April 2001 - Hugs-Bugs

problems????????????? error
by Mark Dacoron 23 Apr '01

23 Apr '01

I am getting an error when trying to run this specific line in hugs the line is " half [1..35] where half x = take (length x 'div' 2) x" and the error that i am getting is "Improperly terminated character constant" I do not know what is going on ???? please help Mark Dacoron _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com

2 1

isAlphanum
by Ryszard Kubiak 22 Apr '01

22 Apr '01

Dear Hugs Maintainers, I would like to ask if it is done for purpose that the function isAlpnanum is called isAlphaNum in Hugs. Here is what the February 2001 version of Hugs tells us: Prelude> :t isAlphanum ERROR - Undefined variable "isAlphanum" Prelude> :t isAlphaNum isAlphaNum :: Char -> Bool The report on the library Char says that the function should be called isAplhanum. Also, Hugs manual mentions it in that form. Various libraries, like Regexp, rely on this statement and one has to manipulate their sources to get them work with Hugs. I would also like to ask when the library Directory is expected to be included into Hugs' distribution. Best Regards, Rysiek

2 2

Re: Haskell 98 non-conformance in qualifiers
by Sigbjorn Finne 22 Apr '01

22 Apr '01

Alastair Reid reid(a)cs.utah.edu writes: > > The following program (from the testsuite) is not legal Haskell'98 but > is accepted by Hugs. The error is that you can't use the qualifier M > inside M (at least, not without first explicitly importing M). > > tests/static/mod75.hs: > --!!! Qualifying with local module name > module M where > f x = M.f x That's legal Haskell 98 and should be accepted; see Section 5.5.1 of the Report; bullet 1. --sigbjorn

1 0

background colour on Hugs home page
by Alastair Reid 21 Apr '01

21 Apr '01

The Hugs homepage used to explicitly set the background colour to white. A couple of years back, this got removed with the result that anyone viewing the home page with a browser which uses grey as the default background colour (i.e., users of Netscape) see a page that just looks crap. Can someone please add an explicit background colour back in? -- Alastair Reid reid(a)cs.utah.edu http://www.cs.utah.edu/~reid/

1 0

Closed out some errors in the testsuite
by Alastair Reid 21 Apr '01

21 Apr '01

[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(a)cs.utah.edu http://www.cs.utah.edu/~reid/

1 0

Haskell98 nonconformance: contexts on member functions
by Alastair Reid 21 Apr '01

21 Apr '01

The H98 report explicitly says (section 4.3.1) that the type of member functions may only add constraints to the type variables which are local to the member. Hugs accepts this (illegal program from the testsuite) without reporting an error. tests/static/mod39.hs: --!!! Illegal constraints on member funs module M where class C a where f :: Eq a => a If someone can do one of the following, I'll update the testsuite (so it isn't reported as an error) accordingly: 1) Add this to the list of Hugs' non-conformance to the standard. 2) Make this report an error in Haskell 98 mode. 3) Make this report an error in both Haskell 98 and Hugs mode -- Alastair Reid reid(a)cs.utah.edu http://www.cs.utah.edu/~reid/

1 0

Haskell 98 non-conformance in qualifiers
by Alastair Reid 21 Apr '01

21 Apr '01

The following program (from the testsuite) is not legal Haskell'98 but is accepted by Hugs. The error is that you can't use the qualifier M inside M (at least, not without first explicitly importing M). tests/static/mod75.hs: --!!! Qualifying with local module name module M where f x = M.f x Looking at the code (static.c:findQualifier) it is clear that this "bug" is the intended behaviour. If someone (e.g., Johan) could add this to the list of known errors, I'll make the test suite stop reporting this as a bug. -- Alastair Reid reid(a)cs.utah.edu http://www.cs.utah.edu/~reid/

1 0

RE: Syntax for implicit parameters
by Simon Peyton-Jones 21 Apr '01

21 Apr '01

I only added 'with' because I did not want to steal *two* new keywords. One is bad enough! I proposed using 'let' (not dlet), with the '?' to distinguish dynamic from lexical bindings, but did not achieve consensus. Lack of consensus => the status quo stays. My order of preference: 1. [happy]. Use 'let' 2. [consent]. Use 'dlet' or 'with' 3. [hate] Use both 'dlet' and 'with' Would the Hugs folk be willing to adopt (2)? Simon | -----Original Message----- | From: Alastair Reid [mailto:reid@cs.utah.edu] | Sent: 18 April 2001 18:31 | To: hugs-bugs(a)haskell.org; glasgow-haskell-bugs(a)haskell.org | Subject: Syntax for implicit parameters | | | | Some months ago, there was talk about making sure GHC and Hugs use the | same syntax for implicit parameters and (most importantly) that that | syntax should not introduce the keyword "with". | | As far as I can see (from looking at both parsers and trying | examples), this discussion has not been acted on. Hugs seems to | allow: | | dlet ?x = 'a' in ?x + 1 | ?x + 1 with ?x = 'a' | | and GHC 5.0 only seems to support: | | ?x + 1 with ?x = 'a' | | Can the GHC people, the Hugs people and the implicit parameter | designers come to some sort of agreement and implement the result? | | | -- | Alastair Reid reid(a)cs.utah.edu http://www.cs.utah.edu/~reid/ _______________________________________________ Glasgow-haskell-bugs mailing list Glasgow-haskell-bugs(a)haskell.org http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

4 5

maybe a bug in hugs version February 2001
by paul.stoeber＠in.stud.tu-ilmenau.de 20 Apr '01

20 Apr '01

Prelude says: data Either a b = Left a | Right b deriving (Eq, Ord, Read, Show) ^^^^ Interactive session: Prelude> Left 3 ERROR - Cannot find "show" function for: *** Expression : Left 3 *** Of type : Either Integer a Test script: > data Foo a b = Bar a | Baz b > showFoo (Bar x) = "Bar: " ++ show x > showFoo (Baz x) = "Baz: " ++ show x Interactive session: Main> showFoo (Bar 3) ERROR - Unresolved overloading *** Type : Show a => [Char] *** Expression : showFoo (Bar 3)

1 0

Precompiled binaries for Hugs on MacOS X
by Atze Dijkstra 19 Apr '01

19 Apr '01

Hi, this is not a bugreport but a message to let you know I did make a precompiled version of Hugs for MacOS X (no other appropriate email address being available), as an installable package. See http://www.cs.uu.nl/~atze/Download/Hugs98-Feb2001.pkg.sit (and/or http://www.cs.uu.nl/~atze/Programming/) I did test the install on a clean machine. Feel free to include it on the Hugs download page. -- - Atze - Atze Dijkstra, Dept. of Computer Science, Utrecht University /|\ Tel.: +31-30-2534093/1454 | WWW : http://www.cs.uu.nl/~atze / | \ Fax : +31-30-2513971 | Email: atze(a)cs.uu.nl /--| \ atze.dijkstra(a)hetnet.nl / |___|

2 1