+1 This was already discussed when I noticed the the Generic instances were missing. https://ghc.haskell.org/trac/ghc/ticket/8797 Edward suggested that: "The Generic and Generic1 instances clearly belong with their definitions. They also deserve Typeable and Data." (...) "FWIW- The committee is in favor of adding the Generic, Generic1, Data and Typeable instances along with Num for Sum and Product." But I guess part of that was lost in the discussion, and did not end up on the patch. My understanding was that it was supposed to. There also seemed to be a consensus in raising a new issue (that I then forgot to raise :( ) to do this for all newtypes in base... João 2014-05-27 7:38 GMT+01:00 Michael Snoyman :

On Tue, May 27, 2014 at 9:31 AM, Roman Cheplyaka wrote:

...

I was surprised to find out that wrappers like Last do not have Typeable instances. Let's add them.

+1

_______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries

* Herbert Valerio Riedel [2014-05-27 10:36:46+0200]

On 2014-05-27 at 08:31:14 +0200, Roman Cheplyaka wrote:

...

I was surprised to find out that wrappers like Last do not have Typeable instances. Let's add them.

Fyi, Data.Monoid has been given a `{-# LANGUAGE AutoDeriveTypeable #-}` in

http://git.haskell.org/ghc.git/commitdiff/1d1ff77aaa09efaddc8cfe0dcf92d67632...

which should take care of not forgetting a 'deriving Typeable' in the future for any new entities defined in Data.Monoid.

Cool! I take it this is already solved in HEAD, then? Any reason not to enable AutoDeriveTypeable for the whole base (and other core packages)? Roman

| You can't enable it everywhere in base due to module-import cycles (you | need to import Data.Typeable before you can AutoDeriveTypeable a | module), but other than that, the plan is to enable it wherever possible | in base. Is someone going to execute on that intention? Thanks! Generally speaking, for every data type in base - either AutoDeriveTypeable should be on - or there should be a standalone deriving clause in Typeable (so that the instance isn't an orphan) Simon