Re: [Haskell-cafe] Problems with strictness analysis?

I tried your example in GHC 6.10 and isum appears to work fine. The type of 10000000 gets defaulted to Integer, a specialized version of isum for Integer is then created, the strictness analyzer determines that isum is strict in s, and the code generator produces a loop. (If you want to look at the assembly code, it's easier if you make the type Int.) Thanks for checking it. I'll be curious to see any improvements the new version brings, but this sounds promising already.

Gergely -- http://www.fastmail.fm - The way an email service should be