the joys (?!!) of porting CLISP

[michaelw@karl.cs.su.OZ.AU (Michael Wise)]

So far I have tried porting it to R4000 RISCos 5.01, got almost no way
down the track before it was total confused and hope less (due to the
multiple universes being supported - best with the bsd flavour);
R3000 with RISCos 4.52 - slightly better, but still got bogged.
I then decided to move to the new Sparc Sun (running Solaris 2.XX)
because we're about the move from the MIPS's in any case to a Sun
multiprocessor. This attempt has been more successful.
Main discovery was to define:
LD_LIBRARY_PATH=/usr/ccs/lib:/usr/lib:/usr/openwin/lib
and for good measure, change
 -Y P,/usr/ccs/lib:/usr/lib
to
 -Y P,/usr/ccs/lib:/usr/lib:/usr/openwin/lib
in the definition for lisp.run .

I've now gotten stuck when trying to make interpreted.mem .
After getting started ok and loading defseq.lsp, backquot.lsp,
defmacro.lsp, macros1.lsp, macros2.lsp ... the following happens:

;; Loading file defs1.lsp ...*** Error code 139
make: Fatal error: Command failed for target `interpreted.mem'


Any bright ideas? BTW, any way that the header can be suppressed at run
time. The problem is that when the system will be run in batch mode
(when student assignments are being marked), the extra output generated
will have to be filtered or it will confuse the marking program. It's there-
fore far easier to suppress it in the first place.

Cheers
MichaelW