May 2003 - Glasgow-haskell-users

RE: -package data in ghc-5.04.3
by Simon Marlow 07 May '03

07 May '03

> This refers the issues of `data' package (FiniteMap), > user package setting > > The Manual tells about `base' to replace `data'. > > 1. > If I put to Main.hs > > import FiniteMap > main = putStr $ shows (fmToList $ listToFM $ zip "abc" "abc") "\n" > > , then ghc -package data --make Main > works. > And ghc -package base --make Main > reports > ghc-5.04.3: chasing modules from: Main > ghc-5.04.3: can't find module `FiniteMap' ... The FiniteMap module from package 'data' is now available as Data.FiniteMap from package 'base' (there's no need to explicitly request 'base', it's always available). > 2. > Makefile sets my user package as it was with 5.02 : > ... > packDocon = Package {name = "\"docon\"", \ > import_dirs = ["\"$(e)\""], \ > source_dirs = ["\"$(e)\""], \ > library_dirs = ["\"$(e)\""], \ > hs_libraries = ["\"HSdocon\""], \ > extra_libraries = [], \ > include_dirs = [], \ > c_includes = [], \ > package_deps = [\"data\"], \ > extra_ghc_opts = [$(extraPackageOpts)], \ > extra_cc_opts = [], \ > extra_ld_opts = [] } > ... > > ghc-5.04.3 reports: dependency `data' doesn't exist This is a problem that will be fixed in the next release. For now, just omit the dependency. > Renaming data <- base gives similar report. > After changing it to package_deps = [], > > it reports: > > --------------------------------- > if [ ! -d /home/mechvel/docon/2.06/docon/source/export ]; > then mkdir /home/mechvel/docon/2.06/docon/source/export; fi > ghc-pkg -f /home/mechvel/docon/2.06/docon/source/docon.conf -l > > echo Package {name = "\"docon\"", > import_dirs = > ["\"/home/mechvel/docon/2.06/docon/source/export\""], > source_dirs = > ["\"/home/mechvel/docon/2.06/docon/source/export\""], > library_dirs = > ["\"/home/mechvel/docon/2.06/docon/source/export\""], > hs_libraries = ["\"HSdocon\""], extra_libraries = [], > include_dirs = [], c_includes = [], package_deps = [], > extra_ghc_opts = ["\"-package data\"", > "\"-fno-warn-overlapping-patterns\"", "\"-fwarn-unused-binds\"", > "\"-fwarn-unused-matches\"", "\"-fwarn-unused-imports\"", > "\"-fglasgow-exts\"", "\"-fallow-overlapping-instances\"", > "\"-fallow-undecidable-instances\"" ], extra_cc_opts = [], > extra_ld_opts = [] } > | > ghc-pkg -f /home/mechvel/docon/2.06/docon/source/docon.conf -a > > Reading package info from stdin... done. > Expanding embedded variables...done. > cannot find `libHSdocon.a' on library path The error is self-explanatory: it can't find libHSdocon.a on the path given in the package specification. It looks like you might have some extra quotation marks in there. Also, I wouldn't put '-package data' in extra_ghc_opts. I think that is unlikely to work. Cheers, Simon

