On 2016-04-26 23:37, Martin Pitt wrote: > Package: tzdata > Version: 2016d-2 > Severity: wishlist > > /etc/localtime got turned into a symlink in 2016a-1 (see bug #803144), > now that /usr gets mounted from the initrd. > > This now leaves /etc/timezone completely redundant, as you should get > exactly the same answer by readlink /etc/localtime -- and if you do > not get the same answer, you have an inconsistency. /etc/localtime > obviously "wins" for the actual clock (as that's what programs are > reading), but you presumably get the /etc/timezone value in some > "system configuration tool" packages which read /etc/timezone first. > > https://codesearch.debian.net/perpackage-results/%2Fetc%2Ftimezone > shows a fair number of hits, but it's actually not so bad: as > /etc/timezone is a Debianism and /etc/localtime the distro-agnostic > standard, a lot of software which isn't Debian specific already looks > at both and only falls back to /etc/timezone if /etc/localtime does > not exist. A desktop installation boots fine and with the correct > time/zone after removing /etc/timezone. Unfortunately it seems a lot of software are actually using /etc/timezone to configure the time zone. When switching to a symlink this made people unhappy, see for example bug#813226. This bug points to the following query: https://github.com/search?q=%2Fetc%2Ftimezone+dpkg-reconfigure+noninteractive+tzdata&type=Code&utf8=%E2%9C%93 > If you are generally open to the idea, I can look through the above > codesearch results more closely to see which packages need fixing, > file bugs, and block this bug on those. But as that's quite some work, > I'd first like to discuss this with you. What we can probably do is to stop looking or creating /etc/timezone if it doesn't exist, but keep updating it if it exists. What do you think? Aurelien -- Aurelien Jarno GPG: 4096R/1DDD8C9B aurelien@aurel32.net http://www.aurel32.net
Attachment:
signature.asc
Description: PGP signature