I am see more and more discussion on usefulness of Orbit project. But there is no discussion on how we are going to maintain the same. With Roland taking other responsibilities in his organisation and not having much time for Orbit project(See: https://www.eclipse.org/lists/orbit-dev/msg05359.html), with no one else stepping in/up I don’t see a point in discussing usefulness of Orbit project, if there is no one willing to maintain it. -Sravan From: Wim Jongman <wim.jongman@xxxxxxxxx>
Sent: 02 June 2021 23:44
To: eclipse-pmc@xxxxxxxxxxx
Cc: Discussions on Eclipse IDE Working Group <eclipse-ide-wg@xxxxxxxxxxx>
Subject: [EXTERNAL] Re: [eclipse-pmc] [eclipse-ide-wg] Fwd: [cross-project-issues-dev] IMHO serious issues in the Eclipse Java IDE 2021-06 M3
On the other hand, my most pressing reason for using Orbit is the fact that it helps mitigate the "fat jar" problem where everyone bundles their own subversion of a particular library. Consuming directly from PDE/M2E abandons this concept. In addition, if the OSGi version number wiring is not correctly followed, which is true in many cases, this could cause issues that would otherwise not surface. I personally love the new option of PDE/ M2e to use Maven libraries directly in plugins. This functionality make converting libs to OSGI very simple. Orbit was a great way to source external libraries for OSGi applications but unfortunately it did not always had all the desired libs in the desired version. For me Orbit solved nicely the issue to get osgified libs but as I can do this now directly I rarely find the need to use Orbit these days for RCP clients. Maybe Orbit was a great solution which has now been made less valuable as PDE and M2e makes direct usage simpler? Mickael, why would the case of not solving this particular point prove that Orbit is not profitable? The case of the logging framework not being able to host two versions is IMO not related to the value of Orbit. The short-term solution to this particular issue is that m2e moves to 2.0. Apparently, you already concluded that Orbit is not profitable even before this issue arose. Are you saying Orbit has no intrinsic value or that it takes too much time to maintain it? I missed the discussion and arguments on why Platform should not use Orbit anymore, so I can't say if I agree or not. So far, I find Orbit very useful. Let me start by rejecting your attempt of framing Orbit benefits by using "hypothetical" as well as denying all the successful, fruitful and helpful collaboration in Orbit which happened in the past and still happens today.
So if those benefits are not hypothetical, please answer to this question I have to ask again: how would publishing slf4j 2.0 to Orbit would have prevented this issue in m2e? As long as there is no good answer to that, I don't think requesting Platform committers to do extra work for Orbit when better workflows were identified has any chance to be successful. You're asking Platform committers to do basically spend much more time for some tasks, so what's the concrete value to expect in exchange of this extra effort? The slf4j 2.0 vs m2e bug is a concrete opportunity to emphasize how Orbit can avoid issues and increase quality/collaboration and so on. So please use this opportunity to build a case in favor of Orbit by explicating how Orbit would have prevented this issue from happening. If you're not successful in building such a case, then we'll gladly let you join the club of those who think Orbit is not so profitable as it used to and that the future can be brighter without Orbit ;) _______________________________________________
eclipse-pmc mailing list
eclipse-pmc@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/eclipse-pmc
_______________________________________________
eclipse-pmc mailing list
eclipse-pmc@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/eclipse-pmc
_______________________________________________
eclipse-pmc mailing list
eclipse-pmc@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/eclipse-pmc
|
