-
Notifications
You must be signed in to change notification settings - Fork 1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Sort out GitHub issues for gitbox repos #499
Comments
I have gone through and tagged all the open PRs against apache/couchdb as a test of the labelling system. It seems to make sense to me. |
I'm generally +1 to everything here. I'd say we should add a |
Yes, I think you've got exactly the right thing for "next minor release," except that we'd simply edit the label to become "2.2.0" once we knew what the numbering might be. (major.minor.patch right?) Alternately we can simply start with defining 2.1.0, 2.2.0, 3.0.0 etc and just assign things as necessary, and move them out to another milestone (or to no milestone if we can't make a commitment). |
Would it make more sense to use tags for that? I'm not super familiar with the release process, but it seems like we'd want "next major/minor/patch release" tags so the release-maker-person would know which issues are going into the next release (and then the tags could be changed to the release name). Milestones might be nice if we decide that certain issues are going to block a release from being pushed. Does that still happen post-2.0? I'm not sold on anything, just thinking. |
The point of GitHub's milestones is to associate a milestone name, description, and optional due date with one or more issues. Milestones can also be open or closed. Typically I see milestones showing what work remains to get a release out the door - you get a nice little bargraph in the UI, for instance. But milestones could also be feature related. Tags don't give you the bar graph or due date, and the tag space can get cluttered quickly (we already have 30 tags to choose from.) I'm nominally against using tags to coordinate releases as a result. If we mirror what we've done in the past in JIRA, there are 2 scenarios where we've used the "Fixed in release" field that would match the GH Milestone function:
In that way I'm in favour of porting over exactly those 3 scenarios from JIRA "Fixed in Release" to GH Milestones. |
OK, labels are live on this repo. I have only normalized the pre-existing labels on the other top-level repos and not implemented the full scheme on them. Fauxton also had a few of their own labels that mostly match this system, but I'd encourage @garrensmith @michellephung @robertkowalski to complete the process in the future if desired. Thanks to everyone (including the CouchDB PMC) for the feedback! |
We've got 'em, we want to use 'em instead of JIRA. Everyone on the project is +1, we just need a system. Here's some proposed rules; I'm keeping them simple.
Assignees
Self-explanatory.
Tagging
I particularly like what Robin have done with their colouring-and-categorizing system. So, inspired by that, here's the proposal. Remember an issue can (and probably should) have more than one tag.
search
,geo
, etc. in the future. (light blue)Projects
Not useful for us right now.
Milestones
We should be using this for version targeting and mass-updating as necessarily. I.e., bugs get targeted for the next minor release, new features that break the API for the next major release, etc. Think v2.1.0, v3.0.0, etc. (Rathole: v or no v?)
Issue template
Going to start with this one.
PR template
Update to change reference from JIRA to GH Issues. Remind people to include text like "Fixes #472" to link a PR to an issue.
Contributing
We've had requests for this before, should we add one?
Multi-repo
We still have a number of repos that aren't likely to be merged into apache/couchdb, like fauxton. I am suggesting we keep primary issue reporting here (in apache/couchdb) for all of our repos, and reference issues here from PRs in other repos (via #472). We shouldn't disallow issues in other repos, but my guess is that they'll be limited to people 'in the know.'
The text was updated successfully, but these errors were encountered: