validate fails for PPC Mac OS X 10.4

[thorkilnaur]

Lately ( http://www.haskell.org/pipermail/cvs-ghc/2008-February/041135.html) validate has been failing for PPC Mac OS X 10.4 as follows:

../compiler/stage1/ghc-inplace -no-user-package-conf -Werror -H64m -Onot -fasm  -istage2/utils  -istage2/basicTypes  -istage2/types  -istage2/hsSyn  -istage2/prelude  -istage2/rename  -istage2/typecheck  -istage2/deSugar  -istage2/coreSyn  -istage2/vectorise  -istage2/specialise  -istage2/simplCore  -istage2/stranal  -istage2/stgSyn  -istage2/simplStg  -istage2/codeGen  -istage2/main  -istage2/profiling  -istage2/parser  -istage2/cprAnalysis  -istage2/ndpFlatten  -istage2/iface  -istage2/cmm  -istage2/nativeGen  -istage2/ghci -Wall -fno-warn-name-shadowing -fno-warn-orphans -Istage2 -package hpc -package bytestring -DGHCI -package template-haskell -DGHCI_TABLES_NEXT_TO_CODE -package readline -DUSE_READLINE -cpp -fglasgow-exts -fno-generics -Rghc-timing -I. -Iparser -package unix -package Cabal -ignore-package lang -recomp -Rghc-timing -Onot -fasm -H16M '-#include "cutils.h"' -package-name  ghc-6.9.20080208   -fgenerics    -c ghci/ByteCodeFFI.lhs -o stage2/ghci/ByteCodeFFI.o  -ohi stage2/ghci/ByteCodeFFI.hi

ghci/ByteCodeFFI.lhs:690:18:
    Couldn't match expected type `PrimRep'
           against inferred type `CgRep'
    In the pattern: FloatArg
    In a case alternative:
        FloatArg
          | nextFPR < 14
          -> (3223257088 .|. (fromIntegral haskellArgOffset .&. 65535)
          .|.
            (fromIntegral nextFPR `shiftL` 21))
           : pass_parameters args (nextFPR + 1) offsetW'
    In the expression:
        let
          haskellArgOffset = a_offW * bytes_per_word
          offsetW' = offsetW + primRepSizeW a_rep
          pass_word w | offsetW + w < 8 = [...]
                      | otherwise = [...]
                      where
                          src = ...
                          ....
        in
          case a_rep of
            FloatArg
              | nextFPR < 14
              -> (3223257088 .|. (fromIntegral haskellArgOffset .&. 65535)
              .|.
                (fromIntegral nextFPR `shiftL` 21))
               : pass_parameters args (nextFPR + 1) offsetW'
            DoubleArg
              | nextFPR < 14
              -> (3357474816 .|. (fromIntegral haskellArgOffset .&. 65535)
              .|.
                (fromIntegral nextFPR `shiftL` 21))
               : pass_parameters args (nextFPR + 1) offsetW'
            _ -> concatMap pass_word ([0 .. primRepSizeW a_rep - 1])
               ++
                 pass_parameters args nextFPR offsetW'

ghci/ByteCodeFFI.lhs:705:12:
    Couldn't match expected type `PrimRep'
           against inferred type `CgRep'
    In the pattern: VoidArg
    In a case alternative: VoidArg -> []
    In the expression:
        case r_rep of
          VoidArg -> []
          FloatArg -> [3493789696 .|. (fromIntegral result_off .&. 65535)]
          DoubleArg -> [3628007424 .|. (fromIntegral result_off .&. 65535)]
          _ | primRepSizeW r_rep == 2
            -> [2424242176 .|. (fromIntegral result_off .&. 65535),
                2426339328 .|. (fromIntegral (result_off + 4) .&. 65535)]
          _ | primRepSizeW r_rep == 1
            -> [2424242176 .|. (fromIntegral result_off .&. 65535)]
<<ghc: 47687720 bytes, 9 GCs, 2690342/5255332 avg/max bytes residency (2 samples), 17M in use, 0.01 INIT (0.00 elapsed), 0.87 MUT (2.73 elapsed), 0.61 GC (1.47 elapsed) :ghc>>
make[2]: *** [stage2/ghci/ByteCodeFFI.o] Error 1
make[2]: *** Waiting for unfinished jobs....
<<ghc: 208607448 bytes, 28 GCs, 3604469/5941884 avg/max bytes residency (3 samples), 20M in use, 0.01 INIT (0.00 elapsed), 4.79 MUT (11.39 elapsed), 1.36 GC (2.55 elapsed) :ghc>>
make[1]: *** [stage2] Error 2
make: *** [bootstrap2] Error 2

Best regards Thorkil

[simonmar]

I just pushed this patch, could someone with a PPC please test and modify if necessary?

Mon Feb 18 03:57:48 PST 2008  Simon Marlow <simonmar@microsoft.com>
  * attempt to fix #2098 (PPC pepple please test & fix)

[thorkilnaur]

On a PPC Mac OS X 10.4, validate now completes with this summary:

OVERALL SUMMARY for test run started at Wed Feb 20 22:52:28 CET 2008
    2091 total tests, which gave rise to
    7883 test cases, of which
       0 caused framework failures
    6100 were skipped

    1702 expected passes
      79 expected failures
       0 unexpected passes
       2 unexpected failures

Unexpected failures:
   divbyzero(normal)
   ghci024(ghci)

So the patch has removed the earlier failure.

Thanks and best regards Thorkil