1 0

-package data in ghc-5.04.3
by Serge D. Mechveliani 07 May '03

07 May '03

Dear GHC team, I am trying to build my DoCon program under ghc-5.04.3, built from RPM And the installation Makefile for ghc-5.02 does not work for 5.04.3. Could you, please, tell me how to correct it? Thank you in advance for the help. ----------------- Serge Mechveliani mechvel(a)botik.ru ----------------------------------------------------------------- This refers the issues of `data' package (FiniteMap), user package setting The Manual tells about `base' to replace `data'. 1. If I put to Main.hs import FiniteMap main = putStr $ shows (fmToList $ listToFM $ zip "abc" "abc") "\n" , then ghc -package data --make Main works. And ghc -package base --make Main reports ghc-5.04.3: chasing modules from: Main ghc-5.04.3: can't find module `FiniteMap' ... 2. Makefile sets my user package as it was with 5.02 : ... packDocon = Package {name = "\"docon\"", \ import_dirs = ["\"$(e)\""], \ source_dirs = ["\"$(e)\""], \ library_dirs = ["\"$(e)\""], \ hs_libraries = ["\"HSdocon\""], \ extra_libraries = [], \ include_dirs = [], \ c_includes = [], \ package_deps = [\"data\"], \ extra_ghc_opts = [$(extraPackageOpts)], \ extra_cc_opts = [], \ extra_ld_opts = [] } ... ghc-5.04.3 reports: dependency `data' doesn't exist Renaming data <- base gives similar report. After changing it to package_deps = [], it reports: --------------------------------- if [ ! -d /home/mechvel/docon/2.06/docon/source/export ]; then mkdir /home/mechvel/docon/2.06/docon/source/export; fi ghc-pkg -f /home/mechvel/docon/2.06/docon/source/docon.conf -l echo Package {name = "\"docon\"", import_dirs = ["\"/home/mechvel/docon/2.06/docon/source/export\""], source_dirs = ["\"/home/mechvel/docon/2.06/docon/source/export\""], library_dirs = ["\"/home/mechvel/docon/2.06/docon/source/export\""], hs_libraries = ["\"HSdocon\""], extra_libraries = [], include_dirs = [], c_includes = [], package_deps = [], extra_ghc_opts = ["\"-package data\"", "\"-fno-warn-overlapping-patterns\"", "\"-fwarn-unused-binds\"", "\"-fwarn-unused-matches\"", "\"-fwarn-unused-imports\"", "\"-fglasgow-exts\"", "\"-fallow-overlapping-instances\"", "\"-fallow-undecidable-instances\"" ], extra_cc_opts = [], extra_ld_opts = [] } | ghc-pkg -f /home/mechvel/docon/2.06/docon/source/docon.conf -a Reading package info from stdin... done. Expanding embedded variables...done. cannot find `libHSdocon.a' on library path make: *** [init] Error 1

1 0

RE: import PrelAddr
by Simon Marlow 06 May '03

06 May '03

> I'm rewriting some old code,originally written for ghc-4.08.2 . > > 1)Now, when I compile with ghc-5.02 , fails: > > ghc -cpp -fglasgow-exts -O -fvia-C -syslib lang -syslib concurrent -c > foo.hs > > foo.hs:47: > failed to load interface for `PrelAddr': > Could not find interface file for `PrelAddr' > gmake: *** [foo.o] Error 1 > > > 2)if commented, > --import PerlAddr > > ghc -cpp -fglasgow-exts -O -fvia-C -syslib lang -syslib concurrent -c > foo.hs > > foo.hs:112: Type constructor or class not in scope: `Addr' > > foo.hs:514: Data constructor not in scope: `A#' > > gmake: *** [foo.o] Error 1 > > > 3)In doc available I'm recommended to import Foreign > > import Foreign > > ghc -cpp -fglasgow-exts -O -fvia-C -syslib lang -syslib concurrent -c > foo.hs > > foo.hs:511: Data constructor not in scope: `A#' > > gmake: *** [foo.o] Error 1 The Addr type is now deprecated, and has been replaced by the Ptr type. The 'A#' constructor belonged to the Addr type - the code shouldn't have been using the internal representation. I suggest you change it to use published interfaces instead because it will end up being more portable to future versions of GHC. Also, the -syslib option will be removed at some point; nowadays -package is the right option to use. Cheers, Simon

