i should also add that the same command is executed whether 'make' is
executed in libraries or libraries/base, so something is either overwriting
Prelude.hi in libraries/Makefile, or maybe an environment variable from
libraries/Makefile is affecting the compiler.

On 3/18/08, Dustin DeWeese <[EMAIL PROTECTED]> wrote:
>
> i started with a clean tree.
>
> This is the command line used to build Prelude.hi (from libraries/base):
>
> ./../compiler/stage1/ghc-inplace -package-name base-3.0.1.0-hide-all-packages 
> -i -idist/build/autogen -idist/build -i. -Idist/build
> -Iinclude -#include "HsBase.h" -odir dist/build -hidir dist/build -stubdir
> dist/build -package rts-1.0 -O -fglasgow-exts -package-name base -XCPP
> -idist/build  -H16m -O -optc-mcpu=arm1136j-s -optc-mfpu=vfp
> -optc-mfloat-abi=softfp -opta-mcpu=arm1136j-s -opta-mfpu=vfp
> -opta-mfloat-abi=softfp -O -Rghc-timing -fgenerics -c Prelude.hs -o
> dist/build/Prelude.o  -ohi dist/build/Prelude.hi
>
> Interestingly, this produces an unregisterised Prelude.hi when repeated.
> If I run 'make' in libraries/base, Prelude.hi is registerised.  If I run
> 'make' in libraries, it is unregisterised.  I set
> GhcRTSWays=
> and
> GhcLibWays=
> beforehand, so the unregisterised versions shouldn't be built.
>
> I do have cGhcUnregisterised = "NO" in compiler/main/Config.hs
>
> I think the issue is in libraries/Makefile somewhere.
>
> Thanks,
> Dusty
>
> On 3/17/08, Ian Lynagh <[EMAIL PROTECTED]> wrote:
> >
> > On Mon, Mar 17, 2008 at 11:28:54AM -0400, Dustin DeWeese wrote:
> > >
> > > So ghc seems to be writing the wrong information to the .hi
> > file...  Any
> > > idea why, or what I should look at that could be causing this?
> >
> >
> > Did you start off with a clean tree?
> >
> > What does the command line used to build Prelude.hi look like?
> >
> > If you repeat the command (with -fforce-recomp added), is it still
> > unregisterised?
> >
> > In compiler/main/Config.hs, you presumably have
> >     cGhcUnregisterised    = "NO"
> > ?
> >
> >
> > Thanks
> >
> > Ian
> >
> >
>
_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc

Reply via email to