apparent bug in 7/26/06 version of 'tzload()'

Robert Elz kre at munnari.OZ.AU
Fri Jul 28 07:56:47 UTC 2006

    Date:        Thu, 27 Jul 2006 14:22:24 -0400
    From:        "Olson, Arthur David \(NIH/NCI\) [E]" <olsona at>
    Message-ID:  <B410D30A78C6404C9DABEA31B54A2813997ADE at>

  | Now if only C had an &&= operator...

Sometimes that would be useful, but it turns out, it would not be
needed here.   What's more, now those inappropriate &= can be removed
as well - the "if" that is added assures us that sp->goback (and
sp->goahead of course) are true before those expressions are evaluated.
And of course, true && x is identical to x, so all you need in those
expressions is a simple assignment now.

I'd also add extra parentheses in the "if".  Well, actually, I probably
wouldn't, but with the absurd pickiness (programmer protection, even when
neither wanted nor desired) of C compilers these days, it is probably a good
idea to make it be

	if ((sp->goback = sp->goahead = (sp->timecnt > i))) {

And even perhaps insert a != 0 test, just to show that it really is intended.

Or, if you were willing to make it be a little different style, write it
as ...

	sp->goback = sp->goahead = FALSE /* or 0 */ ;
	if (sp->timecnt > i) {
		sp->goback = sp->types[i] == sp->types[0] &&
 			differ_by_repeat(sp->ats[i], sp->ats[0]);
		sp->goahead =
 			sp->types[sp->timecnt - 1] == sp->types[sp->timecnt - 1 - i] &&
 			differ_by_repeat(sp->ats[sp->timecnt - 1],
 					 sp->ats[sp->timecnt - 1 - i]);


More information about the tz mailing list