1 0

import PrelAddr
by Rafael Martinez Torres 06 May '03

06 May '03

I'm rewriting some old code,originally written for ghc-4.08.2 . 1)Now, when I compile with ghc-5.02 , fails: ghc -cpp -fglasgow-exts -O -fvia-C -syslib lang -syslib concurrent -c foo.hs foo.hs:47: failed to load interface for `PrelAddr': Could not find interface file for `PrelAddr' gmake: *** [foo.o] Error 1 2)if commented, --import PerlAddr ghc -cpp -fglasgow-exts -O -fvia-C -syslib lang -syslib concurrent -c foo.hs foo.hs:112: Type constructor or class not in scope: `Addr' foo.hs:514: Data constructor not in scope: `A#' gmake: *** [foo.o] Error 1 3)In doc available I'm recommended to import Foreign import Foreign ghc -cpp -fglasgow-exts -O -fvia-C -syslib lang -syslib concurrent -c foo.hs foo.hs:511: Data constructor not in scope: `A#' gmake: *** [foo.o] Error 1 What should I import to use unboxed types and constructors ? You see the "-fglasgow-exts" .

1 0

RE: Haddock code block problems
by Simon Marlow 06 May '03

06 May '03

> I'm using the latest Haddock release (0.4) and seem to > be having a lot of trouble getting line breaks to appear > in code blocks. I've tried it so many ways that if I > were to enumerate them all this would be a very long email. > I can't find any consistency in when it decides to work > (occasionally) and when it doesn't (usually). Here's > some examples.. > > This gives no line breaks: > {-$CStuff > @ > test f [] = 0 > test f (x:xs) = 0 > @ > -} > > Neither does this: > {-$CStuff > >test f [] = 0 > >test f (x:xs) = 0 > -} This is a result of the fact that Haddock removes the <p>..</p> around single paragraphs to eliminate extra whitespace around short documentation comments. In the same vein it turns a single code paragraph, normally <pre>..</pre>, into <tt>..</tt>, which has the side effect of removing line breaks. On reflection that seems to be the wrong thing to do, so I'll change the behaviour for the next release. > This gives a parse error:-( > {-$CStuff > @ > test f [] = 0 > test f (x:xs) = 0 > @ > >test f [] = 0 > >test f (x:xs) = 0 > -} because there's no paragraph break between these paragraphs. A paragraph break is two consecutive newlines. > This doesn't, but still no line breaks. > {-$CStuff > >test f [] = 0 > >test f (x:xs) = 0 > @ > test f [] = 0 > test f (x:xs) = 0 > @ > -} That's a bug (one that I'm aware of). I'm planning to rewrite Haddock's lexer using Alex and get rid of this strange behaviour at the same time. > This gives me line breaks at the end of each birdtracked > line (only). > {-$CStuff > >test f [] = 0 > >test f (x:xs) = 0 > > @ > test f [] = 0 > test f (x:xs) = 0 > @ > -} Urk, that's just a bug. Thanks! Cheers, Simon

1 0

I/O Question
by Seth Kurtzberg 06 May '03

06 May '03

I need to find a way to determine whether data is available, for reading, on a socket connection, without blocking a thread. The equivalent of select() or poll() would do the trick. I don't need to get the information for all open files, as you can in select and poll, so a way to find out if there is data available on a single socket (or handle) is sufficient. TIA, -- Seth Kurtzberg MIS Corp 480-661-1849 seth(a)cql.com

