Milestones

[wilzbach]

Currently we use GH milestones to track PRs that should receive a higher priority and be included in the next release, e.g.

https://github.com/dlang/dmd/milestones https://github.com/dlang/druntime/milestones https://github.com/dlang/phobos/milestones ...

I know that there's the "regression" and "blocker" category, but so far no one really bothers checking Bugzilla for what should be fixed in the next release. The idea of the milestone feature would be

• to have an overview of which bugs are super-critical and should be addressed ASAP
• allow other contributors to help out and fix bugs that are important to us

So this is essentially about roadmapping and being able to put the "high-level vision items" into smaller bug chunks that can be incrementally added.

There's one big problem with this though and that's how the labelling is done and more precisely by whom. At the moment I label PRs that

• are passing, but just need approval
• push deprecations
• fix regressions or other critical bugs and aren't controversial

with the next milestone. An example:

https://github.com/dlang/dmd/milestone/22 https://github.com/dlang/dmd/milestone/20?closed=1 https://github.com/dlang/phobos/milestone/19?closed=1 https://github.com/dlang/phobos/milestone/17?closed=1

Anyhow, I don't know how much Bugzilla supports already out of the box, but ideally the list of bugs that should be fixed for the next release (or next "vision period") should be rather prominent and one click away.

[CyberShadow]

I know that there's the "regression" and "blocker" category, but so far no one really bothers checking Bugzilla for what should be fixed in the next release.

At some point, we used to use the open list of regressions as an important signal.

Not sure if we'll actually use this, but I guess it wouln't hurt to enable it & see what happens.