https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120810
--- Comment #3 from Jonathan Wakely <redi at gcc dot gnu.org> --- (In reply to Jonathan Wakely from comment #1) > Frankly, std::time_get::do_date_order() seems unusable and should be > deprecated. "Returns no_order if the date format specified by ’x’ contains other variable components (e.g., Julian day, week number, week day)." Or even just a 4-digit year. And what if it contains just the %d and %m and %y components, but with different separators, e.g. %d.%m.%y? The spec doesn't seem to permit this to return no_order, but it's not one of the formats in table 100 (which do_date_order doesn't actually refer to). There's a non-normative footnote saying that no_order is a possible result for valid locales, which should be normative. I think we should remove the footnote and say "Returns no_order if the date format specified by ’x’ does not match one of the formats in Table [tab:locale.time.get.dogetdate]." Once we have an LWG issue I'll move these comments there.
