Hello!
On Tue, Jul 28, 1998 at 07:51:25PM +0800, Shin-Cheng Mu wrote:
> Hello, my Hero Haskell Hackers, :)
> Compiling GHC 3.02 for i386 Linux, the make process stops
> at the point when compiling ghc/lib/posix/PosixProcPrim.lhs.
> The error message generated is attached below.
> I tried the patch suggested by Carl R. Witty (I'm not sure
> whether it was to my problem, through), it doesn't work either.
> Do you have any idea what went wrong?
Yes, I stumbled over the same thing a few days before, but was too
lazy to report it.
That module must be compiled using -monly-2-regs with gcc-2.7.*.
With 2.8.1 (on OpenBSD, by the way, the bootstrap *was* some work,
but succeeded, with help from OpenBSD's Linux binary emulation and
ghc2-Linux-binaries) that option is *not* needed, though!
Regards, Felix.
PS: Just stumbled over another buglet:
in ghc/lib/std/PrelIOBase.lhs, around line 176:
isEOFError (IOError _ EOF _) = True
isEOFError _ = True
In the second line, there should be False instead of True...
> --- error message generated by ghc ------
>
> ../../../ghc/driver/ghc -i../misc -recomp -cpp -fglasgow-exts -fvia-C
> -Rghc-timing -O -split-objs -odir PosixProcPrim '-#include"cbits/libposix.h"' -c
> PosixProcPrim.lhs -o PosixProcPrim.o -osuf o
>
> NOTE: Simplifier still going after 4 iterations; bailing out.
> <<ghc: 384309900 bytes, 312 GCs, 2426211/2767904 avg/max bytes residency (20
> samples), 0.02 INIT (0.01 elapsed), 17.97 MUT (18.26 elapsed), 10.59 GC (10.66
> elapsed) :ghc>>
> ghc: module version changed to 1; reason: no old .hi file
> /tmp/ghc2094.hc:2311: fixed or forbidden register was spilled.
> This may be due to a compiler bug or to impossible asm
> statements or clauses.
> gmake[3]: *** [PosixProcPrim.o] Error 1
> gmake[2]: *** [all] Error 1
> gmake[1]: *** [all] Error 1
> gmake: *** [all] Error 1
