hat Auxfixity.hs:92 error?

When attempting to use hat[0] on a reasonably simple program that compiles fine with ghc I am faced with the following error. Could anyone tell me what may be the cause of the error? Could there be a way in which I could alter my program to a semantically equivalent version for which hat would work? Or am I just using hat in the wrong way? $ hmake -hat quineMcCluskey.lhs hat-trans quineMcCluskey.lhs Fail: AuxFixity.hs:92: Non-exhaustive patterns in case Rather than flood inboxes I've put my program (quineMcCluskey.lhs) online at: http://www.srcf.ucam.org/~ddp23/quineMcCluskey.lhs I suspect someone may recognise the bug though, even without the code? Perhaps this problem would go away with the new version of hat? Kind regards and thanks for any help, Daniel [0] Installed from: hat-2.00-1.i386.rpm -- --|------| Daniel Parry (Caius) (Compsci Part II) ddp23@cam.ac.uk |-------|-- "Nobody will ever need more than 640k RAM!" [Bill Gates, 1981] "Windows 95 needs at least 8MB RAM." [Bill Gates, 1996] -----------------------------------------------------------------------------

Daniel Parry
When attempting to use hat[0] on a reasonably simple program that compiles fine with ghc I am faced with the following error.
$ hmake -hat quineMcCluskey.lhs hat-trans quineMcCluskey.lhs
Fail: AuxFixity.hs:92: Non-exhaustive patterns in case
This is a bug in hat-trans, which appears to be (sort of) fixed in the current CVS version. The next release of Hat is due in a couple of weeks, but if you can't wait, and would prefer not to use CVS, there is a workaround: in your program, replace uses of the constant -1 with (negate 1), excepting patterns of course. Regards, Malcolm

On Mon, Mar 03, 2003 at 12:20:23PM +0000, Malcolm Wallace wrote:
Daniel Parry
writes: When attempting to use hat[0] on a reasonably simple program that compiles fine with ghc I am faced with the following error.
$ hmake -hat quineMcCluskey.lhs hat-trans quineMcCluskey.lhs
Fail: AuxFixity.hs:92: Non-exhaustive patterns in case
This is a bug in hat-trans, which appears to be (sort of) fixed in the current CVS version. The next release of Hat is due in a couple of weeks, but if you can't wait, and would prefer not to use CVS, there is a workaround: in your program, replace uses of the constant -1 with (negate 1), excepting patterns of course.
Thanks, the debugger worked fine with (negate 1) replacing -1. I would just point out though that hat-trail will fail on systems that don't have x-term installed, with: sh: resize: command not found Fail: Command (resize -u) failed I admit most systems will have xterm installed (though mind didn't - I use rxvt exclusively ^_^), but it may be a problem that could easily be ironed out? Kind regards, Daniel -- --|------| Daniel Parry (Caius) (Compsci Part II) ddp23@cam.ac.uk |-------|-- "Let's not escape into mathematics. Let's stay with reality." [Michael Crichton] -----------------------------------------------------------------------------

Daniel Parry
Thanks, the debugger worked fine with (negate 1) replacing -1. I would just point out though that hat-trail will fail on systems that don't have x-term installed, with:
sh: resize: command not found Fail: Command (resize -u) failed
This is mentioned in the documentation at http://www.cs.york.ac.uk/fp/hat/requirements.html All that is really required is any terminal emulator that does ANSI colour codes, but the problem is to discover the user's preference at configuration time. We decided to choose the most commonly available one - xterm - as a de facto standard, likely to be available almost everywhere. Regards, Malcolm
participants (2)
-
Daniel Parry
-
Malcolm Wallace