[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

clisp-1996-05-30 Installation

Help needed:

I have been trying to install clisp in my computer (Sun 386 running 
SunOS-4.0.2, with 16 MB memory), but after numerous compile runs,
when I arrive to 'make interpreted.mem', the line:

./lisp.run -m 800kw -x "(load \"init.lsp\") (sys::%saveinitmem) (exit)"

fails with:

sh: 1009 Memory fault - core dumped
*** Error code 139
make: Fatal error: Command failed for target `interpreted.mem'


I am compiling clisp with gcc-2.7.2

This is the CFLAGS from the makefile:

CFLAGS=-W -Wswitch -Wpointer-arith -Wimplicit -Wreturn-type -fomit-frame-pointer -DNO_READLINE -Dunix -Dsun -Dsun386 -DNO_GENERATIONAL_GC -DNO_MULTIMAP_SHM -DNO_MULTIMAP_FILE -DNO_SINGLEMAP -DNO_TRIVIALMAP -DSAFETY=3

(I delete -O2 to avoid possible adverse strength-reduce effects)

I have also tried to compile with generational GC, multimap, shared memory 
and dynamic_ffi, in ALL possible combinations, without success. In fact,
when I try to use multimap, I get a SIGSEGV handle fault at the same
point, even after modifying 'unix.d' to make the fault handler arguments
consistent with the declaration of the function mmap() in SunOS4.0.2:

#if defined(UNIX_SUNOS4)
  #define FAULT_HANDLER_ARGLIST  sig, code, scp, addr
  #define FAULT_HANDLER_ARGDECL  int sig; int code; struct sigcontext* scp; char* addr;
  #define FAULT_ADDRESS  scp->sc_eax

During configure (configure --host=i386-sun-sunos4.0.2)  I see: (among others)

checking for the code address range... 0x00000000
checking for the malloc address range... 0x00000000
checking for the shared library address range... 0x00000000
checking for working mmap... yes
checking for mprotect... yes
checking for working shared memory...  yes

I have alse tested the definitions resulting from the flow of #if define's 
in 'lispbibl.d' where SPVW_PURE_BLOCKS is selected in my configuration
(which apparently is an installation bug if the comment for SPVW_PURE_PAGES 
in case of SUN4 and SUN386, in 'spvw.d' is true)

Anyway, I have tested PURE_PAGES, PURE_BLOCKS, and other combinations
with the same negative results. The same fault messages (with core)
appearing in making interpreted.mem.

Could any of you perhaps with the same machine, give me some help?. 
This thing is driving me nuts.

    /   __  __      /__ _ __/__  __  _
 __//_//_/_/ /   __//_// /_//_/_/ //_/_