November 2006 - Hugs-Bugs

[Hugs] #30: Windows and Hugs, fails to complete build of base
by Hugs 21 Mar '07

21 Mar '07

#30: Windows and Hugs, fails to complete build of base ---------------------+------------------------------------------------------ Reporter: neil | Owner: nobody Type: defect | Status: new Priority: blocker | Milestone: next release Component: hugs | Version: 200605 Keywords: | ---------------------+------------------------------------------------------ Under Windows/Mingw32, building hugs and the base package gives the following issue: ..\..\src\ffihugs -98 -P..\..\hugsdir\packages\base; -i"HsBase.h" ..\..\hugsdir\packages\base\System\Process\Internals.hs -Iinclude -I..\..\ghc\includes cbits\execvpe.c Skipping base package This bug is probably related to: > ERROR "C:\Program > Files\WinHugs\packages\base\System\Process\Internals.hs":747 - > Undefined variable "addMVarFinalizer" Ross says: Hmm, it uses addMVarFinalizer under Windows, and Hugs doesn't have that. Have to think about that one. This build failure causes the Regex module not to be built, since it comes after in the list of builds. -- Ticket URL: <http://hackage.haskell.org/trac/hugs/ticket/30> Hugs <http://www.haskell.org/hugs/> Hugs 98, an interpreter for Haskell

1 3

[Hugs] #43: WinHugs installer does not provide "Install for All Users" option
by Hugs 21 Mar '07

21 Mar '07

#43: WinHugs installer does not provide "Install for All Users" option ------------------------+--------------------------------------------------- Reporter: briansmith | Owner: neil Type: defect | Status: new Priority: major | Milestone: Component: winhugs | Version: 200609 Keywords: | ------------------------+--------------------------------------------------- The WinHugs installer does not provide an "Install for All Users" option. If the user is logged in with a limited account (which is the recommended practice in Windows), they will have to use the "Run as" command to run WinHugs as an administrator. But, the installer will then install the shortcuts to WinHugs in the Administrator account's Startup menu, but not any other users'. -- Ticket URL: <http://hackage.haskell.org/trac/hugs/ticket/43> Hugs <http://www.haskell.org/hugs/> Hugs 98, an interpreter for Haskell

1 3

[Hugs] #50: Disparity between GHCi and Hugs: Outstanding context : Show b
by Hugs 21 Mar '07

21 Mar '07

#50: Disparity between GHCi and Hugs: Outstanding context : Show b -----------------------------------+---------------------------------------- Reporter: guest | Owner: nobody Type: defect | Status: new Priority: minor | Milestone: next release Component: hugs | Version: 200503 Keywords: top level overloading | -----------------------------------+---------------------------------------- Possibly Hugs bug, or possible GHCi feature? {{{ showln = (++ "\n") . show main = interact (showln . length) }}} In GHCi: {{{ $ ghci Z.hs *Main> :t showln showln :: Int -> [Char] }}} In Hugs (Version: March 2005): {{{ ERROR "Z.hs":1 - Unresolved top-level overloading *** Binding : showln *** Outstanding context : Show b }}} I'm not sure which is right. -- Don -- Ticket URL: <http://hackage.haskell.org/trac/hugs/ticket/50> Hugs <http://www.haskell.org/hugs/> Hugs 98, an interpreter for Haskell

1 2

[Hugs] #35: WinHugs: foreground/background colours not from Windows
by Hugs 21 Mar '07

