Mercurial > hg > xemacs-beta
diff tests/automated/lisp-tests.el @ 5912:47ffa085a9ad
Fix a bug when passing a float to truncate and giving a bignum result.
src/ChangeLog addition:
2015-05-10 Aidan Kehoe <kehoea@parhasard.net>
* floatfns.c (truncate_one_float):
Fix a bug here when double_to_integer() (and previously
float_to_int()) returned a bignum; this was silently treated as a
fixnum, giving consistently incorrect remainders and less
consistently-incorrect but still buggy divisors.
tests/ChangeLog addition:
Check that a bug when supplying #'truncate with one float that
rounds to a bignum, has been fixed.
| author | Aidan Kehoe <kehoea@parhasard.net> |
|---|---|
| date | Sun, 10 May 2015 23:55:41 +0100 |
| parents | 48386fd60fd0 |
| children | 1af53d35dd53 |
line wrap: on
line diff
--- a/tests/automated/lisp-tests.el Sun May 10 19:07:09 2015 +0100 +++ b/tests/automated/lisp-tests.el Sun May 10 23:55:41 2015 +0100 @@ -1986,6 +1986,9 @@ (when (featurep 'bignum) (assert (not (evenp most-positive-fixnum)) t "In the unlikely event that most-positive-fixnum is even, rewrite this.") + (Assert (equal (multiple-value-list (truncate (+ most-positive-fixnum 2.0))) + (list (+ most-positive-fixnum 2) 0.0)) + "checking a bug in single-argument truncate's remainder fixed") (Assert-rounding (1+ most-positive-fixnum) (* 2 most-positive-fixnum) :one-floor-result `(,(1+ most-positive-fixnum) 0) :two-floor-result `(0 ,(1+ most-positive-fixnum))