cvs commit: fptools/hslibs/OpenGL/examples/redbook_HS AAPoly. hs Cube.hsPolys.hs Unproject.hs Double.hs Makefile

[Simon Marlow]

> > panne       2000/12/09 08:04:04 PST
> > [...]
> >   Revision  Changes    Path
> >   1.1.2.2   +1 -1      
> fptools/hslibs/OpenGL/examples/redbook_HS/Attic/Double.hs
> >   1.1.2.4   +2 -2      
> fptools/hslibs/OpenGL/examples/redbook_HS/Attic/Makefile
> 
> Huh?? Has anybody an idea why there is "/Attic" in the above path?

My fault, I forgot to merge this directory from the branch.  Fixed now.

BTW, we'd like to encourage people to start using the HEAD.  As a
compiler, it is pretty reliable now; I've done a complete nofib run at
least.  I'll be doing a build or two today to iron out various niggles.

Things you're likely to run into:

	- to specify which compiler to use for bootstrapping,
	  you need to use --with-ghc=<blah> to configure.  4.08.1
	  works fine.

	- the compiler doesn't recognise -H, -K, and -Rghc-timing.
	  We're discussing what to do about these, but for now
	  you can replace -H with +RTS -H<whatever> -RTS.

	- -fglsagow-exts no longer implies -package lang.  This is
	  a painful change, but it's one that was long overdue and
	  had to be made for techinical reasons in the driver
	  (-fglasgow-exts is per-module, but -package lang applies to
	  the whole compilation).

as for GHCi, the current meta-circular interpreter works but is going to
be removed in favour of a bytecode based interpreter (basically a
re-working of the old STG Hugs bytecode evaluator).  To get GHCi you
need to bootstrap (and that's all; to turn it off you need to specify
GhcWithInterpreter=NO in build.mk).

Cheers,
	Simon