planning for lts-25

[juhp]

It feels we are at the time when it would be reasonable to release LTS 25 with ghc-9.12.
Also I think we are no longer seriously blocked on 9.14 for Nightly, though a 9.14.2 release would be welcome.

However there is one contentious (or rather serious) bug/regression in 9.12.4:

• https://gitlab.haskell.org/ghc/ghc/-/issues/26936

so seems we really need to wait for 9.12.5.

Anyone aware of any other serious issues?

[alexfmpe]

https://gitlab.haskell.org/ghc/ghc/-/issues/27121

[alaendle]

Perhaps I'm being too conservative, but I'd wait and see if a 9.12.5 version is released. Anyone wanting to upgrade to 9.12 beforehand can use the nightly build without any issues. Besides, the ecosystem for 9.14 seems only partially ready to me – but that's the usual chicken-and-egg problem. Long story short, it's not really important, but if I had to choose, I'd wait.

[juhp]

Yes, I guess we'd be lucky to even get an 9.12.5 RC this month. (I think we agree about waiting for it...)

What do you think is the main lib(s) missing for 9.14 btw? (Anyway I think we invariably end up branching so late it is never really an issue.)

[alaendle]

I think a truly solid 9.12 version is highly desirable for the next LTS release, so our users can switch as smoothly as possible—but I think we've already agreed on that. The package set for 9.12 is really good, so we're just waiting for the compiler.

Regarding the packages for which I'm missing 9.14 support, constraints-extras (obsidiansystems/constraints-extras#61) immediately comes to mind—which, via dependent-sum/map, has a considerable network of dependent packages.

TLDR: From current perspective, let's ship the next major LTS and move nightly to 9.14 as soon as we think the 9.12 compiler is ready (ghc/ghc#26936?).

[juhp]

https://gitlab.haskell.org/ghc/ghc/-/issues/27286 is the ghc-9.12.5 release tracker

[juhp]

https://gitlab.haskell.org/ghc/ghc/-/merge_requests/16097/commits