August 2008 - Glasgow-haskell-users

ghci history.
by Venkat Ramanan 21 Aug '08

21 Aug '08

Hi all, I've just started learning haskell. I encountered a strange behaviour in ghci. 1. ghci doesn't seem to do pruning of duplicate expressions. If the same expression is entered twice, both of them are stored in the history instead of just one. Bash uses environmental variables (HISTCONTROL and HISTIGNORE ) to modify this behavior. Is there a way to get the same behaviour in ghci? 2. Does the history of ghci get stored somewhere in a file? That would be useful. I'm using ghc-6.8.2 (ubuntu 8.04). Also, is there an updated ghc package (6.8.3) for ubuntu? If not, I'll try to install the binary. Thanks, Venkat.

2 2

#ghc connection issue
by Claus Reinke 20 Aug '08

20 Aug '08

I seem to be unable to join the ghc chatroom at irc.freenode.net at the moment (using Opera). Is that an issue with my irc client or a general problem? 15:47 Joining chat room... Disconnected from chat Claus

4 4

invoking a Haskell script without a .hs extension
by prj＠po.cwru.edu 18 Aug '08

18 Aug '08

I have a Haskell script called "notify", without a .hs extension, which causes some problems. (I'm using ghc 6.8.3.) First attempt: runhaskell notify Without the .hs extension, ghc doesn't know it's a Haskell script, and so I get "Could not find module `notify'". Maybe runhaskell should automatically add "-x hs" to the ghc command? Second attempt: runhaskell -x hs notify This get me "Not in scope: `main'". runhaskell is invoking ghc with these arguments: -ignore-dot-ghci -x -e ':set prog "hs"' -e ':main ["notify"]' hs So it looks like runhaskell it treating "-x" as an argument to be relayed to ghc, "hs" as the name of the script, and "notify" as an argument to the script. I guess I need to use "--" to make it clear where the ghc arguments end and where the script and its arguments begin. Third attempt: runhaskell -- -x hs -- notify This gets me "Not in scope: `main'" again. runhaskell is invoking ghc with these arguments: -ignore-dot-ghci -x -e ':set prog "hs"' -e ':main ["--","notify"]' hs This looks like a bug in the "--" handling, unless I'm misinterpreting the usage message I get from running plain "runhaskell". Fourth attempt: ghc -ignore-dot-ghci -e ':set prog "notify"' \ -e ':main []' -x hs notify This works, but passing arguments becomes rather cumbersome. If there's a way to get runhaskell to pass "-x hs" in the right place, that would be much better. A somewhat related issue: I'd like to avoid hard-coding the path to runhaskell or ghc in the #! line. Instead, I'd like to use #!/bin/sh, and have the shell find runhaskell or ghc in $PATH. This means the first few lines of the script would have to be executed by sh, but ignored by ghc. Most scripting langauges have some way of doing this, although it's often by accident. The best way I've seen is in Guile, where "#!" starts a multi-line comment, and "!#" ends it. For Haskell, this is the best I could come up with: #!/bin/sh {- > /dev/null 2>&1 exec ghc -ignore-dot-ghci \ -e ":set prog \"$0\"" \ -e ':main []' -x hs "$0" -} But this depends on "{-" not existing as an executable command, or at least not doing anything harmful when invoked like this. I'd like to avoid depending on anything like that. Does anyone have any better ideas? paul

4 5

Any release date for GHC 6.8.3 for Mac OS X 10.5 (Leopard) on PowerPC G4 machines?
by Benjamin L.Russell 18 Aug '08

18 Aug '08

Is there any scheduled release date for GHC 6.8.3 for Mac OS X 10.5 (Leopard) on PowerPC G4 machines? I have a G4 PowerBook, currently running Mac OS X 10.2.8 (Jaguar), which I plan to upgrade next week. I also have the upgrade discs for both Mac OS X 10.4 (Tiger) and Mac OS X 10.5 (Leopard). Backing up my hard disk before upgrading takes about 20 hours; it seems a pity to upgrade to Tiger before eventually upgrading to Leopard (unless GHC is never going to be released for Tiger on PowerPC G4 machines, which would be much worse). -- Benjamin L. Russell

