Thursday, July 22, 2010

Eclipse Graphical Modeling Project (GMP) Releases

I had a couple of comments post the GMF restructure on the release line up. Why did the GMF Runtime go back to version 1.4.0. Which version of EMF was needed for which version of GMF Runtime. And so on.

I created a table posted on the GMP Website that lists the releases of the GMF Tooling, GMF Runtime and GMF notation projects over the last five Eclipse simultaneous releases.

Hopefully the table makes sense and gives a clear picture on how all the various releases line up. Let me know if there are any comments.

Thursday, May 20, 2010

Helios RC1 +1, +2 +3, go

We are working though the final stages of the Helios Eclipse Simultaneous Release and this week is RC1.

Helios is organized in layers, Eclipse projects gets a number like +1, +3, -1. It is meant to indicate that a +2 project depends on the builds of a +1 and lower projects.

For example, GEF is +1 and depends on the 0 layer Eclipse platform build.

Now that the GMF Restructure is completed, I was keen to make sure I adopt and build with all the RC1 builds for all the dependencies.

Historically, some projects do not do this. They build with the latest available. This leads to some problems and you will see an RC1a build here and there. I wanted to make sure GMF was clean.

GMF has historically been a +2 component. If you look at the diagram below, +2 never fit really well, but given there are +3 components depending on the GMF Runtime, we made it work.



The projects in green are the builds I produced this week. Unfortunately, I was "mostly late" on all of them.

I produced GEF RC1 late on Monday 9:20 EST, we were late pushing some fixes for a few critical bugs.

I was delayed with the EMF project builds. I was waiting for UML2 RC1, but decided I could not wait anymore. EMF Transaction was the last one after EMF Query and EMF Validation on Tuesday at 3:11 EST. (The UML2 team did confirm they were not going to produce an RC1).

Since Tuesday was supposed to be +2, GMF Runtime was going to be a little late as well. It actually was built at midnight on Wednesday, the +3 day.

I decided not to wait for M2M QVTOML RC1 and pushed the last GMF Tooling build. So GMF Tooling got done Wednesday morning 9:09 AM in time for the RC1 Helios packages.

Apologies for the anyone out there waiting for any of the projects I am responsible for.

Thursday, April 22, 2010

Graphical Modeling at Eclipse

Earlier this week we received word that the Graphical Modeling Framework (GMF) Project received EMO approval for it's restructure request. For those who were unaware of the restructure, read further.

The GMF restructure is mainly a reaction to the Graphiti project proposal. The original version of the proposal was to start a new modeling project.

The scope, architecture and dependencies of Graphiti align directly with the GMF Runtime. Given this fact, it did not make a lot of sense to create a second community. The Graphiti project belongs in the same umbrella project as GMF.

To this end, we proposed the creation of the new Graphical Modeling Project that includes the existing GMF project and will include Graphiti.

A restructure of GMF also solves the following:
  • GMF is currently restricted to one runtime and one tooling project.

    • It makes sense in terms of our project scope to allow multiple runtimes and multiple tooling projects.

    • Expand the Graphical Modeling community at Eclipse

  • GMF can only have one project lead, even though since day one there was a project and committer split between GMF Tooling and GMF Runtime components.

    • It makes sense to split the project between GMF Tooling and GMF Runtime.

    • We will have leadership and committers for each subproject.

    • It will fit into the existing model of umbrella projects in modeling such as EMF, MDT, M2M, etc.

    • Easier defect management, release management, etc.

    • Expand the community working in graphical modeling.

  • Clients have requested a separate notational metamodel.

    • Notational metamodel is aligned with the OMG diagram exchange specifications.

    • GMF Runtime has separate features that deliver the notational metamodel, but end users have to extract manually.

    • Makes sense to create a separate GMF Notation sub-component so other projects can reuse without requiring all of the GMF Runtime
The Existing GMF project will be restructured into GMP. GMP will include:
  • GMF Tooling

  • GMF Runtime

  • GMF Notation

  • Graphiti
We plan to complete for the Helios release with Graphiti an incubating component.

We are now starting the actual restructure. The ongoing work will be discussed in Graphical Modeling Framework (GMF) Restructuring if you wish to follow along.

As well, Graphiti's Creation Review is potentially scheduled for Wed, 28 Apr 2010 at 1500 UTC

Tuesday, March 23, 2010

Naming the next Eclipse Simultaneous Release after Helios

The Eclipse Planning Council has a call out to the community for ideas for the name of the next Eclipse Simultaneous Release in June 2011. For those keeping track, it will be our sixth simultaneous release:

Callisto : 2006
Europa : 2007
Ganymede : 2008
Galileo : 2009
Helios : 2010

The name will be roughly alphabetical; must be greater than "H" (but not too much greater, to leave room for later names).

Preference will be given to "I" names, but there is no strict rule that others would not be considered.

Preference given to names that fit the "moon", "heavenly body gods", or "scientists" themes we've had so far, but no strict rule again.

A few ideas so far would be : Isaac, Ion, Isis, Iris,...

Add your ideas and opinions until April 8 on Bugzilla 306864 .

Monday, February 22, 2010

GEF turns 10

While reviewing and committing some small features and bug fixes this month ([1], [2], [3], [4]) , we noticed that the copyright header on some of the GEF source required an update to:

* Copyright (c) 2000, 2010 IBM Corporation and others.

Happy birthday, GEF is turning 10 years old.

It occurred to me that the year 2000 was somewhat wacky given that the GEF project started in 2002 at Eclipse.org. Have we really been working with GEF that long?

I went back to one of our old Rational Software intranet servers, pre IBM acquision. Sure enough I found an original drop of GEF provided by IBM dated 2000.

This drop of GEF has the familiar GEF Logic Editor Example. Anyone old enough to remember the other original GEF example?

-------
  • [1] Bugzilla 245182 - Draw2d Geometry API should support elementary calculations related to Euclidean Geometry
  • [2] Bugzilla 241767 -ToolTipHelper should be able to be subclassed for enhanced tooltips
  • [4] Bugzilla 301995 - Support option to set only positive coordinates for children figures in FreeformLayout