I’ve asked Zubin about this, to which he replied:
A deprecation warning is not a critical bug that warrants the
effort and pain of a whole new release, including costs for
tooling developers, CI systems and general ecosystem fragmentation
over and above the direct costs of producing a release.
I still think a deprecation coming in 9.14.2 is worth it, even if released only a few months from now.
To add to my points before, many folks wait for the .2 release before upgrading to a new major version.
Cheers,
Rodrigo
On 15 Dec 2025, at 09:10, Simon Peyton Jones <simon.peytonjones@gmail.com> wrote:
Is there a reason we can’t have a 9.14.2 release with just the deprecation warning?
Rodrigo can you ask Zubin about this? There are non-trivial overheads associated with a release. But a very minor release like this might have much smaller overheads?
I suppose the goal would be that the deprecation is embodied in a released GHC, much sooner than awaiting for a critical mass of bug-fixes to accumulate.
Simon
Is there a reason we can’t have a 9.14.2 release with just the deprecation warning?
Sounds like we're converging on that compromise. Any objections to accepting the proposal with these provisions?
- Deprecation would start whenever 9.14.2 comes out (probably not for some months)
- the change would take effect with 9.16 in about six month
If not, I'll take this as consensus by December 16.
_______________________________________________
ghc-steering-committee mailing list -- ghc-steering-committee@haskell.org
To unsubscribe send an email to ghc-steering-committee-leave@haskell.org
_______________________________________________
ghc-steering-committee mailing list -- ghc-steering-committee@haskell.org
To unsubscribe send an email to ghc-steering-committee-leave@haskell.org