The title should be Eclipse Simultaneous Release and the blurb can be this:
SimRel provides infrastructure for coordinating the release of projects used to build Eclipse IDE/RCP applications.
I.e., make the organization page look similar to this:
To avoid disruption to the 2023-09 release, please schedule the migration shortly after the September 13th release. I.e., September 14th or of 15 would be ideal.
Priority
Urgent
High
Medium
Low
Severity
Blocker
Major
Normal
Low
Impact
To reiterate, please do this right after the September 13th release.
@emerks can you please elaborate what you mean? We can't make the migrated repository work together with our tooling, unless we correctly identify it in Project Management Interface.
We may need to create a placeholder "project" in PMI. Simrel is not a project per se; but a project of projects, and operates just like a project on every level.
@wbeaton if my reality matches the real reality (?), do you agree with our aligning PMI & the projects structure to reflect this?
IMHO making SimRel a real PMI project is a good idea - it is an unusual project in the same way that e.g. Eclipse Orbit or Eclipse Packaging is, for example Orbit and Packaging both special rules on committers and SimRel does too. Choosing a PMC to place it under could be contentious, but either Tools (same as Orbit) or Technology (same as Packaging) would work.
Note that https://projects.eclipse.org/releases "namespace" on PMI is essentially owned by "non-project" SimRel project too and I suspect @wbeaton would be delighted if the project could manage this resource too.
Just to add some extra data(or confusion): simrel (as a project) is tracked via an internal only placeholder, but that's just in our backend..
I agree with @jograham and @droy that simrel needs to be a little more 'real' as a project(we should also use this opportunity to update the seriously old-school 'callisto-dev' permissions group)
As current PL of EPP bringing SimRel into EPP is ok - not sure how to manage distinct sets of committers on SimRel vs (current) EPP. Maybe we shouldn't worry about that and combine it together?
thats the way the adoptium project is setup. They have 1 GitHub organization with the name of the top level project, which also hosts all the subprojects, and each subproject has a distinct committers team that is configured via the PMI entry for each subproject.
The downside of that approach is that tooling is a bit more complicated wrt repositories as the assumption that all repos of a GitHub organization belong to a single project is not true anymore (we have a field in the PMI for that purpose, but that can not be used in such a scenario). But if the number of repos is low and not supposed to change a lot, this is not really a concern imho.
Personally I'd just as soon (rather) that it be its own top-level organization in GitHub; nested projects was something we had in Modeling and we still haven't completely been able to get rid of that. I certainly don't want see these pushed together or become conflated:
In the end, EPP and SimRel are currently separate so it just seems cleaner to keep it that way such that each GitHub organization can focus on its own domain. So if it's my decision to make, then that would be my decision.
I'm perfectly fine and good with a SimRel PMI entry being created. I have no strong opinion about whether this be parented by tools versus technology; probably the later makes the most sense though.
If it walks like a duck, and quacks like a duck... for all intents and purposes, simrel is a project and has been behaving like one for many, many years.
It would be great to hear from the projects team on this. In the interim, let's assume we're doing the Right Thing(tm) and go with the above proposal, even if the new project provisioning process is not done yet.
+1 to a technology.simrel project. I believe we need a new project provisioning process to happen - can someone at the webmaster desk kick this off?
Same behavior, this is to ignore/skip local check, but I'm pushing to github and the commit is still present when i'm pushing. I'm trying with git replace --edit 1d52ed1f2c9846cb66bf446984bc1c816893f29e fixed the error in the email address, now I need to push to the old repo, and then redo a clone locally and then push to Github after.
https://github.com/eclipse-simrel has been created, but there is still no simrel project in PMI. So the sync script won't be able to automatically assign permissions to project members.
@wbeaton can we fast-track the creation of this project? Does someone (e.g. @emerks) need to create a EMO ticket for that?
The Simrel has never been an Eclipse open source project. AFAIK the Simrel is "owned" Eclipse IDE Working Group and all resources required for its operation are managed as such.
If we now feel that the Simrel should be a proper Eclipse project, then we need to follow the usual process. I am happy to help expedite that process, but we need a project team and project lead to drive the process.
Note that "create a project" was only one proposed solution to the "permissions problems". I don't suppose there are any other solution that you could suggest? I don't care how we make it work as long as we get it working sooner rather than later because very soon folks will need to start contributing for 2023-12 M1 and I still need to get all the infrastructure working again with the new locations...
@wbeaton I don't want a temporary solution to become a permanent solution, but would it make sense to give simrel committers access to the GitHub repos manually until the simrel Eclipse project is up and running?
This would lower the level of urgency at least.
Note that "create a project" was only one proposed solution to the "permissions problems". I don't suppose there are any other solution that you could suggest?
All of the simultaneous release infrastructure is current managed as working group resources (I believe that this includes the builds, for example). Managing this as an Eclipse open source project makes a lot of sense to me; it should make a lot of the management tasks a lot easier.
To the best of my knowledge we have two options: leverage our existing project infrastructure by making it a project, or continue to treat it as resource that we manage as a special case on behalf of the working group.
Eliminating the special case will make ongoing support easier for everybody involved.
BTW, will "expediting" help with:
I'll see what I can do. I believe, however, that if we can just leave it as a special case a little longer as Fred suggests, it doesn't really matter.
would it make sense to give simrel committers access to the GitHub repos manually until the simrel Eclipse project is up and running?
That all sounds very good! Making it less special is going to be a good thing in the long term.
I'm also very happy with Fred's suggestion so that I can ensure that infrastructure is working in time for people starting to use it. This makes expediting the other things kind of a non-issue.
I will write a project proposal as soon as possible...
Thanks everyone for the constructive suggestions and for the help moving this forward!!
btw. is there a way to create a "private" Eclipse project simply for our tooling to be able to handle that gracefully, but in fact its not a real Eclipse project with all the required paperwork or technicalities.
Our tooling to a large degree relies on some conventions which is a good thing, but that also requires that things are always the same, which would be easier to achieve with something like that. With private I mean that projects like that would not be visible e.g. via projects.eclipse.org and the like.
Things are really progressing well thanks to you guys' great help! I've already gotten the aggregation and the validation working via this two-in-one job: