GHCi 5.00 on HPUX question

[BENNETT,ANDY (HP-Unitedkingdom,ex1)]

Hi,

I've managed to build the GHC 5.00 compiler on HP-UX, but when I try to
build the interactive bit, subsequent to this, I'm getting some complaints
about a function "mkJumpToAddr" in ByteCodeItBls.lhs. My question is, is
this only relevant to platforms with an NCG available and hence could I put
in some sort of empty stub, or do I need to provide a proper routine to
build the machine codes for a jump?

Thanks,
Andy.

> -----Original Message-----
> From: Simon Marlow [mailto:simonmar at microsoft.com]
> Sent: 10 May 2001 10:19
> To: William Chesters; glasgow-haskell-bugs at haskell.org
> Subject: RE: No luck installing on SuSE 7.1
> 
> 
> Hi William,
> 
> It's not your day, is it? :)  You seem to have tripped over just about
> all the nasty bugs in 5.00.
> 
> > Hi, I'm trying to get GHC 5.00 installed on my SuSE 7.1 box.  I have
> > tried (a) the binary distribution, (b) the source 
> distribution, (c) a
> > CVS snapshot.
> > 
> > 
> >    The binary distribution 
> ghc-5.00-i386-unknown-linux.tar.bz2 does to
> > some extent work if I either install libreadline.so.3 
> manually or just
> > symlink it to libreadline.so.4.  I can compile and run trivial
> > programs and use ghci a bit.
> 
> Yep, we'll be removing the dependency on readline in the 
> standard binary
> distribution for 5.00.1 - too many people were burned by this.
> 
> >    However the toy exe I build terminates with an "Illegal
> > instruction" signal (whether or not I use -via-C), and the 
> same quite
> > quickly happens to ghci if I play around.  If I copy the 
> toy exe onto
> > a Mandrake (~ RedHat) 6.0 machine I get the same.  (If I 
> copy it onto
> > an RH5 machine I get segfault instead.)
> 
> This was caused by a misconfiguration problem with the GMP library in
> the binary distribution (it configured itself for i686, with 
> the result
> that it doesn't work on i586 and lower), 5.00.1 will have the fix.
> 
> >    The source distribution ghc-5.00-src.tar.bz2, with
> > ghc-5.00-i386-hc.tar.bz2 overlaid on it, gets some way through
> > compilation (using distrib/hc-build) and then fails:
> 
> There were several problems with bootstrapping from .hc files in 5.00,
> and this is something we're concentrating on for 5.00.1.
> 
> >    The CVS snapshot I obtained tonight, with 
> ghc-5.00-i386-hc.tar.bz2
> > overlaid on it, gets some way through compilation (using 
> autoconf and
> > distrib/hc-build), and then fails as follows:
> 
> again, another problem with .hc bootstrapping.
> 
> > P.S.  I can see now why the ocaml team keep the bytecode interpreter
> > alive---it does make bootstrapping an awful lot easier ...
> > 
> > P.P.S.  By the way I __love__ `echo  | sed 
> > 's/^$/-DSTOLEN_X86_REGS=4/'`  ;-)
> 
> [polishes fingernails]
> 
> Cheers, and sorry for all the hassle - 5.00.1 will be greatly 
> improved,
> trust me :)
> 
> Simon
> 
> _______________________________________________
> Glasgow-haskell-bugs mailing list
> Glasgow-haskell-bugs at haskell.org
> http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
>