- Glasgow-haskell-users

Re: efficiency question
by Kirsten Chevalier 11 Feb '02

11 Feb '02

On Sat, Feb 09, 2002 at 12:01:02PM -0500, glasgow-haskell-users-request(a)haskell.org wrote: > > > I'd say that's because in the second case you also got to apply the (,), > > besides the (+)/(-) constructor during the transversing... > > Am I right? > > opss... I meant to write: the (,) constructor besides the (+)/(-)... > J.A. > I'd guess that it's not just that you have to apply the (,) constructor -- it also has to do with the fact that the tuples it's constructing here are boxed. -- Kirsten Chevalier * krc(a)cs.berkeley.edu * Often in error, never in doubt "When nothing remains of me, / Not a particle, / I shall sparkle in the footnote of an article." -- Daniel Aaron http://www.cs.berkeley.edu/~krc/

4 4

RE: Increasing stack size
by Simon Marlow 11 Feb '02

11 Feb '02

> I'm having trouble adjusting the stack size, viz: > > + cacheprof +RTS -K8M -RTS gcc -I. -I. -c /tmp/ghc5664.s -o > PrelTup.c_o > Stack space overflow: current size 1048576 bytes. > Use `+RTS -Ksize' to increase it. The -K option definitely works, at least on simple examples I've tried here. It looks like the option somehow isn't getting through to the compiled program; is it perhaps being gobbled up by an intermediate stage (Makefile, script, whatever)? Cheers, Simon

2 1

Increasing stack size
by Nicholas Nethercote 11 Feb '02

11 Feb '02

Hi, I'm having trouble adjusting the stack size, viz: + cacheprof +RTS -K8M -RTS gcc -I. -I. -c /tmp/ghc5664.s -o PrelTup.c_o Stack space overflow: current size 1048576 bytes. Use `+RTS -Ksize' to increase it. I'm trying to increase it! The program was compiled with GHC 5.02.2. Any ideas why the stack size isn't increasing? Thanks. -- Nick Nethercote njn25(a)cam.ac.uk

1 0

RE: efficiency question
by Julian Seward (Intl Vendor) 11 Feb '02

11 Feb '02

The transformations that the compiler does, especially with -O, are sufficiently complex that trying to second-guess the performance effects of source changes is difficult. You'd get a much better handle on this by reading the post-optimised program, which you can get with -ddump-simpl. Good luck! This is not for the faint of heart. J | -----Original Message----- | From: Jorge Adriano [mailto:jadrian@mat.uc.pt] | Sent: Sunday, February 10, 2002 7:18 PM | To: glasgow-haskell-users(a)haskell.org | Subject: Re: efficiency question | | | On Sunday 10 February 2002 18:48, Kirsten Chevalier wrote: | > I'd guess that it's not just that you have to apply the (,) | constructor -- | > it also has to do with the fact that the tuples it's | constructing here are | > boxed. | | could you elaborate a little more on that (boxed / unboxed) | or provide a link | on the subject? | | thanks | J.A. | _______________________________________________ | Glasgow-haskell-users mailing list | Glasgow-haskell-users(a)haskell.org | http://www.haskell.org/mailman/listinfo/glasgow-haskell-users |

1 0

socket question
by Hal Daume III 11 Feb '02

11 Feb '02

suppose i want to write a stupid ftp client, i want to connect to the ftp server, wait for it to give it's intro stuff (welcome to blah, username: ) and then it wants input. how do i do something like this? so far, i have (don't laugh, this is the first socket haskell program i've written. ever.) > connect cfg = withSocketsDo $ > do h <- connectTo (remoteAddr cfg) (Service (remoteType cfg)) > hWaitForInput h 1000 > s <- readWhileAvailable h > return s > > readWhileAvailable h = > do b <- hReady h > if b > then do c <- hGetChar h > s <- readWhileAvailable h > return (c : s) > else do return [] but this doesn't seem to ever halt :) if i get rid of the hWaitForInput, it returns immediately with nothing. if i replace hWaitForInput...readWhileAvailable with "s <- getContents h ; return (take 10 s)" i get the beginning of the intro message (so it is connecting properly). please someone help me :) -- Hal Daume III "Computer science is no more about computers | hdaume(a)isi.edu than astronomy is about telescopes." -Dijkstra | www.isi.edu/~hdaume

