- Glasgow-haskell-users

efficiency of UArray
by Hal Daume III 15 May '02

15 May '02

can we expect a function like: sum [x*y | (x,y) <- zip (elems v) (elems u)] to be as efficient as, say: sum = 0 for i=1, n sum = sum + v[i] * u[i] ? Basically, will any intermediate lists be created here? -- Hal Daume III "Computer science is no more about computers | hdaume(a)isi.edu than astronomy is about telescopes." -Dijkstra | www.isi.edu/~hdaume

1 0

RE: using error x as a placeholder
by Simon Marlow 15 May '02

15 May '02

[ moved to glasgow-haskell-users(a)haskell.org ] > The timing results were: > > mkArray1: 11.42u 0.79s 0:13.47 90.6% > mkArray2: 24.55u 2.31s 0:30.12 89.1% > > Which is actually *slower*. Any ideas why? (These were > compiled with ghc > 5.02.3 -O2 -fvia-c -fall-strict) What are the results without -fall-strict? (-fall-strict is an old experimental flag and almost certainly doesn't do anything reasonable. We've de-documented it in the HEAD). Cheers, Simon

2 1

RE: Haddock installing problems
by Simon Marlow 15 May '02

15 May '02

[ moved to glasgow-haskell-users(a)haskell.org, which is slightly more appropriate than haskell ] > Looking the output produced by ./configure, I noticed that: > > checking for ghc... /PROGS/GHC5023/BIN/ghc > checking version of ghc... /PROGS/GHC5023/BIN/ghc: not found > > I wonder why haddock thinks my ghc dir is > "/PROGS/GHC5023/BIN/ghc" when it > should be "/progs/ghc5023/bin/ghc". Perhaps that's why it > can't find perl: > ghc itself was not found. Pathnames aren't case-sensitive on Win32 are they? So why should that make a difference? What is your cygwin mount setup? What is mounted on /? Cheers, Simon

1 0

RE: Replacing the Prelude
by Simon Peyton-Jones 14 May '02

14 May '02

Ashley writes | > I was hoping to do something similar for 'do' notation by redefining | > (>>), (>>=) etc., but unfortunately GHC is quite insistent | that 'do' notation quite specifically refers to GHC.Base.Monad Dylan replies | I'm surprised that ghc uses the fromInteger and fromRational | that are in scope; I thought there was general agreement that | it should use the Prelude.from{Integer,Rational} in scope. Ashley is referring to a GHC extension. Normally, GHC uses Prelude.fromInteger etc, regardless of what is in scope. But if you say -fno-implicit-prelude, GHC will instead use whatver fromInteger is in scope. (This is documented in the manual.) Ashley's question, as I understand is whether something similar could be done for monads. Answer: I think so, without much bother. I'm beavering away on a Haskell workshop paper at the moment, but ping me in a fortnight to do it. Simon

1 0

Replacing the Prelude
by Ashley Yakeley 14 May '02

14 May '02

I have recently been experimenting writing code that replaces large chunks of the Prelude, compiling with -fno-implicit-prelude. I notice that I can happily redefine numeric literals simply by creating functions called 'fromInteger' and 'fromRational': GHC will use whatever is in scope for those names. I was hoping to do something similar for 'do' notation by redefining (>>), (>>=) etc., but unfortunately GHC is quite insistent that 'do' notation quite specifically refers to GHC.Base.Monad (i.e. Prelude.Monad, as the Report seems to require). I don't suppose there's any way of fooling it, is there? I was rather hoping 'do' notation would work like a macro in rewriting its block, and not worry about types at all. I accept that this might be a slightly bizarre request. There are a number of things I don't like about the way the Prelude.Monad class and 'do' notation are set up, and it would be nice to be able to experiment with alternatives. -- Ashley Yakeley, Seattle WA

2 1

RE: Prelude.catch vs. Exception.catch
by Ashley Yakeley 14 May '02

14 May '02

At 2002-05-14 03:57, Simon Marlow wrote: >It turns out that the compiler "bug" is really just the compiler being a >bit loose with the IO monad - it freely translates the original >definition of evaluate using 'seq' into the slightly less strict version >by pushing the 'seq' through the state lambda of the IO monad (this only >happens for the IO monad, and strictly speaking it's a deviation from >the semantics but it has important performance benefits for IO code). Under what circumstance would the compiler do this to my code? I really need to know about any deviation from the semantics. >so your catch can be defined in terms of the current catch like so: > > catch' a h = a `seq` (catch a h) Oh yeah. >what's the motivation for this change? Well I think it's more intuitive. The documentation (sec. 5.12.3) says this: The simplest version is catch. It takes a single argument, runs it, and if an exception is raised the "handler" is executed passing the value of the exception. Otherwise, the result is returned as normal. ...leading to the question, how do you 'run' bottom? But since I can do what I want with your "catch'" function, it's not really that important. -- Ashley Yakeley, Seattle WA

