On Mar 18, 2024, at 8:12 PM, Matt Johnson-Pint via tz <tz@iana.org> wrote:
What do y'all think?
This appears to be a Debianism picked up by Ubuntu, and Debian may not have informed users - and upstreams? - of the change clearly enough: https://groups.google.com/g/linux.debian.bugs.dist/c/H8u-kxaEB2Y https://github.com/nextcloud/server/issues/39767 https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1043250 It appears that Ubuntu may not have done so either. That, by itself, was probably not handled well - and, given the various places where TZ settings get stored, perhaps that change needed to be done more slowly, e.g. with some mechanism by which the C library's time zone loading code could catch references to old time zones and, somehow, log a message that says "this program tried using an old TZ setting, you might want to change it to use this setting rather than that old setting", so people have a chance to fix the deprecated settings before the flag day. (BTW, one of the comments in the Debian bug had “It is inappropriate to require that a time represented as seconds since the Epoch precisely represent the number of seconds between the referenced time and the Epoch.” -- IEEE Std 1003.1b-1993 (POSIX) Section B.2.2.2 as its signature line. I wonder whether my gripes back in the old days about "POSIX days are specified as being exactly 86400 seconds long" provoked that comment in the Rationale?)