Home » Modeling » Graphiti » Plans on Graphiti future?
| Plans on Graphiti future? [message #732774] |
Tue, 04 October 2011 15:24  |
 Hernan Gonzalez
Messages: 188
Registered: October 2010
Location: Buenos Aires, Argentina |
Senior Member |
|
|
I've been coding quite a bit lately modifying Graphiti sources, changing some ot its architecture, and I'm wondering where I should be aiming... My changes are motivated partly by my own needs, but also by the desire to make Graphiti more cleanly extensible (in particular, I intend that alternative diagram+model storage scenarios - eg, diagram and model in separated resources, or model-only persistence- can be supported out of the box, or with clean classes extensions), and I'm pleased with my (partial) results, but the truth is that my changes imply some relevant API changes: for example, I've removed the DiagramEditorInput class and introduced a DiagramProvider instead; the EditorInternal is now abstract, the DiagramEditor and the DiagramBehaviour are intended to be extensible, etc.
So, I'm reaching a point in which I must make my mind, whether I resign to maintain my own "private fork" of Graphiti, or what. I really like Graphiti, I think it has a lot of potential; of course, I'd like to contribute to Graphiti, but I know very little about the Graphiti development team processes and plans (I'm not even sure if this is the right place to ask). In particular, I wonder if such architecural changes (which require careful consideration and that will break some things, at least for users that have introduced dependence on internal code) are potentially welcomed or rather out of the question.
Is there some development roadmap or discussions in this regard? Graphiti is approaching a version 1.0.0, and (though I'm not sure how relevant that numbering is), I think it's the appropiate time to discuss this. I suspect that there are quite a few users that are touching internal Graphiti code for its own needs (current Graphiti docs seem to admit that the api is half-baked, in that there are many internal classes/methods that should be not modified or even called from the outside, but at the same time it's often necesary to do that in order to adapt Graphiti to user's needs). This is not a nice state of affairs, will make things more painful to change-polish later. Any comments?
|
|
|
| Re: Plans on Graphiti future? [message #733611 is a reply to message #732774] |
Wed, 05 October 2011 07:58  |
Michael Wenz
Messages: 1931
Registered: July 2009
Location: Walldorf, Germany |
Senior Member |
|
|
Hernan,
great to see you asking...
I was also wondering if you had interest in doing more contributions to
Graphiti and becoming a committer for the project. We would definitely like
to get more interested people involved in the project. So if you are
interested in becoming a committer please let me know.
Regarding the team processes and plans, there's a project plan out for Juno
(http://www.eclipse.org/projects/project-plan.php?projectid=modeling.gmp.graphiti)
that mentions the topics we want to tangle for our Juno release. Besides
there's some developer docu in the wiki
(http://wiki.eclipse.org/GMP/Graphiti) where some details are described, but
there's gaps in there, because until now there was no need to document
things that neatly because our team was co-located in Germany. If you (or
potentially others) will join this surely needs to change and we will add
more stuff to the wiki pages. Just let me know what is missing (or add it
youself, it's a wiki... ;-) )
Regarding API changes: while we always strive for being compatible with
previous versions, we know that there are parts in the framework we will
need to change API to get them cleaned-up; especially with our 1.0 release
in mind. One large area here is the editor and what you are doing is highly
interesting to us. Actually we have just started to think about changes that
need to be done here, so please continue to share your thoughts on this. A
good place for more detailed discussions is either the bug for the editor
API you already commented or the graphiti-dev mailing list
(https://dev.eclipse.org/mailman/listinfo/graphiti-dev).
Forking would be no good alternative in my view and yes, this is the perfect
time to discuss this topic.
Michael
|
|
|
| Re: Plans on Graphiti future? [message #734357 is a reply to message #733611] |
Fri, 07 October 2011 14:30 |
Hernan Gonzalez
Messages: 188
Registered: October 2010
Location: Buenos Aires, Argentina |
Senior Member |
|
|
Thanks for the openness, I'm deeply honoured and all that  I'm certainly interested in contributing to Graphiti. Perhaps currently I don't know enough about the whole framework and the underlying platforms (I'm very novice at EMF -and Git- and I have a rather cursory knowledge of GEF - and I'm not very strong at the whole Eclipse stack), but I'm always glad to learn. I will soon post some thoughts and doubts about what I've been coding and musing on (related to the DiagramEditor, mainly).
For now, a quick question: I've branched from the current head (http://git.eclipse.org/c/gmp/org.eclipse.gmp.graphiti.git/) which would correspond to Graphiti 0.9.0, target for Juno. I see that org.eclipse.graphiti.mm requires (MANIFEST.MF) org.eclipse.core.resources v 3.8.0 (which I don't have, as I'm developing on Indigo). I wonder if this dependency is necessary; considering that Graphiti 0.9.0 should run on Juno and Indigo..., version 3.7.0 would seem enough, isn't it?
|
|
| |
Goto Forum:
Current Time: Thu Apr 25 14:42:42 GMT 2024
Powered by FUDForum. Page generated in 0.03260 seconds |
] 
Search
Help
Register
Login
Home