21 Mar '07

#35: WinHugs: foreground/background colours not from Windows ---------------------+------------------------------------------------------ Reporter: neil | Owner: neil Type: defect | Status: new Priority: minor | Milestone: Component: winhugs | Version: 200605 Keywords: | ---------------------+------------------------------------------------------ I just started using the new WinHugs GUI you developed and maintain, and I have a small issue with its appearance. I usually try to tone down all bright colors on the desktop, and hence the default window background color on my machines are not white but some shade of light gray. The issue is then that text portions seem to always have a white background while the rest of the WinHugs window conforms to the global color setting. I attach a window-shot for illustration. I suggest that it would be much nicer if the global background color setting was honored by the written text portion as well. --- Good point! The issue is to do with escape character, you can programatically change the background color of the text by outputting some magic, and this means that by default the background gets "forced" to white, rather than getting its default setting. I think the solution is to interpret white = the users background colour, black = the users foreground colour, and honour all other colours. I'll hack that up in the next few days and the next release should have this fixed. -- Ticket URL: <http://hackage.haskell.org/trac/hugs/ticket/35> Hugs <http://www.haskell.org/hugs/> Hugs 98, an interpreter for Haskell

1 2

[Hugs] #37: Equal Ints are not equal
by Hugs 18 Mar '07

18 Mar '07

#37: Equal Ints are not equal ---------------------+------------------------------------------------------ Reporter: guest | Owner: nobody Type: defect | Status: new Priority: blocker | Milestone: Component: hugs | Version: current Keywords: | ---------------------+------------------------------------------------------ Hi, a few days ago I wrote a program which handles my own integers. Then I check my own implementation against the normal Int implementation in Hugs. Thereby I wonder about inequality of Ints which are definitely equal. This is a test program I wrote after I recognized that hugs has probably a bug in the inequality check of Ints. data MInt = Zero | Succ MInt | Pred MInt deriving Show tn :: Int -> MInt tn x | x<0 = Pred (tn (x+1)) tn 0 = Zero tn (n+1) = Succ (tn n) ti :: MInt -> Int ti Zero = 0 ti (Succ x) = 1+(ti x) ti (Pred x) = (ti x) -1 testi :: (MInt -> MInt -> MInt) -> (Int -> Int -> Int) -> Int -> Int -> Bool testi f g x y = (ti (f (tn x) (tn y))) /= (g x y) myMul x y = tn ((ti x) * (ti y)) test = [(x,y,ti (myMul (tn x) (tn y)),x * y)| x<-[-100..100], y<-([-100..(-1)]++[1..100]),(testi myMul (*) x y) ] the list "test" should be empty in any case but unfortunately hugs returns this (copy of the console): __ __ __ __ ____ ___ _________________________________________ || || || || || || ||__ Hugs 98: Based on the Haskell 98 standard ||___|| ||__|| ||__|| __|| Copyright (c) 1994-2003 ||---|| ___|| World Wide Web: http://haskell.org/hugs || || Report bugs to: hugs-bugs(a)haskell.org || || Version: November 2003 _________________________________________ Haskell 98 mode: Restart with command line option -98 to enable extensions Type :? for help Main> test [(-100,80,-8000,-8000),(-100,99,-9900,-9900), (-99,58,-5742,-5742),(-99,83,-8217,-8217), (-98,71,-6958,-6958),(-98,86,-8428,-8428), (-97,46,-4462,-4462),(-97,87,-8439,-8439), (-96,76,-7296,-7296),(-95,88,-8360,-8360),(-94,73,-6862,-6862) ERROR - Control stack overflow Main> I have checked this program with Hugs on different plattforms: Gentoo Linux x86 64-Bit and Suse Linux 32-Bit. On both plattforms the same bug occurs with different Ints for each run (i.e. the list contains different tuples). I have also checked my program with other Haskell implementations and there it returns an empty list as expected. Best regards, Stephan Swiderski (swiderski(a)informatik.rwth-aachen.de) P.S.: We have tested Sep2006 with similar results: __ __ __ __ ____ ___ _________________________________________ || || || || || || ||__ Hugs 98: Based on the Haskell 98 standard ||___|| ||__|| ||__|| __|| Copyright (c) 1994-2005 ||---|| ___|| World Wide Web: http://haskell.org/hugs || || Bugs: http://hackage.haskell.org/trac/hugs || || Version: September 2006 _________________________________________ Haskell 98 mode: Restart with command line option -98 to enable extensions Type :? for help Main> :e Main> test [(-100,90,-9000,-9000),(-98,40,-3920,-3920),(-97,63,-6111,-6111),(-97,69,-6693,- 6693) ERROR - Control stack overflow -- Ticket URL: <http://hackage.haskell.org/trac/hugs/ticket/37> Hugs <http://www.haskell.org/hugs/> Hugs 98, an interpreter for Haskell

