On 25/04/2009, at 17:32, Max Bolingbroke wrote:
2009/4/25 Roman Leshchinskiy <r...@cse.unsw.edu.au>:
On 25/04/2009, at 04:20, Ian Lynagh wrote:
Fri Apr 24 05:47:54 PDT 2009 Ian Lynagh <ig...@earth.li>
* Require a bang pattern when unlifted types are where/let bound;
#3182
FWIW, I find this quite counterintuitive and it also breaks
existing code.
#3182 seems to be an unrelated bug report. What is the real ticket
number?
Yeah, I'm also not convinced about this change. The real ticket is
#2806. http://hackage.haskell.org/trac/ghc/ticket/2806
Thanks Max, I now understand the reasoning behind this change a bit
better. However, I'm still not convinced that it is a good idea,
mainly because it introduces a tight coupling between two otherwise
unrelated language extensions, breaks backwards compatibility and also
perhaps compatibility to other compilers (AFAIK, jhc implements
unboxed types but not bang patterns, for instance). I also never
considered the issue it is intended to solve as a problem, to be
honest. If you understand the semantics of unlifted types it should be
quite obvious. I know I have to be careful about strictness when
working with unlifted types. I don't think I have ever had a bug
related to binding unlifted variables. On the other hand, I'm sure I
will routinely forget to add the bangs and burn cycles on syntax that
is, strictly speaking, not necessary.
As to uniformity and forgetting about patterns for the moment, this
change breaks the following equality:
let v = e1 in e2 = (\v -> e1) e2
This is quite unfortunate, IMO.
Roman
_______________________________________________
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
