Re: [Haskell-cafe] Solved but strange error in type inference

Thanks all, I finally give up to put Ord in the LegGram type. What was annoying me was that @legsome@ was in fact an instance of a class I defined. So I changed its signature to make it depend on Ord. That is not very nice, since at first glance, there could be implementations which does not require "nt" to be ordered. But as in some other method of this class I need it, I guess it isn't that annoying. I still do find the reported error strange, but thanks to your remarks, I can remove some typing annotation in the body of the function (which I forgot to remove and used while I was debugging my types) to avoid this ugly redundancy (I guess I could also remove the global type signature, but I like to keep them).