On 25/01/07, Brent Fulgham <[EMAIL PROTECTED]> wrote:
3. I don't recall which test was changed to its less-efficient counterpart, but as I recall this was due to some sniping that the lazy analysis was avoiding doing some of the work. After comparing the tests, we ended up agreeing and the Haskell implementation was modified.
How can it be that some of the work is avoided if the final result can be printed and is correct? If there are requirements to produce specific intermediate values, it may be a good idea to add those to the required output. - Cale _______________________________________________ Haskell mailing list Haskell@haskell.org http://www.haskell.org/mailman/listinfo/haskell
