that will fail at least one of the tests (Feb 29th in a non-leap year), but i don't think i really care about that. the case i care about is not accepting ridiculous input. the chances of any given seconds-since-1970 passing all these tests seems basically zero, so that should be fine.
On Tue, Mar 28, 2017 at 1:51 PM, Rob Landley <r...@landley.net> wrote: > There was a thread back in January where I tried to figure out why %s > wasn't working in date -D, which resulted in me being mad at mktime and > strptime for being stupidly designed, petering out around here: > > http://lists.landley.net/pipermail/toybox-landley.net/2017-January/008829.html > > Along the way, a problem I hit (repeatedly) was that chkmktime() is > really brittle, most recently that timezones can drive it nuts. > > Here's the patch I have lying around to remove chkmktime and instead > just check the posix boundaries required by > http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/time.h.html but > I dunno if that fixes Elliott's original reasons for adding chkmktime in > the first place. > > Opinions? > > Rob > > _______________________________________________ > Toybox mailing list > Toybox@lists.landley.net > http://lists.landley.net/listinfo.cgi/toybox-landley.net > -- Elliott Hughes - http://who/enh - http://jessies.org/~enh/ Android native code/tools questions? Mail me/drop by/add me as a reviewer. _______________________________________________ Toybox mailing list Toybox@lists.landley.net http://lists.landley.net/listinfo.cgi/toybox-landley.net
