new lognot

* new lognot
@ 2003-08-25 22:11 Kevin Ryde
  2003-08-31 17:52 ` Marius Vollmer
  0 siblings, 1 reply; 2+ messages in thread
From: Kevin Ryde @ 2003-08-25 22:11 UTC (permalink / raw)

I'd like to suggest the rewrite below for lognot.  I think it'd be a
good thing if lognot behaved like logand etc and accepted only
integers.  The current use of scm_difference means a flonum or complex
will work, but they're really not sensible operands.

The use of mpz_com is a bit more direct too, though in fact it won't
be much different speed-wise from mpz_sub.

SCM_DEFINE (scm_lognot, "lognot", 1, 0, 0,
            (SCM n),
	    "Return the integer which is the ones-complement of the integer\n"
	    "argument.\n"
	    "\n"
	    "@lisp\n"
	    "(number->string (lognot #b10000000) 2)\n"
	    "   @result{} \"-10000001\"\n"
	    "(number->string (lognot #b0) 2)\n"
	    "   @result{} \"-1\"\n"
	    "@end lisp")
#define FUNC_NAME s_scm_lognot
{
  if (SCM_INUMP (n)) {
    /* No overflow here, just need to toggle all the bits making up the inum.
       Enhancement: No need to unpack and put a new tag etc, could just xor
       a block of 1 bits.  */
    return SCM_MAKINUM (~ SCM_INUM (n));

  } else if (SCM_BIGP (n)) {
    SCM result = scm_i_mkbig ();
    mpz_com (SCM_I_BIG_MPZ (result), SCM_I_BIG_MPZ (n));
    scm_remember_upto_here_1 (n);
    return result;

  } else {
    SCM_WRONG_TYPE_ARG (SCM_ARG1, n);
  }
}
#undef FUNC_NAME

_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

^ permalink raw reply	[flat|nested] 2+ messages in thread