RE: target all fails

2000-01-05 Thread Simon Marlow
Compiling some libraries whilst making target "all" fails because sometimes "ghc" is used instead of "../../ghc/driver/ghc". If there's already some old, buggy ghc in the PATH, this one is used and so compilation fails. Can you be more specific? i.e. where exactly in the tree are you

RE: -ansi switch doesn't work

2000-01-05 Thread Simon Marlow
The -ansi switch doesn't take effect on preprocessing -- the preprocessor is still being called with -traditional: Yep, the -ansi switch basically doesn't work. The RTS isn't ANSI C - it uses some GCC extensions. And we have to call the preprocessor with -traditional to preprocess Haskell

RE: Socket module does not export Socket (datatype) !

2000-01-05 Thread Simon Marlow
I'm writing a program which does import Socket but not import SocketPrim GHC complains that the type constructor Socket isn't in scope. Indeed, in Socket.lhs from GHC 4.04.19990916 I see that Socket isn't exported, even though a function listenOn :: PortID - IO Socket is. When I say

RE: source binary packages bugs; shared libraries

2000-01-05 Thread Simon Marlow
While I'm on the subject of shared libraries,why is it that the Win32 distribution of GHC includes the RTS, Prelude, and perhaps other libraries, as DLL's, but (as far as I can tell) none of the distributions for Unix-like platforms include any of these as shared libraries? Is there some

RE: RegexString: was it really meant to be that way ?

2000-01-05 Thread Simon Marlow
You must think I'm making a terrible nuisance of myself. Please tell me to go away and real manuals if I'm being annoying or clueless. Anyway, I have another problem, with RegexString this time. I was trying to use mkRegex and matchRegex and I get this error: Fail: Ix{Int}.index:

Mysterious missing symbols in object files

2000-01-05 Thread George Russell
Whenever I try to compile anything non-trivial, for example green-card, I get lots of complaints from the linker about missing symbols. It looks as if this is due to some flakey handling in the latest version of supposedly inlined functions. Example - the symbol

RE: Lack of error message with ``sth``

2000-01-05 Thread Simon Marlow
When compiling a module A, contained in a file called F, GHC produces A.hi F.o Notice that there is no F.hi. (Or, in your case, 1.hi.) You should have got an error message saying that Main.main isn't in scope (main is reqd in module Main). This is what happens when

RE: Deadlock again (was Re: GHC 4.04.19990916 produces coredumping executable)

2000-01-05 Thread Ian Jackson
Simon Marlow writes ("RE: Deadlock again (was Re: GHC 4.04.19990916 produces coredumpin g executable)"): No doubt you're quite tired of seeing my mails by now, but here, have another one :-). I now have a couple of problems which look like they can be summarised as `race