April 2001 - Hugs-Users

Re: a runhugs sandbox?
by Keith Wansbrough 26 Apr '01

26 Apr '01

> I am using (run)hugs to grade students' homework: > they send some Haskell code (by email), > this gets imported from a test program which is run by `runhugs' > (and its output is mailed back to the sender). > > Now, running arbitrary programs is potentially dangerous, > but Haskell's type system ensures that they don't do nasty IO stuff - > as long as students don't know about unsafePerformIO. > > What is a realiable way to prohibit the usage of such functions? You might try a Unix-style way of doing it... use chroot(1) or chroot(2) to provide an environment within which the code can do what it likes, but be insulated from the rest of the system. Just an idea. I haven't tried it. --KW 8-) -- Keith Wansbrough <kw217(a)cl.cam.ac.uk> http://www.cl.cam.ac.uk/users/kw217/ Cambridge University Computer Laboratory.

3 2

Debug Script?
by David Genziuk 26 Apr '01

26 Apr '01

Is it possible to step through a hugs script, as in C where it is possible to debug by stepping in and evaluate variables? Is there any method to evaluate variables during recursion? Adding watches? Any help would be greatly appreciated, Thank You _________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

2 1

a runhugs sandbox?
by Johannes Waldmann 26 Apr '01

26 Apr '01

Dear all, I am using (run)hugs to grade students' homework: they send some Haskell code (by email), this gets imported from a test program which is run by `runhugs' (and its output is mailed back to the sender). Now, running arbitrary programs is potentially dangerous, but Haskell's type system ensures that they don't do nasty IO stuff - as long as students don't know about unsafePerformIO. What is a realiable way to prohibit the usage of such functions? I tried the brute force method of removing share/hugs/lib/exts/* altogether, but runhugs somehow needs it (why?) (and Random needs it anyway). The next thing would be to disallow any import statements in students' code. But this would as well prohibit the import of "pure" modules (like List). A more far reaching idea would be a "tainted" mode (a la Perl) for (run)hugs (and a "tainted" attribute, defined in the language). Any thoughts? -- -- Johannes Waldmann ---- http://www.informatik.uni-leipzig.de/~joe/ -- -- joe(a)informatik.uni-leipzig.de -- phone/fax (+49) 341 9732 204/252 --

1 0

multi-parametric type class
by Javier Torres Salgado 20 Apr '01

20 Apr '01

Hi, Nowadays I am using Hugs 98 interpreter (2001 version) to finish my final work to get my Computer Master Degree at Instituto Tecnológico de Costa Rica. Next, I have transcribed part of the code from Hugs's Help in Multiple Parameter Classes and Dependencies Section --- module KK where class E a b | a -> b, b -> a where go :: a -> b -- line 5, personal addition back :: b -> Maybe a -- personal addition --- With the previous declaration Hugs gives me the following message: ERROR KK.hs:5 - Haskell 98 does not support dependent parameters. And I tried too: --- module KK where class C a b where -- line 3 ok :: a -> b ko :: b -> Maybe a --- and Hugs gives me the message: ERROR KK.hs:3 - Haskell 98 does not support multiple parameter classes Comment: Hugs 98 User Manual of September 1999, says that the facility of Multiple parameter type classes already is available in Hugs 98. Then, how can I use this? I installed the following files again again (from Web site) * Win32.zip * Win32exes.zip * Hugs98-Feb2001.zip -- Hugs for Windows 32 * Hugs.hlp Thank you for paying attention to this, Javier. ********************************************************************* Po.Box TELEFONOS (PANAMA) Javier Torres Salgado Panama: (507)770-5042 Estafeta Universitaria Costa Rica: (506)551-4336* Universidad Autonoma de Chiriqui (506)551-6503 Chiriqui, Panama (COSTA RICA) E-MAIL Javier Torres Salgado Apdo 13307050 jatosa(a)mail.chiriqui.com Cartago, Costa Rica jtorres(a)ganimedes.cic.itcr.ac.cr *********************************************************************

2 1

Re: WinHelp file
by Sigbjorn Finne 18 Apr '01

18 Apr '01

John.Velman(a)HSC.com writes: > ... long and painful process deleted... > This is all down to a bug/problem with the web browser (Netscape) you're using. My suggestion would be to use a different one or download a .zip'ed up version of hugs.hlp from http://www.dcs.gla.ac.uk/~sof/hugs-hlp.zip hth --sigbjorn

