Malcolm Wallace wrote:
Simon Marlow <[EMAIL PROTECTED]> wrote:
Wed Aug 9 08:38:10 PDT 2006 Simon Marlow <[EMAIL PROTECTED]>
* Remove deprecated Data.FiniteMap and Data.Set interfaces
Is it necessary to remove Data.FiniteMap quite so soon?
Data.FiniteMap was just bloat in the base package. There's nothing
stopping you (or someone else) from packaging Data.FiniteMap
separately, or providing a compatibility API that works with GHC
6.2-6.6. This whole centralised library thing is less of an issue
now that we have packages and Cabal.
OK. Attached is a new module Data.FiniteMap, containing (most of) the
old API, but actually using the newer Data.Map to implement everything.
It is very useful for me, since it means all my old software continues
to work with all of ghc-6.2.2 and ghc-6.4.x and ghc-6.6 (forthcoming).
The only question is where should I put it? At the moment, it is local
to my darcs repo of the base package. Should I commit it back to the
main branch? Should I distribute it as a separate (very tiny) package?
Should I just keep it internally for use in my own projects, since we
should not be publically encouraging the continued use of old APIs?
Distribute it as a separate package, please. We can add a pointer from the 6.6
release notes to help people porting code from 6.4.x.
You may want to add DEPRECATED pragmas too (it's up to you, of course).
Cheers,
SImon
_______________________________________________
Cvs-ghc mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/cvs-ghc
