That’s an interesting point Ian. My guess is that the main safeguard we have right now is the commit rights vote. So, hypothetically speaking, if someone were suspicious of this or other cases, they should vote -1 until they were comfortable with the move. In that sense the current EDP Move Review Process seems sound.
My previous direct doing a move is when we decided to move Zest from Mylyn to GEF in 2006 (bug 164387). If I recall, I pushed for the very talented Zest creator to get commit rights on the new Zest component of GEF without directly earning them on GEF, thinking that it would be onerous to end up being moved without commit rights on the target. It made sense and worked in that case, which is why I didn’t question the process when Wayne suggested this process for the P2/Discovery move.
But I still think that we need to fix something here. I had one of the Tasktop guys ask me “why am I getting commit rights on P2 when I didn’t earn them?”. He didn’t require more explanation than a pointer to the Move Review, since the description there is pretty clear. Wayne, assuming that the process is OK and that the vote is sufficient to avoid gaming, I wonder if we need a special procedure and link to EDP, or at least a convention, for committer votes that result from a move review. Also, somethign I didn’t see specified is if the commit rights should not be assigned until after the move review happens, in case anyone from outside the project’s mailing list wants to object or discuss.
Dr. Mik Kersten
Tasktop CEO, Mylyn Lead
From: Ian Bull [mailto:irbull@xxxxxxxxxxxxxxxxx]
Sent: December-08-09 2:12 PM
To: P2 developer discussions
Cc: mik.kersten@xxxxxxxxxxx; p2-dev-bounces@xxxxxxxxxxx
Subject: Re: [p2-dev] committer votes and fumbled fingers.
Mik, Susan, Thanks for taking the time to consider this issue. I personally don't feel "dis'd", but I can see how that could happen.
In this case there is absolutely no problem as both Shawn and Steffen are very strong developers and established members of the Eclipse community (and we would be lucky to have them help out with p2). However, I think this particular policy could be gamed. Do we see fundamental difference between someone committing 6 months worth of patches (both bug fixes and new features) and someone else creating a new enhancement in a separate project and "moving" it? It would seem that a project *could* manipulate this policy to fast track commit status. (Give a developer commit rights on an incubator, have them fix one bug there, move the code).
It's an interesting discussion, clearly not relevant in this case, but interesting none-the-less.
I'm looking forward to the code in the p2 repository :-).
On Tue, Dec 8, 2009 at 1:33 PM, Susan Franklin McCourt <susan_franklin@xxxxxxxxxx> wrote:
Even though I was aware of the move and the commit vote (having participated in calls about it and following the bug report), I must say I hadn't read the specific move review link before casting my vote. It does explain quite clearly that the commit rights are part of the move.
The most important thing to me was that other committers understand the nature of this nomination so it didn't seem like we were subjectively "fast tracking" folks.
"Mik Kersten" ---12/08/2009 12:24:48 PM---Hi Susan,
I see your concern, and was wondering about this myself when we initiated
the move discussions. The oddity being that individuals get commit rights
on a project that they may not have participated in directly to a sufficient
degree. My understanding of it from Wayne's post on bug 295273 is that this
kind of move and commit rights assignment follows the Move Review part of
ve_Review My reading of that is that the P2 project is deciding to give
commit rights to those committers, who have commit rights on another Eclipse
project, because it wants the component in P2, not because those committers
have earned their rights on P2. If that were not the case, I agree that it
would not do right by the others who have earned their commit rights.
> -----Original Message-----
> From: Steffen Pingel [mailto:steffen.pingel@xxxxxxxxxxx]
> Sent: December-08-09 11:05 AM
> To: Mik Kersten; Shawn Minto
> Subject: Fwd: [p2-dev] committer votes and fumbled fingers.
> I agree with Susan's points and don't think it's right that me and
> Shawn are getting commit rights on a project that we have not
> participated in. I think it's worth considering to withdraw and submit
> patches instead of committing directly to the P2 repository. The least
> we should do is to state that we expect the discovery component to be
> part of the incubator and will limit our activity to that one
> particular component.
> ---------- Forwarded message ----------
> From: Susan Franklin McCourt <susan_franklin@xxxxxxxxxx>
> Date: Tue, Dec 8, 2009 at 10:48 AM
> Subject: [p2-dev] committer votes and fumbled fingers.
> To: P2 developer discussions <p2-dev@xxxxxxxxxxx>
> Hi, everyone.
> I felt the need to explain a couple of things with respect to my
> committer votes for Steffen and Shawn.
> I thought I was voting +1 for Daniel as an equinox incubator committer
> (after all, he's already a regular committer), but apparently I
> inadvertently assigned a +1 and "sure" to one of Steffen's
> I was planning a longer explanation for voting +1 for Steffen and Shawn.
> So my votes look a bit confused, one saying "sure" the other with a
> long explanation, and a bunch of +1's after that (after getting
> frustrated with fumbling the portal interface).
> I thought it was important to say that:
> - I've seen patches and bug reports from both Steffen and Shawn in the
> course of my platform UI work, or watching SWT bugs, etc.
> - these guys are talented committers
> - I think it's a bit odd to vote in a committer who hasn't yet
> actively participated in the project itself
> - the Mylyn discovery work fills an important gap in our p2 story and
> has proven itself through adoption by other projects
> - it is best for p2 if we can integrate the work and have a cohesive
> story about how these pieces fit together and why
> - we need those guys to support their code in p2-land, hence they need
> to be committers
> I'm making this lengthy (public) explanation because I don't want to
> "dis" committers like Ian, Thomas, Henrik, Daniel who attended a bunch
> of calls and submitted loads of patches before we even started a
> p2-dev mailing list
> Steffen Pingel
> Committer, http://eclipse.org/mylyn
> Senior Developer, http://tasktop.com
p2-dev mailing list
p2-dev mailing list
R. Ian Bull | EclipseSource Victoria | +1 250 477 7484
http://eclipsesource.com | http://twitter.com/eclipsesource