2 1

RE: GHC on Win32 strangeness
by Simon Peyton-Jones 11 Feb '02

11 Feb '02

I've updated the instructions on win32 building, but they havn't got to the web site yet. You have to say ./configure --host=i386-unknown-mingw32 --with-gcc=/mingw/bin/gcc Without this, things go badly arwy. For the --with-gcc part you need to install a mingwin bundle from http://www.mingw.org Simon | -----Original Message----- | From: Manuel M. T. Chakravarty [mailto:chak@cse.unsw.edu.au] | Sent: 11 February 2002 03:24 | To: glasgow-haskell-users(a)haskell.org | Cc: brk(a)jenkon.com | Subject: GHC on Win32 strangeness | | | Hi! | | Bryn <brk(a)jenkon.com> is trying to compile C->Haskell with | GHC on Win32, but ran into some problems. Firstly, | configure says, | | $ ./configure | loading cache ./config.cache | checking host system type... i686-pc-cygwin | checking target system type... i686-pc-cygwin | checking build system type... i686-pc-cygwin | checking for Cygwin environment... (cached) no | checking for gcc... (cached) gcc | [..] | | So, it says no Cygwin, but identifies the system as | i686-pc-cygwin. AFAIK, GHC uses mingw32 these days, doesn't | it? How can this be reliably tested for? (To exclude | -package posix on that platform.) | | Secondly, are there any issues with symlinks? I append a | sequence of commands that seems to indicate that GHC doesn't | find a file, because it is symlinked. | | Any ideas? | | Cheers, | Manuel | | -=- | | Administrator@BRK ~/hipar/build/ghc5/base/admin | $ file BaseVersion.hs | BaseVersion.hs: symbolic link to ../../../../base/admin/BaseVersion.hs | | So, change it to an absolute: | | Administrator@BRK ~/hipar/build/ghc5/base/admin | $ ln -s /home/hipar/base/admin/BaseVersion.hs BaseVersion.hs | | Administrator@BRK ~/hipar/build/ghc5/base/admin | $ ls | BaseVersion.hs Common.hs Config.hs Makefile | | Administrator@BRK ~/hipar/build/ghc5/base/admin | $ file BaseVersion.hs | BaseVersion.hs: symbolic link to /home/hipar/base/admin/BaseVersion.hs | | The build fails the same way, copy the command line and | restrict it to just the file we changed, to be sure: | | Administrator@BRK ~/hipar/build/ghc5/base | $ ghc -M -syslib lang -optdep-f -optdep.depend -iadmin | -ierrors -igeneral -igra phs -istate -isyms -isyntax -isysdep | admin/BaseVersion.hs | ghc.exe: file `admin/BaseVersion.hs' does not exist | _______________________________________________ | Glasgow-haskell-users mailing list | Glasgow-haskell-users(a)haskell.org | http://www.haskell.org/mailman/listinfo/glasgow-| haskell-users |

