On 13.12.2016 21:24, David Feuer wrote:

On Dec 13, 2016 1:26 PM, "Oleg Grenrus" mailto:oleg.grenrus@iki.fi> wrote:

I'm not sure we want `transformers` to depend on `containers` either, not even mentioning `vector` and `unordered-containers`. Also there are people stuck with `transformers-0.4` (GHC 7.10).

Good point, although practically speaking, containers currently *can't* depend on transformers, because transformers is not currently a GHC boot package. There'll probably be room for your compat package for a while, but you'll have to chase those bounds :-/.

transformers *is* boot package, `ghc` package depends on transformers directly (since GHC 7.6): https://ghc.haskell.org/trac/ghc/wiki/Commentary/Libraries/VersionHistory $ ghc-pkg describe ghc ... depends: array-0.5.1.1 base-4.9.0.0 binary-0.8.3.0 bytestring-0.10.8.1 containers-0.5.7.1 directory-1.2.6.2 filepath-1.4.1.0 ghc-boot-8.0.1 ghci-8.0.1 hoopl-3.10.2.1 hpc-0.6.0.3 process-1.4.2.0 template-haskell-2.11.0.0 time-1.6.0.1 transformers-0.5.2.0 unix-2.7.2.0

> Current implementations are naive, but more obviously correct. > > More than what?

It seems that implementation for Data.Map is indeed "check size, compare lists", but e.g for vector it uses streaming framework to make things fuse, I still use "check size, compare lists" approach.

I believe the only containers structures with non-boring instances of any of those classes are Data.IntMap, Data.IntSet, and Data.Tree.

IntSet is :: *

Side question: why doesn't Read1 have the ReadPrec-based methods that Read does?

I have been wondering the same