[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
trouble porting clisp
- To: clisp-list@ma2s2.mathematik.uni-karlsruhe.de
- Subject: trouble porting clisp
- From: marcus@ee.pdx.edu (Marcus Daniels)
- Date: Thu, 2 Dec 93 09:39:20 PST
I'm trying to bring up clisp-1993-11-08 on m68k NEXSTEP 3.1.
I've configured it so it compiles with few complaints. I can
compile clisp with the `strict bsd' and posix configurations. To compile
several files, it was necessary to use SAFETY=3 (with -g, the compiler
(gcc 2.2.2) generated local symbols that weren't defined, which caused the
assembler to fail). Also, for some reason, NeXT's gcc wouldn't grok
the register ... __asm__("register_name") directives in eval.d...
I've tried the various memory-management options, but none work `out
of the box'. NeXT doesn't document the mmap functions, and several
of the relevant constants aren't defined. So the configuration that
I've started working with is the SPVW_MIXED_BLOCKS -- neither the
HAVE_MMAP_ANON or HAVE_MMAP_DEVZERO are defined.
Here is what happens when lisp.run is invoked initially:
*** - The value of *TERMINAL-IO* was not a stream: #<UNBOUND>. It has been changed to #<TERMINAL-STREAM>.
*** - The value of *DEBUG-IO* was not a stream: #<UNBOUND>. It has been changed to #<SYNONYM-STREAM *TERMINAL-IO*>.
*** - No more room for LISP objects
*** - No more room for LISP objects
*** - No more room for LISP objects
*** - No more room for LISP objects
*** - No more room for LISP objects
*** - No more room for LISP objects
*** - No more room for LISP objects
*** - No more room for LISP objects
*** - No more room for LISP objects: RESET
*** - The value of *STANDARD-OUTPUT* was not a stream: #<UNBOUND>. It has been changed to #<SYNONYM-STREAM *TERMINAL-IO*>.
*** - No more room for LISP objects: RESET
*** - The value of *STANDARD-INPUT* was not a stream: #<UNBOUND>. It has been changed to #<SYNONYM-STREAM *TERMINAL-IO*>.
*** - No more room for LISP objects: RESET
The garbage collection that triggers the `no room' messages is
triggered by an attempt to display the `breakprompt_string', which
apparently isn't initialized yet (the length field in the sstring structure
is 537564620, and the characters arent `Break>'. ).
I'm guessing there is a obvious oversight I'm making that explains this
behavior. (BTW, the NeXT installation note isn't valid for 3.0).
Any suggestions?