Community
Participate
Working Groups
Short version: Anyone inactive in Simultaneous Release aggregation files for the past year will be removed from the access group on 10/14, I propose. Long, interesting, entertaining version: As most readers of this bugzilla component will know, to commit changes to the yearly, simultaneous "common repo" input files, existing committers need to ask for access, give their project affiliation, and they are pretty much "in" -- and they technically access (write) those files by becoming members of the Linux group named 'callisto-dev'. (It is named that since "callisto" was our first simultaneous release). So far, so good. Not too hard. Not much overhead. But ... over the years ... over 6 Simultaneous Releases, to be exact ... the list gets very large, and currently there is no process to remove or reduce people. There is no _pressing_ reason to remove people ... but, just seems like a good hygiene to remove anyone no longer involved ... at least have a process to do so every now and then, such as once a year, or so. One option is to remove everyone and have them re-apply ... JUST KIDDING! :) Here's my proposal ... others welcome. I suggest just about this time every year (shortly after SR1) we see who has been active to the current and "past year" Simultaneous Release aggregation ... and have them automatically continue as-is ... but anyone else, in the callisto-dev list, that has not committed anything in the previous year would be removed from the callisto-dev list --- maybe give a chance to say "wait, leave me on the list". Sound like a reasonable approach? Looking at the cvs change logs, I see that 59 people have committed to "indigo.build" project (or, to the one milestone of "juno.build" project). This goes back to the beginning of Indigo input, August 2010. In contrast, there are 109 people in the callisto-dev group. That leaves 50 people that have not committed to these "input files" since the very beginning of Indigo. I suspect of those 50, some are still "involved" in some way, and might want to stay in the list as backup for someone. But, I'd guess, 20 to 40 are no longer involved enough to care. So, where to go from here? I could publish the list of 50 committers ids, here to this bugzilla, so they'd know if they have to speak up, or otherwise, they will be removed. Is that easiest approach? I sort of hate publishing a bunch of user IDs, but guess they are not secret and easily accessible by many other means. (And, as far as I know, there is no easy way to go from committer id to email, such as to notify them via email ... let me know if there is.) I guess we could literally just remove those that have not committed in the past year, and literally have them re-apply if needed ... but ... a) I could make errors in my change-log-to-group-membership comparisons, and b) in my experience, people wouldn't notice they no longer had access until 15 minutes before their deadline. To give a deadline, I'd like to say to have this resolved by 10/14 ... that is, have identified people to remove on 10/14 ... a few weeks after SR1 and Juno M2 ... so it would not be too stressful if someone was accidentally prematurely removed and had to get re-added before Juno M3 or Indigo SR2. Or ... this not a concern at all, and we should just skip it, let the group grow and grow? (I have been told I am a bit of a worrier :) Thanks for any suggestions. But in the absence of clear alternatives, I'll publish the list of unused access here, give a short while for people to speak up for exceptions or errors ... and then ask webmasters to remove inactives on 10/14.
+1 to remove people that are not involved in simultaneous release since some time. As you suggested, may the webmasters can provide some APIs to send an email from the eclipse portal ? I suspect all the information is there ( id, email )
> One option is to remove everyone and have them re-apply ... JUST KIDDING! :) har har har :)
As someone who would no doubt be on the removal list (and rightly so), that sounds totally reasonable to me.
Created attachment 203717 [details] ids to remove from callisto-dev group on 10/14 This is my computation of users who have not committed anything to "callisto-dev" files, since Indigo began (August, 2010). I could have made mistakes, since I do some "cleanup" (editing) of change logs files before diffing ID lists). So, feel free to say if I've erred. Also, this year, anyone on the list who wants to stay on the list, may, by saying so. I think in future years (when number involved will be much smaller) we'll just remove anyone not committing for previous years, partially justified that "people should know" that if they want to be backups, etc., for others, that they'll have to touch the files at least occasionally. After all, and remaining justification, out of the 59 remaining committers to callisto-dev, surely one could be found that could "pitch in" to commit someone else's changes if a backup was needed. I have made a note in the wiki "contributing to aggregation build" to make note of this removal procedure. http://wiki.eclipse.org/Juno/Contributing_to_Juno_Build#Checkout_CVS_juno.build_project
I didn't check, but assume the names still need to be removed? So, will assign to "webmaster" to do the removal, and then we can close this. Thanks!
The list has been pruned. Thanks for the housecleaning.
Note for future: in the list of ids to be removed, 'genie' was included ... as 'genie' has not committed anything to callisto-dev ... but ... 'genie' needs to be there, for signing purposes (see bug 362436) so, in future, we can't "automatically exclude all ids that have not committed anything" ... there will be at least 'genie' that should not be removed. Just adding as a reminder.