From oleg.grenrus at iki.fi  Thu May 14 18:56:30 2020
From: oleg.grenrus at iki.fi (Oleg Grenrus)
Date: Thu, 14 May 2020 21:56:30 +0300
Subject: Release 3.4 planning
Message-ID: <95bad923-47ca-46c3-c204-edcd29af4c17@iki.fi>

Dear cabal developers and users

As GHC-8.12 release plan was announced [1].
I suggest that we cut the 3.4 branch at the end of June,
i.e. after the first GHC-8.12 alpha release.

I opened an issue for discussion: [2]

https://github.com/haskell/cabal/issues/6757

---

I renamed previously existing 3.4 milestone to "Considered for 3.4". [3]
At the moment there are 64 open issues.

I will go though them ASAP.

- if there is an assignee ping them, whether they are working on the issue
   - if positive, we'll add it to 3.4 milestone
- assign few unassigned issues to myself, which I think I could complete
   - I already started work on the last mile/yard/foot of public 
multi-libraries.

At the end of May (in two weeks), I'll remove "Considered for 3.4"
issues.

The goal is that the issues in release milestones have someone working
on them (i.e. they are assigned to someone).

The reasoning for above is that we most likely won't resolve all of the 
60 issues,
and we have  "priority: low" and "priority: high" labels
to hint "how soon we should start working on them".

Please bring other issues (not in "Considered for 3.4") milestone in
issues #6757 [2], if you will work on them.
Especially if they are clear tasks, which don't require bikeshedding.

---

Also the special highlight: We removed sandboxes for cabal-install-3.4
https://github.com/haskell/cabal/issues/6445

Related to the issued of the `cabal install --lib` which is not exactly 
bug free.
That might prevent transition to v2-build for someone.

There is an issue about separating that functionality into a subcommand
for package environment management. The comment [4] sketches the design.
Please read though the issue and feel free to ask questions like "how I
would..." if you have any.  This way we can make good initial design.
(I'm spoiled by cabal-env design, but I haven't found an issue
which won't fit into its way of working).
Unfortunately I'd leave this feature out of 3.4 as we don't have clear and
agreed design.

- Oleg

[1]: https://mail.haskell.org/pipermail/ghc-devs/2020-May/018851.html
[2]: https://github.com/haskell/cabal/issues/6757
[3]: 
https://github.com/haskell/cabal/issues?q=is%3Aopen+is%3Aissue+milestone%3A%22Considered+for+3.4%22
[4]: https://github.com/haskell/cabal/issues/6481#issuecomment-620865817