th-abstraction-0.6.0.0

[alaendle]

th-abstraction-0.6.0.0 (changelog) (Ryan Scott ryan.gl.scott@gmail.com @RyanGlScott) is out of bounds for:

• [ ] boomerang-1.4.9 (>=0.4 && < 0.6). Jeremy Shaw jeremy@n-heptane.com @stepcut. Used by: library
• [ ] cleff-0.3.3.0 (>=0.2 && < 0.6). Xy Ren xy.r@outlook.com @re-xyr. Used by: library
• [ ] generics-sop-0.5.1.3 (>=0.4 && < 0.6). Andres Löh andres@well-typed.com @kosmikus. Used by: library
• [ ] linear-generics-0.2.2 (>=0.5 && < 0.6). Arnaud Spiwack arnaud.spiwack@tweag.io @aspiwack. Used by: library
• [x] th-extras-0.0.0.6 (>=0.4 && < 0.6). Grandfathered dependencies. Used by: library
• [x] universe-some-1.2.1 (>=0.4.2.0 && < 0.6). Grandfathered dependencies. Used by: library

[arybczak]

I made a revision with updated bounds for effectful-th.

[serras]

kind-generics-th-0.2.3.3 uploaded with relaxed version bounds.

[Topsii]

I think it makes sense to point out that th-abstraction-0.6.0.0 contains breaking changes that will not show unless you compile with GHC 9.8 or later.

From the changelog:

Because BndrVis is a synonym for () on pre-9.8 versions of GHC, this change is unlikely to break any existing code, provided that you build it with GHC 9.6 or earlier. If you build with GHC 9.8 or later, on the other hand, it is likely that you will need to update your existing code. Here are some possible ways that your code might fail to compile with GHC 9.8, along with some migration strategies: ...

[tomjaguarpaw]

Ah, that's pretty critical information, thanks @Topsii. It seems to violate (what I thought was) the point of th-abstraction, however.

[gelisam]

Bumped in recursion-schemes-5.2.2.4 revision 3.

It seems to violate (what I thought was) the point of th-abstraction, however.

I don't think the point of th-abstraction is that if your code compiles against th-abstraction with one version of ghc, then your code will also compile against th-abstraction with every other version of ghc; I think the point is that th-abstraction does the CPP dance around TemplateHaskell so you don't have to. It is easier, but unfortunately not trivially-easy, to write code against a single version of th-abstraction which works with multiple versions of ghc, than it is to write code against the TemplateHaskell API which works with multiple versions of ghc.

[ysangkok]

@alaendle:

• aeson is fixed
• bifunctors is fixed
• bound is fixed
• free is fixed
• lens is fixed
• structs is fixed

[alaendle]

@ysangkok I've updated the issue to only show the remaining packages.

[ysangkok]

I think this can be closed as Nightly is now on tasty-1.5. As was done with aeson-2.2.