Drop Support for GHC < 8.6 in Preparation for PrimUnaligned

[andrewthad]

Before GHC 8.6, there are no primitives for unaligned access to byte arrays. These have now been available for a long time, and I have a typeclass and functions for this in the primitive-unaligned library at (Data.Primitive.ByteArray.Unaligned)[https://hackage.haskell.org/package/primitive-unaligned-0.1.1.2/docs/Data-Primitive-ByteArray-Unaligned.html]. The design of the typeclass needs to be changed soon since GHC is gaining primops for unaligned Addr# access, but I'd like to go ahead and start down the path of getting these consolidated. The first step toward that is removing support for GHC < 8.6.

Please keep discussing here about dropping support for GHC 8.0, 8.2, and 8.4. Any discussion about unaligned access to primitive arrays should happen at #409

[raehik]

A C header macro file in bytestring bytestring-cpp-macros.h notes that unaligned byte array access primops were not always safe before GHC 9.2.2:

#define HS_UNALIGNED_ByteArray_OPS_OK \
  MIN_VERSION_base(4,12,0) \
  && (MIN_VERSION_base(4,16,1) || HS_UNALIGNED_POKES_OK)
/*
The unaligned ByteArray# primops became available with base-4.12.0,
but require an unaligned-friendly host architecture to be safe to use
until ghc-9.2.2; see https://gitlab.haskell.org/ghc/ghc/-/issues/21015
*/

(clickable GHC issue link: https://gitlab.haskell.org/ghc/ghc/-/issues/21015 )

Does this change anything for us?