Re: Subject: CFV: New JDK 10 Committer: Erik Österlund (original) (raw)

Daniel D. Daugherty daniel.daugherty at oracle.com
Thu Jun 22 17🔞38 UTC 2017

On 6/22/17 3:26 AM, Thomas Schatzl wrote:

Vote: veto

Justification: I looked through the changes and I do not believe they meet the (informal?) goal of having eight significant commits. Here is an overview of the contents of these changes. Feel free to correct me. trivial: http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/f1ad14991f86 (add volatile + fix compilation) http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/c2ecbb9ee746 (add volatile + fix compilation) http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/4d009502987b (add volatile + fix compilation)

The above three changesets are specifically described as being split off from http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/8fcdd3cc8da0 below because the fix for:

8033552: Fix missing missing volatile specifiers in CAS operations in GC code

had gotten too large.

Personally, I would not flag appropriately researched additions of "volatile" as either trivial or borderline, but I would count all four changesets together as one.

I'm not going to argue about what "significant" means. That's like arguing religion and we're not going to get anywhere with that. The fact of the matter is that OpenJDK (and not Oracle) requires 8 significant changesets for some unspecified definition of "significant". Thomas has legitimately raised a "veto" because the list of changesets does not match his definition of "significant". I have to respect Thomas' opinion as an OpenJDK (R)eviewer.

If I had taken the time to read each changeset and counted them, I would count the changesets only slightly differently and I wouldn't reach the magic 8. However, I don't typically do that unless I have some past experience with the person that makes me nervous about the person being granted the role of Committer.

Full disclosure: I also know Erik and am working with him on a project. I believe he meets what I'm looking for in a Committer and even a Reviewer. Please notice the difference between "Reviewer" and "(R)eviewer"; I use the parenthesized version when I'm talking about the OpenJDK role... More than once, I have had Erik provide articulate reasons and carefully written rationale for why some piece of code I was working on was wrong. He's not afraid to step forward and offer an opinion. He's not afraid to be wrong and has no problem with evolving his opinion as things change.

Yes, I would have issued:

Vote: yes

because I wouldn't have taken the time to go check the changesets if I had not seen Thomas' veto, but I cannot and will not fault the folks that voted "veto".

Yes, Erik has some very significant work coming to a repo near you! He'll easily cross the "magic 8" administrative barrier and I'm enjoying being part of that process.

I hope that Erik realizes that all this discussion is not about his work; it is about OpenJDK processes and the indisputable fact that those processes are applied by humans in the typical human fashion: inconsistently in the general case.

Dan

http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/bbf76069d7fc (remove obsolete code)

borderline: http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/bb5c32e2d31a (fix compilation) http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/8fcdd3cc8da0 (add volatile + fix compilation, this one is a bit larger than the others) significant: http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/5793813a17dd (build changes) http://hg.openjdk.java.net/jdk10/hs/rev/55bf5464b0e6 (build changes) http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/8a4e011d99be http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/5398ffa1a419 I counted the build changes as significant because I feel unable to judge them. That sums up to, including the ones I considered borderline, as six significant commits. I do know Erik, and I know that he is trustworthy to be committer, and I know that he has lots of good changes about to be committed, but in addition to this understanding some formal requirements must be met (or at least very close to) at the time of nomination. This is, in my view not the case here at this time. We have in the past also been pretty strict about getting the Committer role, and I think it would devalue every others' hard work to attain that goal. I also, just a few months ago, privately looked through commits of somebody else as trustworthy in the same situation, and I asked him to come back after another two commits. I would like to ask the same in this case. Thanks, Thomas

On Thu, 2017-06-22 at 09:51 +0200, Stefan Karlsson wrote: I hereby nominate Erik Österlund to JDK 10 Committer.

Erik is a member of the Garbage Collection team in Oracle. He likes to rewrite and fix bugs in the OrderAccess and Atomics layers, loves template metaprogramming, and is currently working on creating a GC Barrier Interface as a part of JEP 304 - Garbage-Collector Interface [3][4] Votes are due by 2017-07-06T09:50+02:00. Only current JDK 10 Committers [1] are eligible to vote on this nomination. Votes must be cast in the open by replying to this mailing list. For Lazy Consensus voting instructions, see [2]. Stefan Karlsson Contributions: http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/bb5c32e2d31a http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/5793813a17dd http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/f1ad14991f86 http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/c2ecbb9ee746 http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/8fcdd3cc8da0 http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/4d009502987b http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/8a4e011d99be http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/5398ffa1a419 http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/bbf76069d7fc (missing attribution) http://hg.openjdk.java.net/jdk10/hs/rev/55bf5464b0e6

[1] http://openjdk.java.net/census [2] http://openjdk.java.net/projects/#committer-vote [3] http://openjdk.java.net/jeps/304 [4] http://mail.openjdk.java.net/pipermail/hotspot-gc-dev/2017-April/0197 41.html

More information about the jdk10-dev mailing list