1 0

Re: hugs on w2k
by Sigbjorn Finne 18 Apr '01

18 Apr '01

Shirley Yang Shirley.Yang(a)insearch.edu.au writes: > > Any problems if I install your latest hugs98 (Feb2001) on > Windows 2000 professional or server? Hi, no, I reckon you won't run into any installation issues - Win2k is supported quite well. I haven't tried installing it on Win2k Server, but would be surprised if it adds anything new over Professional as far as Hugs installation and use, is concerned. hth --sigbjorn

1 0

RE: Black Hole's in Hugs?
by pablocardenal＠laplatavive.com 17 Apr '01

17 Apr '01

Hello again, Thanks to Alastair Reid for your excellent response for my previous question !!! The concrete question now is: The order of constraints in a class definition are important? ======================================================= ========= More info can look here ===================== ======================================================= I'm try to look for the infinte loop in my previous example and i can't find it anywhere, but i discover that investing the order of the constranints in a class definition, the infinite loop disappears and program now works fine, i.e.: if i replace >class (Rule e r, SubResult r) => SubRule e r where ... by >class (SubResult r, Rule e r) => SubRule e r where ... the infinite loop error disappears Then, the order of constraints in a class definition are important? Saludos, Pablo_pc Cardenal Student at UNLP, La Plata, Buenos Aires, Argentina.

1 0

RE: Question about hugs´s DEBUG_SHOWSC
by Chris Angus 13 Apr '01

13 Apr '01

is this not the dictionary passing (or whatever)to achieve overloading. > -----Original Message----- > From: dario.bahena(a)correo.unam.mx [mailto:dario.bahena@correo.unam.mx] > Sent: 13 April 2001 06:48 > To: hugs-users(a)haskell.org > Cc: hugs-bugs(a)haskell.org; haskell(a)haskell.org > Subject: Question about hugs´s DEBUG_SHOWSC > > > hi haskellers ... > > I´m trying to use the hugs´s flag DEBUG_SHOWSC. Everything > seems to work > fine, except for one(simple?) detail: > > When you use Int or Double literals, hugs always add extra variables, > related to the fromInt/fromDouble aplication. > > For example: > > f = 666 > > should be translated to: > > f = fromInt 666 > > but hugs returns : > > f = fromInt v35 666 > > What is the var. v35 for???, the function fromInt receives > just one argument. > > > If the function has parameters, the resulting function has > extra-parameters, > which seems to be out-of-place in both sides of the binding ... > here you are a couple of examples: > > the function: > > f x = 666 > > generates the output: > > f o2 o1 = fromInt o2 666 > > And the reported type for f is: Num a => b -> a , which is > inconsistent with > the previous result. > > the function: > > f x = 666 + x > > generates : > > f o2 o1 = (+) o2 (fromInt o2 666) o1 > > which has similar troubles to the previous example. > > What are these extra-variables for? Are they a debug-feature? > what´s the > meaning? > > And finally, the obligated question: how can I avoid them? I > suppose that > I have to take a look at the source code, but: what´d be the > part which > is responsible for this(parser,static,type... no please,compiler)? > > Thanks in advance. > > saludos > dario estepario ... > > ------------------------------------------------- > Obtén tu correo en www.correo.unam.mx > UNAMonos Comunicándonos > > _______________________________________________ > Haskell mailing list > Haskell(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell >

1 0

Question about hugs´s DEBUG_SHOWSC
by dario.bahena＠correo.unam.mx 13 Apr '01

13 Apr '01

hi haskellers ... I´m trying to use the hugs´s flag DEBUG_SHOWSC. Everything seems to work fine, except for one(simple?) detail: When you use Int or Double literals, hugs always add extra variables, related to the fromInt/fromDouble aplication. For example: f = 666 should be translated to: f = fromInt 666 but hugs returns : f = fromInt v35 666 What is the var. v35 for???, the function fromInt receives just one argument. If the function has parameters, the resulting function has extra-parameters, which seems to be out-of-place in both sides of the binding ... here you are a couple of examples: the function: f x = 666 generates the output: f o2 o1 = fromInt o2 666 And the reported type for f is: Num a => b -> a , which is inconsistent with the previous result. the function: f x = 666 + x generates : f o2 o1 = (+) o2 (fromInt o2 666) o1 which has similar troubles to the previous example. What are these extra-variables for? Are they a debug-feature? what´s the meaning? And finally, the obligated question: how can I avoid them? I suppose that I have to take a look at the source code, but: what´d be the part which is responsible for this(parser,static,type... no please,compiler)? Thanks in advance. saludos dario estepario ... ------------------------------------------------- Obtén tu correo en www.correo.unam.mx UNAMonos Comunicándonos

