June 2002 - Hugs-Bugs

Say�n hugs-bugs ,
by abdyesilkart.com 16 Jun '02

16 Jun '02

1 0

odd Hugs runtime error
by Scott Turner 14 Jun '02

14 Jun '02

The following program should print "e" but gets an execution-time error: Program error: {slist Bool_True} data Set = Leaf Int | Empty selem Empty = False slist Empty = "e" main = do let t = not $ selem Empty if t || t then do let e = Empty putStrLn (slist e) else return () This is Hugs 98 December 2001 built for Solaris. Current settings: +fewuiR -stgl.qQkI -h2500000 -p"%s> " -r$$ -c40 Compatibility : Haskell 98 (+98) Tried it also on Windows 2000 and got an access violation. Scott Turner Member of Technical Staff, Product Development, MKS Inc., www.mks.com Burlington, MA, USA; tel: 781-359-3309; fax: 781-359-3399 Build Better Software

2 1

too much simplification?
by Levent Erkok 13 Jun '02

13 Jun '02

HUGS> :t (== []) flip (==) [] :: Eq a => [a] -> Bool GHCI> :t (== []) forall a. (Eq [a]) => [a] -> Bool I think ghc is doing the right thing. -Levent.

4 4

Weird quantification in error message
by Josef Svenningsson 13 Jun '02

13 Jun '02

Hi! I'm observing very strange error messages from Hugs involving rank-2 types. The whole thing could just be an error in the pretty printer but it is never the less rather disturbing. I'm running the December 2001 release. OK, here's how to reproduce the behaviour. Consider the following (*not* type correct) program: \begin{code} data Plus m1 m2 a = T1 (m1 (Plus m1 m2 a)) | T2 (m2 (Plus m1 m2 a)) | Var a run :: (Functor m1, Functor m2) => (forall a. m1 a -> a) -> (forall a. m2 a -> a) -> Plus m1 m2 a -> a run r1 r2 (T1 m) = run r1 r2 (fmap (run r1 r2) r1 m) run r1 r2 (T2 m) = run r1 r2 (fmap (run r1 r2) r2 m) run r1 r2 (Var a) = a \end{code} Hugs should give a type error for the definition of "run" and does so. But the error message looks really strange. Here's what it looks like: Prelude> :r Reading file "HugsBug.hs": Type checking ERROR "HugsBug.hs":6 - Type error in explicitly typed binding *** Term : run *** Type : (forall a. a d -> d) -> (forall a. b d -> d) -> Plus a b (Plus a b c) -> c *** Does not match : (forall a. a d -> d) -> (forall a. b d -> d) -> Plus a b (Plus a b c) -> Plus a b c *** Because : unification would give infinite type Prelude> Look at the quantification. It quantifies over the wrong type variable. Cheers, /Josef

1 0

install win2000
by Bjorn H Bjornsson 12 Jun '02

12 Jun '02

Hi hugs-group I'll be stuying Haskell this summer, and thus installed Hugs-dec2001 for Windows 2000. If I use the Admin-account for installation and then run Hugs as a less privileged user (which is quite normal for Win2000), I find that Hugs is not fully installed, e.g. shortcuts missing from the Start-menu etc. So it seems that I need to run Hugs with the same user that was used for installation (with installation privileges). If this is intended it's not a big problem for me, but it _really_ should be clearly documented in the download-page. Otherwise, you might want to look at software such as Ghostscript for Windows will gives the option to install for "All users" (who may or may not have installation privileges). With kind regards Björn _________________________________________________________________ MSN Foto är det enklaste sättet att dela eller skaffa papperskopior av dina foton: http://photos.msn.com/support/worldwide.aspx

2 1

Re: hugs on powerpc with linux
by Alastair Reid 11 Jun '02

11 Jun '02

[copied to Hugs-bugs] Stefan observes (in a recent cvs snapshot): > Reading file "/usr/local/hgl/lib/x11/Xlib.hs": > Type checking > ERROR "/usr/local/hgl/lib/x11/Xlib.hs":112 - Type error in application > *** Expression : int32ToInt gc_arg2 > *** Term : gc_arg2 > *** Type : Int32 > *** Does not match : Int32 As I guessed (in private mail), this was my screwup. The confusing error message is because the types involved are really Prelude.Int32 and Int.Int32 The CVS repository is fixed again. Update and hopefully things will work. -- Alastair Reid reid(a)cs.utah.edu http://www.cs.utah.edu/~reid/

1 0

hugs on powerpc with linux
by Stefan Heimann 07 Jun '02

07 Jun '02

Hi! I have tried to install Hugs on a Apple Powerbook G3, running debian. It does not work. First I get: gcc -c -g -O2 input.c input.c: In function `nextLine': input.c:516: warning: comparison is always false due to limited range of data type Then I changed the type of lineBuffer in input.c from *char to *int. The compile error disapperead, instead I got 2 warnings: gcc -c -g -O2 input.c input.c: In function `nextLine': input.c:546: warning: passing arg 2 of `linecmp' from incompatible pointer type input.c:558: warning: passing arg 2 of `linecmp' from incompatible pointer type So I added the two necessary casts. Now compilation went fine, but when I try to start a program, the following happens: Reading file "/usr/local/share/hugs/lib/exts/IORef.lhs": ERROR "/usr/local/share/hugs/lib/exts/IORef.lhs":40 - Empty script - perhaps you forgot the `>'s? There seems to be a problem with if (linecmp(BEGINCODE, (char*)lineBuffer)) { input.c, line 545 and else if (linecmp(ENDCODE, (char*)lineBuffer)) { input.c, line 557 Bye, Stefan -- Stefan Heimann | http://www.stefanheimann.net Brandensteinstr. 5 | http://www.cantaloop.org D-79110 Freiburg | http://cvsshell.sourceforge.net

2 1

div error
by Jurek Piotrowski 06 Jun '02

06 Jun '02

> div 1 (-2) -1 > div (-1) 2 -1 but > div (-1) (-2) 0 > div 1 2 0 > div 0 (-2) 0 Regards, Jerzy Piotrowski CIT UWS, Australia

3 2

RE: div error
by Simon Marlow 06 Jun '02

06 Jun '02

> I did the following check (in ghci): > > Prelude> let check a b = (a`div`b,a == > (a`div`b)*b+(a`mod`b) && ((a`mod`b)>=0) && ((a`mod`b)<b)) > Prelude> check 1 2 > (0,True) > Prelude> check (-1) 2 > (-1,True) > > Thus, for positive divisors it seems OK. > However the result for negative divisors I don't understand: > > Prelude> check 1 (-2) > (-1,False) > Prelude> check (-1) (-2) > (0,False) As I understand it, the condition (a`mod`b) >= 0 isn't necessarily true. This condition is true, however: let check a b = let (d,m) = divMod a b in (d, m, a == d * b + m && abs m < abs b && signum m == signum b) Cheers, Simon

1 0

Todo lo que no sabes y querrias saber de Internet
by Webmaster chirona.com 06 Jun '02

06 Jun '02

1 0