[For the benefit of those not on the cvs-hugs list] I've closed out all errors reported by the testsuite except those mentioned in the previous 2 bug reports by the simple expedient of deciding that Hugs was behaving correctly in all cases and updating the expected output files accordingly. One of the suspicious behaviours I decided to accept is, perhaps, a little controversial: the builtin printer prefixes data constructor names with the name of the type that the constructor belongs to. This shows up in error messages and if you turn off "use 'show' to display results" Prelude> case Just 'a' of { Nothing -> 'b' } Program error: {v1294 (Maybe_Just 'a')} Prelude> :set -u Prelude> Just 'a' Maybe_Just 'a' It is not completely clear to me whether: 1) the typename prefix is a good idea 2) the underscore should be replaced by some other punctuation such as '.' or '@' or '/' or whatever. I'd be interested in hearing people's opinions. Historical note: I think the prefix was originally added to improve the quality of heap profiling output. Since then it has been extended to cover more and more kinds of entity (member functions, classes, types, whatever). -- Alastair Reid reid@cs.utah.edu http://www.cs.utah.edu/~reid/