1 0

RE: Prelude.catch vs. Exception.catch
by Simon Marlow 14 May '02

14 May '02

> At 2002-05-14 02:58, Simon Marlow wrote: > > >I must admit I can't think of any compelling reasons for the change, > >other than the fact that this is functionality that we don't > have at the > >moment, and therefore might be useful. Opinions? > > I need a function that does this: > > evaluate :: a -> IO a > evaluate _|_ = fail something > evaluate a = return a > > The idea is that you can take something that might be bottom > and safely > handle it in the IO monad, with the "bottomness" gone. This is what > Exception.evaluate currently does, and I think that's correct. Ok, I'll change the definition of evaluate to reflect the slightly different semantics. It turns out that the compiler "bug" is really just the compiler being a bit loose with the IO monad - it freely translates the original definition of evaluate using 'seq' into the slightly less strict version by pushing the 'seq' through the state lambda of the IO monad (this only happens for the IO monad, and strictly speaking it's a deviation from the semantics but it has important performance benefits for IO code). > I think the behaviour of Exception.catch is wrong. I think it should > behave like this: > > catch _|_ cc = _|_ > catch (failure ex) cc = cc ex > catch (success a) cc = success a > > ...whereas it actually behaves like this: > > catch _|_ cc = cc something > catch (failure ex) cc = cc ex > catch (success a) cc = success a so your catch can be defined in terms of the current catch like so: catch' a h = a `seq` (catch a h) what's the motivation for this change? Cheers, Simon

1 0

RE: Prelude.catch vs. Exception.catch
by Ashley Yakeley 14 May '02

14 May '02

At 2002-05-14 02:58, Simon Marlow wrote: >I must admit I can't think of any compelling reasons for the change, >other than the fact that this is functionality that we don't have at the >moment, and therefore might be useful. Opinions? I need a function that does this: evaluate :: a -> IO a evaluate _|_ = fail something evaluate a = return a The idea is that you can take something that might be bottom and safely handle it in the IO monad, with the "bottomness" gone. This is what Exception.evaluate currently does, and I think that's correct. I think the behaviour of Exception.catch is wrong. I think it should behave like this: catch _|_ cc = _|_ catch (failure ex) cc = cc ex catch (success a) cc = success a ...whereas it actually behaves like this: catch _|_ cc = cc something catch (failure ex) cc = cc ex catch (success a) cc = success a -- Ashley Yakeley, Seattle WA

1 0

RE: Prelude.catch vs. Exception.catch
by Simon Marlow 14 May '02

14 May '02

> >This is bizarre: the definition of evaluate in Exception is > exactly the > >one you gave above, yet they behave differently. You may > have uncovered > >a compiler bug, I'll look into it. > > I might ask which is correct: according to the rules for seq, > "evaluate' > undefined" should be bottom, but we want "Expression.evaluate > undefined" > to be a failing IO action. > > I think the compiler is correct but the definition given in the > documentation is wrong. Well, given the compiler is compiling the definition from the documentation and giving the wrong semantics, I'd say the compiler is definitely wrong :-) But the question of whether Exception.evaluate should have a different semantics is interesting. So to be clear about this, there are three possible definitions of evaluate: evaluate a = return a -- not strict enough evaluate a = a `seq` return a -- too strict (?) evaluate a = IO $ \s -> a `seq` (# s, a #) -- just right :) I had to write out the third one in full, because we don't have any combinators that provide this functionality elsewhere (at least, I can't think of any). I must admit I can't think of any compelling reasons for the change, other than the fact that this is functionality that we don't have at the moment, and therefore might be useful. Opinions? Cheers, Simon

1 0

RE: Prelude.catch vs. Exception.catch
by Ashley Yakeley 14 May '02

14 May '02

At 2002-05-14 02:24, Simon Marlow wrote: >This is bizarre: the definition of evaluate in Exception is exactly the >one you gave above, yet they behave differently. You may have uncovered >a compiler bug, I'll look into it. I might ask which is correct: according to the rules for seq, "evaluate' undefined" should be bottom, but we want "Expression.evaluate undefined" to be a failing IO action. I think the compiler is correct but the definition given in the documentation is wrong. -- Ashley Yakeley, Seattle WA

1 0