2 1

Haddock code block problems
by Adrian Hey 04 May '03

04 May '03

Hello I'm using the latest Haddock release (0.4) and seem to be having a lot of trouble getting line breaks to appear in code blocks. I've tried it so many ways that if I were to enumerate them all this would be a very long email. I can't find any consistency in when it decides to work (occasionally) and when it doesn't (usually). Here's some examples.. This gives no line breaks: {-$CStuff @ test f [] = 0 test f (x:xs) = 0 @ -} Neither does this: {-$CStuff >test f [] = 0 >test f (x:xs) = 0 -} This gives a parse error:-( {-$CStuff @ test f [] = 0 test f (x:xs) = 0 @ >test f [] = 0 >test f (x:xs) = 0 -} This doesn't, but still no line breaks. {-$CStuff >test f [] = 0 >test f (x:xs) = 0 @ test f [] = 0 test f (x:xs) = 0 @ -} This gives me line breaks at the end of each birdtracked line (only). {-$CStuff >test f [] = 0 >test f (x:xs) = 0 @ test f [] = 0 test f (x:xs) = 0 @ -} But this gives me line breaks for both (as I would expect from the Haddock docs): {-$CStuff @ test f [] = 0 test f (x:xs) = 0 @ >test f [] = 0 >test f (x:xs) = 0 -} Strangely this works too (no parse error this time): {-$CStuff >test f [] = 0 >test f (x:xs) = 0 @ test f [] = 0 test f (x:xs) = 0 @ -} So is this possibly a bug in Haddock? Perhaps there's some syntactic requirement I'm failing to observe? Thanks -- Adrian Hey

1 0

Alignment of doubles in ghc
by Matthew Donadio 04 May '03

04 May '03

Hi all, I was wondering if anyone has investigated the alignment of doubles in ghc. On x86 machines, doubles can be aligned on 4 byte boundaries, but there is a performace improvement if they have 8 byte alignment. As far as I can tell, ghc uses 4 byte alignment for doubles. I started to look into the changes needed to go from 4 to 8 byte alignment... >From what I saw, it looks like I need to add -malign-double to the gcc flags in ghc/compiler/main/DriverFlags.hs for the x86 architecture. The only other place I saw is in the autoconf stuff. The FPTOOLS_CHECK_ALIGNMENT macro is defined in aclocal.m4, and I can just set CFLAGS="-malign-double" in that macro (despite the name, that flag also effects other types...). Are there any other places that need attention? Has anyone else looked into this? Thanks. -- Matthew Donadio <m.p.donadio(a)ieee.org>

1 0

RE: Slow compile times
by Simon Marlow 02 May '03

02 May '03

> On Fri, May 02, 2003 at 02:56:45AM +0200, Peter Simons wrote: > > Kirsten Chevalier writes: > > > > > I'm running GHC with the following options: > > > -fallow-overlapping-instances -cpp -O1 -fvia-C > -fglasgow-exts -package text > > > -dcore-lint -prof -auto-all +RTS -K48M -M1024M -RTS > -fno-strictness > > > -Rghc-timing > > > > Did you try setting a bigger "suggested heap size" for GHC with > > the run-time-system option "-H"? I found that setting something > > like "-H128M" affected compile times positively on my machine. > > > > I hope this helps ... > > > > I tried compiling the same file with -H512M -M512M instead of -M1024M, > keeping all other options the same. I got the following results: > > <<ghc: 3763646968 bytes, 33 GCs, 17879484/44290920 avg/max > bytes residency (6 > samples), 528M in use, 0.00 INIT (0.83 elapsed), 54.89 MUT > (116.82 elapsed), > 9.58 GC (11.56 elapsed) :ghc>> > > With -O0, I get: > > <<ghc: 1482033536 bytes, 26 GCs, 9519452/29147696 avg/max > bytes residency (5 > samples), 515M in use, 0.00 INIT (0.28 elapsed), 15.27 MUT > (69.29 elapsed), > 2.63 GC (2.92 elapsed) :ghc>> > > This is better than before, but it still seems pretty slow; any other > suggestions? The main thing to reduce GC time is to use -H<size>, as suggested. We do have some compilation speed "issues" with large source files, which I hope we'll get around to investigating at some point. You're using -prof, which precludes use of the native code generator. If you can do without -prof, then -fasm should improve compilation times a lot without affecting performance of the compiled program too much. Cheers, Simon

1 0

RE: joinPS missing from Data.PackedString
by Simon Marlow 02 May '03

02 May '03

> is this omission intentional? No, and I just added it recently (along with unlinesPS and unwordsPS defined in terms of it). Cheers, Simon

1 0