1 0

GHC on Win32 strangeness
by Manuel M. T. Chakravarty 11 Feb '02

11 Feb '02

Hi! Bryn <brk(a)jenkon.com> is trying to compile C->Haskell with GHC on Win32, but ran into some problems. Firstly, configure says, $ ./configure loading cache ./config.cache checking host system type... i686-pc-cygwin checking target system type... i686-pc-cygwin checking build system type... i686-pc-cygwin checking for Cygwin environment... (cached) no checking for gcc... (cached) gcc [..] So, it says no Cygwin, but identifies the system as i686-pc-cygwin. AFAIK, GHC uses mingw32 these days, doesn't it? How can this be reliably tested for? (To exclude -package posix on that platform.) Secondly, are there any issues with symlinks? I append a sequence of commands that seems to indicate that GHC doesn't find a file, because it is symlinked. Any ideas? Cheers, Manuel -=- Administrator@BRK ~/hipar/build/ghc5/base/admin $ file BaseVersion.hs BaseVersion.hs: symbolic link to ../../../../base/admin/BaseVersion.hs So, change it to an absolute: Administrator@BRK ~/hipar/build/ghc5/base/admin $ ln -s /home/hipar/base/admin/BaseVersion.hs BaseVersion.hs Administrator@BRK ~/hipar/build/ghc5/base/admin $ ls BaseVersion.hs Common.hs Config.hs Makefile Administrator@BRK ~/hipar/build/ghc5/base/admin $ file BaseVersion.hs BaseVersion.hs: symbolic link to /home/hipar/base/admin/BaseVersion.hs The build fails the same way, copy the command line and restrict it to just the file we changed, to be sure: Administrator@BRK ~/hipar/build/ghc5/base $ ghc -M -syslib lang -optdep-f -optdep.depend -iadmin -ierrors -igeneral -igra phs -istate -isyms -isyntax -isysdep admin/BaseVersion.hs ghc.exe: file `admin/BaseVersion.hs' does not exist

2 1

Re: efficiency question
by Jorge Adriano 08 Feb '02

08 Feb '02

> I'd say that's because in the second case you also got to apply the (,), > besides the (+)/(-) constructor during the transversing... > Am I right? opss... I meant to write: the (,) constructor besides the (+)/(-)... J.A.

1 0

efficiency question
by Hal Daume III 08 Feb '02

08 Feb '02

define test1 l = let s1 = foldr (+) 1 l s2 = foldr (-) 1 l in (s1, s2) test2 l = let s = foldr (\x (a,b) -> (x+a,x-b)) (1,1) l in s why is test1 so much faster than test2 for long lists l (eg [1..1000000])? replacing foldr with foldl makes it faster (of course), but test2 is still much slower. i *expected* test2 to be much faster because you're only traversing the list once. presumably the two elements "a" and "b" in test2 could be put in registers and i'd imagine test2 should be faster (it certainly would be if written in c). - hal -- Hal Daume III "Computer science is no more about computers | hdaume(a)isi.edu than astronomy is about telescopes." -Dijkstra | www.isi.edu/~hdaume

2 1

ANNOUNCE: GHC 5.03.20020204 snapshot release
by Julian Seward (Intl Vendor) 08 Feb '02

08 Feb '02

There have been a number of significant improvements made to GHC since the 5.02 sources were branched off the main GHC development tree, but we're not quite ready to make a full 5.04 release yet. However, we're keen to get feedback on the new features from people that don't have easy access to CVS or don't have the time & patience to build GHC from scratch, so we're making a snapshot release of the current GHC, in source and binary form. There are NO GUARANTEES as to the stability of this release, although we do know it has passed our basic three-stage bootstrap test and run the regression tests successfully. In some cases the documentation hasn't been fully updated to reflect the new features yet. Briefly, the changes relative to 5.02.2 are: - The type system now supports arbitrary rank polymorphism, given appropriate type annotations. - Heap profiling has had a major overhaul and now supports retainer profiling and biographical profiling ala nhc98. - Major improvements to the native code generators. You can now compile any and all code through them, including the Prelude. - The FFI syntax has been updated to match the latest version of the FFI Haskell 98 Addendum. - newtypes support deriving *any* class for which the underlying type is also an instance. - Linear implicit parameters: a highly experimental feature. - GHCi has several new commands: - ':show bindings' to list the bindings made on the command line, - ':show modules' to show which modules are loaded, - ':module' has been enhanced as per the discussion on the mailing list (syntax is still experimental - feedback welcome). - ':browse' similar to Hugs' :browse command. And many other minor changes & bugfixes. At the moment there is just the sources and a binary build for x86-linux available. We may add builds for other if they become available. You can get it from http://haskell.cs.yale.edu/ghc/download_ghc_snapshot.html Enjoy! The GHC Development Team.

2 1