2 5

problems running ghc-6.8.2 on solaris 10, sparc
by Daniil Elovkov 16 Aug '08

16 Aug '08

Hello folks, Christian I'm trying to get ghc 6.8.2 running on Solaris 10 and having problems. To be precise, I'm trying to compile a 'hello world' program by ghc 6.8.2 which I got in binary form haskell.org. gcc is 2.95, it uses sun linker. I remember there were problems with that in the past. Is ghc supposed to work only with gnu ld or sun ld as well? So, how it went first I got compiler errors in many places of Reg.hs: global register variable follows a function definition Googling showed that Don Stewart used to fix it by swapping 2 includes in Stg.h - putting MachRegs.h after Regs.h instead of before. It helped Then there was assembler error: cannot use v8plus instructions in a non-v8plus target binary It was caused by -mcpu=v9, which ghc passes to gcc. I blindly added -optc -mcpu=v8 and it helped :) Then the linker complained that it could not resolve aio_fork and __aio_suspend64, referenced from librt.so. -lrt is passed by ghc to the linker. On this machine there is /lib/libaio.so. Linking with it didn't help. It doesn't really contain exactly those functions, only with slightly different names, like _aio_forkinit, _libaio_fork. Also, librt which was linked in was the one lying close to where gcc is installed. Apart from that, there is also librt.so in /lib. I thought, maybe the wrong librt was used and said -optl -L/lib to link against the one in /lib. No complaints, but the resulting binary segfaults. Does anybody have any ideas? -- Daniil Elovkov

4 5

Overlapping instances vs. Haskell98 Report [was: How to disable warning for "export item 'module ...' exports nothing"?]
by Sean Leather 15 Aug '08

15 Aug '08

Ross Paterson wrote: > On Fri, Aug 15, 2008 at 04:17:44PM +0200, Sean Leather wrote: > > module A where > > class A t where > > a :: t > > > > module B where > > import A > > instance A Int where > > a = 0 > > a0 :: Int > > a0 = a > > > > module C where > > import A > > instance A Int where > > a = 1 > > a1 :: Int > > a1 = a > > > > module Main where > > import A > > import B > > import C > > main = do putStrLn $ "a0=" ++ show a0 > > putStrLn $ "a1=" ++ show a1 > > > > This works, because of the way the instances are used. While overlapping > > instances are imported into Main, they are not used in Main. > > Then that is a GHC bug. Haskell 98 Report 4.3.2: "A type may not be > declared as an instance of a particular class more than once in the > program." > That's interesting. So, maybe there should be some language extension or warning (with associated -fno-warn) for this in GHC. Personally, I prefer the way it's done now. (I guess that's obvious, considering I'm developing a library that will take advantage of it. ;) ) But it makes sense that instances are only looked up when needed, instead of globally tracked. Sean

2 1

How to disable warning for "export item 'module ...' exports nothing"?
by Sean Leather 15 Aug '08

15 Aug '08

Hi, I have a module A that re-exports module B, and module B contains only class instances. Since I'm using -Wall, I get this warning about module B exporting nothing. Is there a way to disable this particular warning, since it is unnecessary in this case? No mailing list messages or GHC documentation has told me what to do. Thanks, Sean

3 12

How to produce a statically linked binary with GHC?
by Nicola Squartini 15 Aug '08

15 Aug '08

It seems that the -static flag serves a different purpose: according to the GHC manual it forces the compiler use the static Haskell libraries, but the resulting binary is still dynamically linked to the C libraries. What is the correct flag to create a static binary? Thanks Best Regards Nicola

5 8

Re: Orphan Instances
by Yitzchak Gale 12 Aug '08

12 Aug '08

<Moving this side point to the ghc users list...> Ashley Yakely wrote: >> What is an orphan instance, and why do we care about them? Simon Peyton-Jones wrote: > They are documented in the GHC manual > http://www.haskell.org/ghc/docs/latest/html/users_guide/separate-compilatio… Thanks for the nice explanation there. Since you brought it up, though, the explanation might be made even clearer with two small changes: 1. The term "instance head" is used throughout the explanation, beginning already with the second paragraph ("the head of the instance..."). However, that term is only defined near the very end '(the part after the "=>")'. For those who do not already know the surprise ending, this makes the explanation read rather like a mystery story. 2. I believe that the text "by setting a = Int" was meant to be "by setting a = T". Thanks, Yitz

2 1

RE: Version control systems
by Eric 12 Aug '08

12 Aug '08

A metacomment: As a lurker, and a reader of other languages' mail groups, I just wanted to complement you GHC folks on the quality of your discussion. You're discussing an issue that people clearly have strong opinions about, yet you've all remained polite and respectful and kept the signal-to-noise ratio of the discussion high, unlike what I see on some the mail groups of some other languages. It's a pleasure reading even your disagreements. --Eric

1 0