Kevin Glynn writes:
> 
> Simon Marlow writes:
>  > 
>  > This is pre-release of the long awaited GHC 4.00.  I decided to put
>  > together a release today because the sources are in pretty good shape
>  > at the moment, not because it is at all ready.  In particular, we plan
>  > to have some sort of profiling working for the real 4.00 release.
>  > 
>  > The source snapshot can be found at
>  > 
>  >     ftp://ftp.dcs.gla.ac.uk/pub/haskell/glasgow/working/ghc-pre-4.00.tar.gz
>  > 
>  > Have Fun!
>  > 
> 
> Actually:
> 
> ftp://ftp.dcs.gla.ac.uk/pub/haskell/glasgow/working/ghc-pre-4.00-src.tar.gz
> 

There was a couple of minor problems with that tarball, so I've rolled
another one that hopefully simplifies the path to ghc-4.00 happiness.
Try

ftp://ftp.dcs.gla.ac.uk/pub/haskell/glasgow/working/ghc-pre-4.00.1-src.tar.gz

--Sigbjorn

For those keeping track / don't want to download it all again, the
changes in this bundle were: 

 - removed ghc/driver/ghc
 - removed ghc/compiler/parser/tree.{c,h}
 - removed ghc/compiler/parser/hslexer.c
 - removed ghc/compiler/parser/hsparser.tab.{c,h}
 - added mk/config.h.in

 - turn off the building of profiling libs by default, since
   the profiling bits in the rts doesn't compile cleanly:

*** fptools-4.00/mk/config.mk.in        Fri Aug 21 13:35:43 1998
--- fptools/mk/config.mk.in     Sat Aug 29 16:42:13 1998
***************
*** 208,212 ****
  GhcLibWays=
  else
! GhcLibWays=p
  endif
  
--- 208,212 ----
  GhcLibWays=
  else
! GhcLibWays=
  endif
  

 - configure.in: changed default WithGhcHc from ghc-2.10 to
   ghc-3.02, since the compiler sources doesn't compile
   cleanly with 2.10 at the moment:

*** fptools-4.00/configure.in   Fri Aug 21 13:35:26 1998
--- fptools/configure.in        Sat Aug 29 16:48:13 1998
***************
*** 287,296 ****
  AC_ARG_WITH(ghc-hc,
  [  --with-ghc-hc=<haskell compiler>
!         Use a command different from 'ghc-2.10' to compile up the GHC compiler 
sources.
           (no claims currently made that this will work with a compiler other than a
            recent version of GHC, but you could always try...)
  ],
  [WithGhcHc="$withval"],
! [WithGhcHc="ghc-2.10"]
  )
  AC_SUBST(WithGhcHc)
--- 287,296 ----
  AC_ARG_WITH(ghc-hc,
  [  --with-ghc-hc=<haskell compiler>
!         Use a command different from 'ghc-3.02' to compile up the GHC compiler 
sources.
           (no claims currently made that this will work with a compiler other than a
            recent version of GHC, but you could always try...)
  ],
  [WithGhcHc="$withval"],
! [WithGhcHc="ghc-3.02"]
  )
  AC_SUBST(WithGhcHc)


 - misc ghc/rts/gmp Makefile fixes to make gmp source build properly.

*** fptools-4.00/ghc/rts/gmp/Makefile.in        Fri Aug 21 15:48:47 1998
--- fptools/ghc/rts/gmp/Makefile.in     Sat Aug 29 17:49:45 1998
***************
*** 193,196 ****
  
  force:
! .PNONY: check install install-bsdmp install-info-files install-strip uninstall
  .PHONY: doc clean distclean maintainer-clean force info dvi
--- 193,204 ----
  
  force:
! .PHONY: check install install-bsdmp install-info-files install-strip uninstall
  .PHONY: doc clean distclean maintainer-clean force info dvi
+ 
+ # added sof. 
+ .SUFFIXES: .o .c
+ # In the fptools/ mk setup we clear the setting of .SUFFIXES since we
+ # don't use any pre-supplied suffix rules (if any). This causes a
+ # problem when descending into a directory like gmp/ which does
+ # use suffix rules, so we need to add .c & .o to the .SUFFIXES list
+ # here.


*** fptools-4.00/ghc/rts/gmp/mpz/Makefile.in    Fri Aug 21 15:50:28 1998
--- fptools/ghc/rts/gmp/mpz/Makefile.in Sat Aug 29 15:40:01 1998
***************
*** 78,82 ****
  clean mostlyclean:
        rm -f *.o libmpz.a
!       -cd tests; $(MAKE) $@
  distclean maintainer-clean: clean
        rm -f Makefile config.status
--- 78,82 ----
  clean mostlyclean:
        rm -f *.o libmpz.a
!       #-cd tests; $(MAKE) $@
  distclean maintainer-clean: clean
        rm -f Makefile config.status

Reply via email to