I now believe, after discussion with the binutils developers, that the problem is that -zcombreloc just plain doesn't work in any release of binutils. One implication of this fact is that turning on -znocombreloc when building python is insufficient to make it work if you built gcc with -zcombreloc, because that build will write dynamic libraries that the Sun loader cannot handle no matter what.

So I think you are right -- the correct fix is to warn people that they should not use binutils 2.13 on Solaris, period. I believe that they will fix the problem in 2.13.1 and will let you know.

Meanwhile, might I suggest including the following test somewhere in the build procedure? If it fails, I believe it will not be possible to build Python successfully, so one might as well find about it early....

If you run this and the output includes "core dumped', it failed :-)

-------------------------cut here---------------------- #! /bin/sh

mkdir /tmp/t.$$ || exit 3 cd /tmp/t.$$ || exit 3

cat >main.c <<'EOF' #include <stdio.h> #include <dlfcn.h>

int main(void) { void *handle, *sym; char *error;

puts("calling dlopen"); handle = dlopen("./dyn.so", RTLD_NOW); if (!handle) { printf("%s\n", dlerror()); return 1; }

puts("calling dlsym"); sym = dlsym(handle, "sym"); if ((error = dlerror()) != 0) { printf("%s\n", error); return 1; } puts("calling sym"); ((void (*)(void))sym)(); puts("done"); return 0; } EOF

cat >dyn.c <<'EOF' #include <stdio.h> void sym(void) { puts("in sym"); } EOF

[ -n "$SHFLAGS" ] || SHFLAGS="-fPIC -shared" [ -n "$CC" ] || CC=gcc

set -x

$CC $CFLAGS $SHFLAGS dyn.c -o dyn.so $CC $CFLAGS main.c -o main -ldl

./main || exit $?

cd /tmp rm -rf t.$$