From: "Alvar Penning" <post@0x21.biz>
Subject: Re: newsyslog
To: "Matthias Kilian" <kili@outback.escape.de>
Cc: <tech@openbsd.org>
Date: Thu, 02 Apr 2026 08:20:24 +0200

On Wed Apr 1, 2026 at 11:30 PM CEST, Matthias Kilian wrote:
> Hi,
>
> On Wed, Apr 01, 2026 at 10:41:28PM +0200, Alvar Penning wrote:
>> diff --git a/newsyslog.c b/newsyslog.c
>> index 0be4ed259b9..d4a62194ab3 100644
>> --- a/newsyslog.c
>> +++ b/newsyslog.c
>> @@ -1251,12 +1251,12 @@ parseDWM(char *s)
>>  
>>  	nd = mtab[tm.tm_mon];
>>  
>> -	if (tm.tm_mon == 1) {
>> -		if (((tm.tm_year + 1900) % 4 == 0) &&
>> -		    ((tm.tm_year + 1900) % 100 != 0) &&
>> -		    ((tm.tm_year + 1900) % 400 == 0)) {
>> -			nd++;	/* leap year, 29 days in february */
>> -		}
>> +	/* leap year, 29 days in february */
>> +	if ((tm.tm_mon == 1) && ((tm.tm_year + 1900) % 4 == 0)) {
>> +		if ((tm.tm_year + 1900) % 100 != 0)
>> +			nd++; /* divisible by 4, not divisible by 100 */
>> +		else if ((tm.tm_year + 1900) % 400 == 0)
>> +			nd++; /* divisible by 4, 100 and 400 */
>>  	}
>>  	tm.tm_hour = tm.tm_min = tm.tm_sec = 0;
>
> While this looks correct to me, wouldn't something like this be
> better?

The effect would be the same. However, personally I find it a bit harder
to read. That's why I rewrote the logic. But I am fine with both.

>
> --- newsyslog.c	Tue Sep 23 08:09:56 2025
> +++ newsyslog.c.new	Wed Apr  1 23:28:07 2026
> @@ -1253,8 +1253,8 @@
>  
>  	if (tm.tm_mon == 1) {
>  		if (((tm.tm_year + 1900) % 4 == 0) &&
> -		    ((tm.tm_year + 1900) % 100 != 0) &&
> -		    ((tm.tm_year + 1900) % 400 == 0)) {
> +		    (((tm.tm_year + 1900) % 100 != 0) ||
> +		     ((tm.tm_year + 1900) % 400 == 0))) {
>  			nd++;	/* leap year, 29 days in february */
>  		}
>  	}
>
> Ciao,
> 	Kili