floor(Inf), floor (NaN)

Simon Peyton-Jones simonpj@microsoft.com
Fri, 6 Apr 2001 04:20:25 -0700


Further to my mail re the Haskell 98 report, there is one loose end.

| 2.  George Russell pointed out that floor(Inf) and floor(NaN) are not=20
| defined in Standard Haskell; indeed GHC and Hugs have=20
| different behaviour.=20

This point led to an inconclusive discussion back in March 2000.
Fergus Henderson and George Russell were the main protagonists.

I'm too ignorant to have an opinion, but given the lack of conclusion
I have taken no action.  If anyone wants to revive the discussion, go
ahead, but remember the extremely conservative nature of this revision!

Simon