| Subject: | Simplify dow maths in _rd2ymd [PATCH] |
There's some slightly overcomplicated code in DateTime.xs with a comment
about the author being a "math gimp". This patch replaces it with
something slightly less mysterious which should always produce the same
result, and which passes all the same tests.
Being a math gimp myself I can't actually prove that this is correct,
but I'm fairly sure it is.
| Subject: | datetime-math-gimp.patch |
--- DateTime.xs.orig 2009-10-17 00:03:13.000000000 +0100
+++ DateTime.xs 2009-10-17 00:04:46.000000000 +0100
@@ -59,9 +59,6 @@
305,
335 };
-
-const IV neg_dow[] = { 1, 7, 6, 5, 4, 3, 2 };
-
IV
_real_is_leap_year(IV y) {
/* See http://www.perlmonks.org/?node_id=274247 for where this
@@ -123,17 +120,9 @@
if (extra) {
quarter = ( ( 1.0 / 3.1 ) * m ) + 1;
- dow = ((rd_days + 6) % 7);
-
- if (rd_days < -6) {
- /* if the left side of the modulus was negative we now
- have a bogus answer. This fixes it, though it ccould
- probably be done more elegantly if I wasn't such a math
- gimp. */
-
- dow = neg_dow[ abs(dow) ];
- } else {
- dow++;
+ dow = rd_days % 7;
+ if (dow <= 0) {
+ dow += 7;
}
PUSHs(sv_2mortal(newSViv(dow)));