...
Anthony
--
Anthony Hunter mailto:anthonyh@xxxxxxxxxx
Software Development Manager
IBM Rational Software: Aurora / Modeling Tools
Phone: 613-270-4613

----- Forwarded by Anthony Hunter/Ottawa/IBM on 2009/12/08 10:15 AM -----


From:

Artem Tikhomirov <tikhomirov.artem@xxxxxxxxx>

To:

PMC members mailing list <modeling-pmc@xxxxxxxxxxx>

Cc:

Anthony Hunter/Ottawa/IBM@IBMCA, Artem Tikhomirov <Artem.Tikhomirov@xxxxxxxxxxx>

Date:

2009/12/07 08:01 PM

Subject:

Re: [modeling-pmc] Re: concerned at the state of CVS with the recent changes to OCL




Yes, GMF is planning to use QVTo 3.0, hence OCL 3.0. We already have all the necessary code changes ready in a CVS branch, and were waiting for bug 294549 to be fixed.

Artem.

On Mon, Dec 7, 2009 at 8:33 PM, Ed Willink <ed@xxxxxxxxxxxxx> wrote:
    Hi Anthony

    My understanding is that GMF uses QVTo which is using OCL 3.0. Any attempt to mix and match OCL 3.0 and OCL 1.4 is doomed to be a major time waster.

    OCL 1.4 was a well-intentioned but misguided attempt to maintain exact API and OCL '2.0' specification interpretation compatibility. Ed Merks killed this managerially last week by instructing us not to deliver two OCLs to Helios, and practically by introducing an 'API' change into EMF that makes preservation of exact API compatibility far too difficult. I certainly won't be working on OCL 1.4 and so far I'm the only OCL committer who's had time to make significant code contributions. As far as I'm concerned OCL 1.4 is dead.

    For the record the problem with OCL CVS was trivial and was only apparent on IBM SDK 5.0. Five files contained incorrect UTF-8 encodings of an a-acute in comments. These have now been corrected, and the latest Hudson 3.0 build is green. The Hudson 1.4 builds show a thundery state (316 attempts with no successes); probably still under development.

    As I do not have releng access, I do not have the opportunity to test or monitor builds, so only became aware of a problem when Ed broadcast to the world that everything had collapsed. My builds on Windows were all fine. Over the weekend I found out about the Hudson CBI interface (very useful) and so can now watch it. I hope that I will be granted releng access soon so that I can perform builds more often than every six hours and can promote successful builds.

        Regards

           Ed Willink

    Anthony Hunter wrote:
        Hi Ed,

        I was on vacation last week so apologies for any delay on my part.

        I have no idea what the latest OCL plan is, but I have been building EMF validation and GMF with OCL 1.4.

        I created Helios builds for EMF validation and GMF the last week of November with OCL 1.4 M2. These builds were required due to changes at the platform (
        https://bugs.eclipse.org/bugs/show_bug.cgi?id=296177 and https://bugs.eclipse.org/bugs/show_bug.cgi?id=296176 ).

        Since M4 is almost upon us, I suggest we stick with OCL 1.4.

        Cheers...
        Anthony
        --
        Anthony Hunter
        mailto:anthonyh@xxxxxxxxxx
        Software Development Manager
        IBM Rational Software: Aurora / Modeling Tools
        Phone: 613-270-4613


        Inactive hide details for Ed Merks ---2009/12/04 07:07:28 AM---David,Ed Merks ---2009/12/04 07:07:28 AM---David,

From:

Ed Merks <ed.merks@xxxxxxxxx>

To:

"eclipse.org-planning-council" <eclipse.org-planning-council@xxxxxxxxxxx>, PMC members mailing list <modeling-pmc@xxxxxxxxxxx>, Cross project issues <cross-project-issues-dev@xxxxxxxxxxx>, Artem Tikhomirov <Artem.Tikhomirov@xxxxxxxxxxx>, Ed Willink <ed@xxxxxxxxxxxxx>, "Kolb, Bernd" <bernd.kolb@xxxxxxx>

Date:

2009/12/04 07:07 AM

Subject:

Re: [eclipse.org-planning-council] How to encourage roll-up reviews




        David,

        JWT isn't a modeling project.

        I'm a little concerned to see folks like Christian Damus and Richard Gronback as contact points for the projects given they are no longer committers. I'm also concerned at the state of CVS with the recent changes to OCL. It appears that a large portion of the modeling technology stack simply can't build right now, with GMF, QVT, Query, and Validation all appearing to be broken to me. Some of the folks on the CC list really need to take immediate steps to address the outstanding issues or we're headed for a train wreck.

        Regards,
        Ed


        David M Williams wrote:

                As mentioned on another thread, I've updated the starter "participating projects" page,


                http://wiki.eclipse.org/Helios/Participating_Projects

                with a little more top level organization, and added the following guidance at the
                top of the page:


                [Note: the data that's currently there was just a quick edit of last years data to provide a starting point. As projects fill in the table, be sure to discuss with your PMC and Project Chain and Project Peers. It is desired to consolidate as many "rows" as can meaningfully and legitimately be combined. In the end, each row should correspond to one project review (that is, one set of Docuware for each row).]


                Looks like we have 11 Top Level Projects now! Guess I forgot about that "Runtime" one. :) And, appears to me there would be at least 25 "rows", at a minimum.

                If anyone sees any improvements to make, feel free.

                Thanks,






                _______________________________________________
                eclipse.org-planning-council mailing list

                eclipse.org-planning-council@xxxxxxxxxxx
                https://dev.eclipse.org/mailman/listinfo/eclipse.org-planning-council

                IMPORTANT: Membership in this list is generated by processes internal to the Eclipse Foundation.  To be permanently removed from this list, you must contact
                emo@xxxxxxxxxxx to request removal._______________________________________________
                eclipse.org-planning-council mailing list

                eclipse.org-planning-council@xxxxxxxxxxx
                https://dev.eclipse.org/mailman/listinfo/eclipse.org-planning-council

                IMPORTANT: Membership in this list is generated by processes internal to the Eclipse Foundation.  To be permanently removed from this list, you must contact
                emo@xxxxxxxxxxx to request removal.



        No virus found in this incoming message.
        Checked by AVG -
        www.avg.com 
        Version: 9.0.709 / Virus Database: 270.14.96/2548 - Release Date: 12/06/09 07:30:00

         


_______________________________________________
gmf-dev mailing list
gmf-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/gmf-dev



_______________________________________________
gmf-dev mailing list
gmf-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/gmf-dev


GIF image

GIF image