Raymond Toy pushed to branch issue-156-take-2-nan-comparison at cmucl / cmucl

Commits:

• 64c9296f
  by Raymond Toy at 2023-03-09T19:49:21-08:00

  Fix errors in dd >= and <= implmeentation
  
  For `dd<=`, we accidentally used `dd>` instead of `dd<`.   Oops.
  
  The deftransforms had swapped the args because of a previous
  implementation that we changed, but we forgot to revert the swapping.
  
  Now `(describe decode-float)` correctly returns
  ```
    (VALUES
     (OR (SINGLE-FLOAT 0.5 (1.0))
         (DOUBLE-FLOAT 0.5d0 (1.0d0))
         (DOUBLE-DOUBLE-FLOAT 0.5w0 (1.0w0)))
     KERNEL:DOUBLE-FLOAT-EXPONENT
     (OR (SINGLE-FLOAT 1.0 1.0)
         (SINGLE-FLOAT -1.0 -1.0)
         (DOUBLE-FLOAT -1.0d0 -1.0d0)
         (DOUBLE-FLOAT 1.0d0 1.0d0)
         (DOUBLE-DOUBLE-FLOAT -1.0w0 -1.0w0)
         (DOUBLE-DOUBLE-FLOAT 1.0w0 1.0w0)))
  ```
  like we used to.
  

1 changed file:

• src/compiler/float-tran-dd.lisp

Changes:

 (declaim (inline dd<=))

 (defun dd<= (a0 a1 b0 b1)

-  (or (dd> a0 a1 b0 b1)

+  (or (dd< a0 a1 b0 b1)

       (dd= a0 a1 b0 b1)))

 (declaim (inline dd>=))

 	(kernel:double-double-lo b)))

 (deftransform <= ((a b) (vm::double-double-float vm::double-double-float) *)

-  `(dd<= (kernel:double-double-hi b)

-	 (kernel:double-double-lo b)

-	 (kernel:double-double-hi a)

-	 (kernel:double-double-lo a)))

+  `(dd<= (kernel:double-double-hi a)

+	 (kernel:double-double-lo a)

+	 (kernel:double-double-hi b)

+	 (kernel:double-double-lo b)))

 (deftransform >= ((a b) (vm::double-double-float vm::double-double-float) *)

-  `(dd>= (kernel:double-double-hi b)

-	 (kernel:double-double-lo b)

-	 (kernel:double-double-hi a)

-	 (kernel:double-double-lo a)))

+  `(dd>= (kernel:double-double-hi a)

+	 (kernel:double-double-lo a)

+	 (kernel:double-double-hi b)

+	 (kernel:double-double-lo b)))

 ) ; end progn