2 7

Undefined variable bug
by John Tromp 27 Feb '07

27 Feb '07

I just found out that I can circumvent the problem by adding a line import Char to my loaded file. regards, -john

2 1

maximum [] complains about an empty list
by Neil Mitchell 24 Nov '06

24 Nov '06

Hi, In GHC: Prelude> maximum [] *** Exception: Prelude.maximum: empty list In Hugs: Prelude> maximum [] Program error: pattern match failure: foldl1 (Ord_max instOrd_v29) [] Obviously this is because foldl1 is used to define maximum, but it would be nice if the equation maximum [] = error "pattern match failure in maximum" was issued, otherwise people might end up grepping for foldl1 which doesn't ever occur in their code. Thanks Neil

1 0

[Hugs] #47: Hugs segfaults on black hole.
by Hugs 14 Nov '06

14 Nov '06

#47: Hugs segfaults on black hole. ---------------------------------+------------------------------------------ Reporter: guest | Owner: nobody Type: defect | Status: new Priority: major | Milestone: next release Component: hugs | Version: 200609 Keywords: segfault black hole | ---------------------------------+------------------------------------------ Hugs.Base> let s = 1 + s in s zsh: illegal hardware instruction (core dumped) hugs -- Ticket URL: <http://hackage.haskell.org/trac/hugs/ticket/47> Hugs <http://www.haskell.org/hugs/> Hugs 98, an interpreter for Haskell

1 2

[Hugs] #46: Records aren't working properly when declared using 'newtype' syntax.
by Hugs 14 Nov '06

14 Nov '06

#46: Records aren't working properly when declared using 'newtype' syntax. ----------------------------+----------------------------------------------- Reporter: Elifant | Owner: nobody Type: defect | Status: new Priority: major | Milestone: next release Component: hugs | Version: 200609 Keywords: newtype record | ----------------------------+----------------------------------------------- Test code: {{{ newtype Test = Test { var :: String } main = print (var x) where x = Test { var = "a" } }}} Output: {{{ " Program error: pattern match failure: instShow_v16_v1443 (Test_Test "a") }}} Records update is also broken: {{{ newtype Test = Test { var :: String } main = print (var x{var="b"}) where x = Test "a" }}} Output: {{{ " Program error: pattern match failure: instShow_v16_v1443 (Test_Test "b") }}} It I change "newtype" with "data", all is OK. -- Ticket URL: <http://hackage.haskell.org/trac/hugs/ticket/46> Hugs <http://www.haskell.org/hugs/> Hugs 98, an interpreter for Haskell

1 1

[Hugs] #48: record patterns fail on newtypes
by Hugs 14 Nov '06

14 Nov '06

#48: record patterns fail on newtypes --------------------+------------------------------------------------------- Reporter: ross | Owner: nobody Type: defect | Status: new Priority: major | Milestone: Component: hugs | Version: 200609 Keywords: | --------------------+------------------------------------------------------- Given the declaration {{{ newtype T = C { f :: String } }}} Hugs exhibits the following misbehaviour: {{{ Newtype> case C "abc" of C { f = v } -> v "bc" Newtype> case undefined of C {} -> True Program error: Prelude.undefined }}} The non-record versions are fine: {{{ Newtype> case C "abc" of C v -> v "abc" Newtype> case undefined of C _ -> True True }}} The internal representation of `newtype`s elides the data constructor, but the record pattern-matching code treats them like `data` types (see also #46). -- Ticket URL: <http://hackage.haskell.org/trac/hugs/ticket/48> Hugs <http://www.haskell.org/hugs/> Hugs 98, an interpreter for Haskell

1 0