Thanks,
[Cc'ed to glasgow-haskell-bugs - the people on that list likes to
see this kind of stuff]
the options to hsc loop sane - what version/port of perl are you
using? I've seen this kind of problem with the Hip/ActiveWare port of
perl (system() invokes cmd.exe rather than bash as command shell,
causing the \( \) cmd line arguments to hsc to be mangled by
cygwin.dll
A cygwin32 compiled version of perl5 can be found at
http://www.edv.agrar.tu-muenchen.de/~syring/gnu-win32/
which works a lot better.
Also, make sure your PATH is set up such that you pick GNU rm rather
than the MS one. (same goes for `time')
Let me know if these suggestions this fixes anything.
--Sigbjorn
Peter Caven writes:
> Re: "compiler bug",
>
> Here is the result of running "ghc -v bug.hs" :
>
> ============================================================================
> ==========
> bash$ ghc -v bug.hs
> The Glorious Glasgow Haskell Compilation System, version 2.02, patchlevel 0
>
> Ineffective C pre-processor:
> echo '#line 1 "bug.hs"' > /tmp/ghc172.cpp && cat bug.hs >>
> /tmp/ghc172.cpp
>
> ================================================================
> *** It hangs here until I press enter, and then this follows:
> ================================================================
>
> Haskell compiler:
> /fptools/lib/i386-unknown-cygwin32/ghc-2.02/hsc ,-W
> ,/tmp/ghc172.cpp -h
> isuf-prelude=.hi -hisuf=.hi -fignore-interface-pragmas
> -fomit-interface-pragmas
> -fsimplify \( -ffloat-lets-exposing-whnf -ffloat-primops-ok -fcase-of-case
> -fre
> use-con -fpedantic-bottoms -fsimpl-uf-use-threshold0
> -fessential-unfoldings-only
> -fmax-simplifier-iterations4 \)
> -himap=.:/fptools/lib/i386-unknown-cygwin32/g
> hc-2.02/imports -v -hifile=/tmp/ghc172.hi -C=/tmp/ghc172.hc +RTS
> -H6000000 -K1
> 000000
>
> *** Pattern-matching error within GHC!
>
> This is a compiler bug; please report it to
> [EMAIL PROTECTED]
>
> Fail: "main/CmdLineOpts.lhs", line 451: incomplete pattern(s) to match in
> functi
> on "simpl_sep"
> Command had non-zero exit status 1
> 0.00user 0.00system 0:00.20elapsed 0%CPU (0avgtext+0avgdata 0maxresident)k
> 0inputs+0outputs (0major+0minor)pagefaults 0swaps
> deleting... /tmp/ghc172.hi /tmp/ghc172.hc
>
> rm -f /tmp/ghc172*
> Microsoft File Removal Utility. Version rmj.0rmm.00rup
> Copyright (C) Microsoft Corp 1987-90. All rights reserved.
>
> Usage: RM [/help] [/ikft] [/x [/d]] [/r dir] files
> /help invoke Quick Help for this utility
> /i inquire of user for each file for permission to remove
> /k keep read only files, no prompting to remove them
> /r dir recurse into subdirectories
> /f force delete of read only files without prompting
> /t attempt to take ownership of file if delete fails
> /x dont save deleted files in deleted subdirectory
> /d delay until next reboot.
> bash$
>
> ============================================================================
> =============
> *** The contents of file "ghc172.cpp" (which was not deleted) are:
> ============================================================================
> =============
> The system cannot accept the time entered.
> Enter the new time: -- GHC test on i386-cygwin32 (Windows NT 4.0, Service
> Pack 2)
> fact :: Integer -> Integer -> Integer
> fact n a = if (n == 1) then a else fact (n - 1) (a * n)
> ============================================================================
> =============
>
>
> The use of the "rm" command on Windows NT appears to a problem too.
>
>
> I hope this helps! (Or is my system misconfigured?).
>
> -- Peter.
>