1 0

Black Hole's in Hugs?
by pablocardenal＠laplatavive.com 12 Apr '01

12 Apr '01

Hi, I'm extract this fragment of code from my tesis program. I'm try to reduce more, but i can't. The concrete question is: What does {_Gc Black Hole} mean? ======================================================= ===== Concrete problem here, implementation later ===== ======================================================= I'm execute mainOk and receive the expected result Main> mainOk Dyn Hello World!! (82 reductions, 196 cells) I'm execute main and system stop, not reduction executed (27 reductions in five minutes of wait) Main> main {Interrupted!} (27 reductions, 64 cells) I'm expected Dyn <Hello World !!> I'm break and re run and receive the message Main> main Program error: {_Gc Black Hole} (8 reductions, 51 cells) What does _Gc Black Hole mean? ======================================================= ========== Implementation can look here =============== ======================================================= Tested on Hugs 98 February 2001 in Windows 98 with options -98 +som ======================================================= ================ Class Result ========================= ======================================================= >class Result r where > toString :: r -> String > fromString :: String -> r > toR :: Result r' => r -> r' > fromR :: Result r' => r'-> r ======================================================= ============ Class SubResult ========================== ======================================================= >class Result a => SubResult a where > toHtml :: a -> Html > fromHtml :: Html -> a ======================================================= ========= Html data type ============================== ======================================================= >newtype Html = Html String deriving Show ======================================================= ========= Html are Result and SubResult =============== ======================================================= >instance Result Html where > toString (Html st) = st > fromString st = Html ("<" ++ st ++ ">") > toR (Html st) = toR st > fromR r = toR r >instance SubResult Html where > toHtml = id > fromHtml = id ======================================================= =========== Dyn data type ============================= ======================================================= >data Dyn = forall a . SubResult a => Dyn a >instance Show Dyn where > show (Dyn x) = "Dyn " ++ toString x ======================================================= =========== Dyn are Result and SubResult ============== ======================================================= >instance Result Dyn where > toString (Dyn x) = toString x > fromString st = Dyn st > fromR = toR > toR (Dyn x) = toR x >instance SubResult Dyn where > toHtml (Dyn x) = toHtml x > fromHtml html = Dyn html ======================================================= ============ String are Result and SubResult ========== ======================================================= >instance Result String where > toString = id > fromString = id > fromR = toString > toR = fromString >instance SubResult String where > toHtml st = fromString st > fromHtml html = toString html ======================================================= =============== Rule Class ============================ ======================================================= >class (Result r) => Rule e r where > apply :: e r -> r -> r > rws :: (Result r') => (r -> r') -> e r ======================================================= =============== SubRule Class ========================= ======================================================= >class (Rule e r, SubResult r) => SubRule e r where > rwsS :: (SubResult r') => (r -> r') -> e r ======================================================= ======== TRule data type ============================== ======================================================= >newtype TRule r = TRule (r -> r) ======================================================= ========= TRule are Rule ============================== ======================================================= >instance Result a => Rule TRule a where > apply (TRule f) r = f r > rws f = TRule (\r -> toR (f r)) ======================================================= =========TRule are SubRule ============================ ======================================================= >instance SubResult a => SubRule TRule a where > rwsS f = TRule (\r -> toR $ toHtml (f r)) ======================================================= ========== Examples =================================== ======================================================= >genRule1 :: Rule e r => e r >genRule1 = > rws (\r -> toR r ++ "!!") >genRule2 :: SubRule e r => e r >genRule2 = > rwsS (\r -> toR r ++ "!!") >mainOk:: Dyn >mainOk = apply (genRule1::TRule Dyn) (toR "Hello World") >main :: Dyn >main = apply (genRule2::TRule Dyn) (toR "Hello World") ======================================================== ===================== The end ========================== ======================================================== Saludos Pablo "Yeti" Cardenal from La Plata, Buenos Aires, Argentina

3 2