I just tried GHC HEAD on one of my projects and got ``` test/Main.hs:418:14: warning: [GHC-69797] Type applications in constructor patterns will require the TypeAbstractions extension starting from GHC 9.12. Suggested fix: Perhaps you intended to use TypeAbstractions | 418 | SomeWordN (WordN @n1 i1) == SomeWordN (WordN @n2 i2) = | ``` I assume that the changes have been properly approved by GHC SC, so feel free to skip until "(Not rant)" section. ---- (Rant) I'm quite unhappy when GHC refines language pragmas and forces me to make changes. I have a perfectly valid program using `{-# LANGUAGE TypeApplications #-}`, a part of `GHC2021` standard. Suddenly GHC 9.9 (and presumably future GHC 9.10) decides to change its semantics (why?) and says this is no longer good enough and forces me to add `{-# LANGUAGE TypeAbstractions #-}`. Mind you that the latter has been just introduced in GHC 9.8 and is marked as ["partially implemented"](https://ghc.gitlab.haskell.org/ghc/doc/users_guide/exts/type_abstractions.html), so as a developer I don't even know what I'm opting in. And the warning says that my program will stop compiling with GHC 9.12. This is just a gratuitous breakage out of some religious zeal on behalf of GHC: if it knows what I meant and can compile just fine, please assume `TypeAbstractions` or whatever needed transparently. Breaking existing users for the sake of far-fetched Dependent Haskell (?) benefits is a way to nowhere. ---- (Not rant) Could someone please suggest a migration strategy? How can we maintain backward compatibility, ideally without CPP and warnings, when `TypeAbstractions` has been just introduced in GHC 9.8, but will be mandatory as soon as GHC 9.12?