Re: [9front] date -t vs rfc3339 i which is correct:

[ori@eigenstate.org]

> my interpretation is that this rfc3339 is a SUBSET of iso-8601 and we
> hit edge case in iso-8601 that is outside of rfc3339. we'r correct
> for iso 8601 but not for rfc3339. i think this should be changed
> so we cover both specifications and make date -t more widely usefull.
> 
> --- a/sys/src/cmd/date.c	Mon Dec 09 18:08:02 2019 +0100
> +++ b/sys/src/cmd/date.c	Tue Dec 10 20:15:42 2019 +0100
> @@ -15,7 +15,7 @@
>  char*
>  isodate(Tm *t)
>  {
> -	static char c[25]; /* leave room to append isotime */
> +	static char c[26]; /* leave room to append isotime */
>  	snprint(c, 11, "%04d-%02d-%02d", 
>  		t->year + 1900, t->mon + 1, t->mday);
>  	return c;
> @@ -39,7 +39,7 @@
>  			c[9] = '-';
>  			tz = -tz;
>  		}
> -		snprint(c+10, 5, "%02d%02d", tz / 60, tz % 60);
> +		snprint(c+10, 6, "%02d:%02d", tz / 60, tz % 60);
>  	} else {
>  		c[9] = 'Z';
>  		c[10] = 0;
> 
> --
> cinap

The change looks good to me, but one question:

Is there anything that uses the output of 'date -t' and cares?
(How did we run into this?)