Failed to bootstrap 6.10.4 with itself on Windows (MinGW)

[Simon Marlow]

On 24/07/2009 05:17, leledumbo wrote:
>
> OK, I've tried ghc's supplied gcc, too (not so easy, I need to set some
> environment variables first) and here are the results:
>
> With ghc's gcc:
>
> D:/Sources/ghc/ghc-6.10.4/ghc/stage1-inplace/ghc.exe -package rts-1.0
> -optc-O2 -
> odir dist/build -c cbits/longlong.c -o dist/build/cbits/longlong.o
> (echo dist/build/cbits/longlong.o `find dist/build -name "*_stub.o" -print`;
> fin
> d dist/build/GHC/Bool_split dist/build/GHC/Generics_split
> dist/build/GHC/Orderin
> g_split dist/build/GHC/PrimopWrappers_split dist/build/GHC/IntWord32_split
> dist/
> build/GHC/IntWord64_split dist/build/GHC/Tuple_split
> dist/build/GHC/Types_split
> dist/build/GHC/Unit_split -name '*.o' -print) | xargs c:/ghc/bin/ar.exe q
> dist/
> build/libHSghc-prim-0.1.0.0.a
> xargs: c:/ghc/bin/ar.exe: Bad file number

See this ticket:

http://hackage.haskell.org/trac/ghc/ticket/3201

work-aroundable by adding SplitObjs=NO to mk/build.mk

>> FYI, if you want your GHC build to work on Windows, it is best to use
>> the same tools that we do
>
> It's the same as using ghc's supplied gcc, right (other requirements from
> the page already fulfilled)?

I think so, yes.  Although, as you noticed, using GHC's built-in gcc is 
not terribly convenient.  It will be much easier in 6.12.1, as we've 
made the layout of the bundled gcc files match the standard MinGW layout.

Cheers,
	Simon