<snip> # upgrade timezone file if [[ ! -e ${ROOT}/etc/timezone ]] ; then ( source "${ROOT}"/etc/conf.d/timezone [[ -n ${TIMEZONE} ]] && echo "${TIMEZONE}" > "${ROOT}"/etc/timezone ) fi </snip> Well, /etc/conf.d/timezone never existed, the configuration has always been in /etc/conf.d/clock AFAICR. Plus the new /etc/timezone location is still incompatible w/ sys-libs/timezone-data pkg_{postinst,config}, so I don't understand where's this supposed to be configured now really.
fixed in cvcs http://sources.gentoo.org/sys-apps/openrc/openrc-9999.ebuild?r1=1.8&r2=1.9
Well thanks, but what's the /etc/timezone thing good for? As noted, nothing looks there, sys-libs/timezone-data still wants /etc/conf.d/clock.
I was hoping Mike could actually provide some more info on this file because I'd include it in my OpenRC ebuild. The OpenRC-9999 ebuild should move the creation of this file above the has_version check as well
Reopen; creating another file in /etc which isn't used anywhere doesn't make sense.
it is used elsewhere ... if it wasnt used at all, we wouldnt have created the variable in the first place
jakub, it's possible this file will be used in the future in ebuilds like timezone-data. Before we jump to conclusions, I've already contacted Mike to try and get more info from him on his intentions. E-mail was Message-ID: <47E94737.5030405@gentoo.org>
(In reply to comment #5) > it is used elsewhere ... if it wasnt used at all, we wouldnt have created the > variable in the first place Sigh... I'm a bit old for these riddle games. Could you enlighten us *where* it's being used? timezone-data ebuild mangles the timezone to Factory since the configuration no longer exists in /etc/conf.d/clock for openrc users. Where are people supposed to configure the timezone?
if you cant be bothered to keep up, then dont bother people who do keep up
(In reply to comment #8) > if you cant be bothered to keep up, then dont bother people who do keep up Yeah, great... Care to tell me where's this being used, if I can't keep up? I don't see anything newer anywhere in the tree. http://sources.gentoo.org/viewcvs.py/gentoo-x86/sys-libs/timezone-data/timezone-data-2008b.ebuild?rev=1.1&view=markup
why exactly do you feel you need an explanation in order for this to be fixed? you have never been involved with this sort of development. if you need someone to hold your hand, go ask on irc or the forums.
(In reply to comment #10) > why exactly do you feel you need an explanation in order for this to be fixed? > you have never been involved with this sort of development. if you need > someone to hold your hand, go ask on irc or the forums. Sigh. We obviously need to document it for our *users*, and we need the configuration file installed by default for *everyone* - with appropriate comments. Neither of which happens ATM, even after your last commit to timezone-data ebuild. How about fixing this bug properly finally?
Rather then slinging insults, it might have been worth while to actually check and see if the code was being used. Instead, insults we slung for 6 hours and then silently the issue was fixed... You might also want to stick an elog statement for the /etc/timezone code block since the /etc/conf.d/clock section has one to remove user confusion.
writing up documentation is a separate issue which Doug is tackling. that does not require justification of things to you. not that it really matters to users as the upgrade path is taken care of them, so they shouldnt really care. this bug is fixed properly: there was a typo in the openrc ebuild, now there isnt.
(In reply to comment #13) - that file won't be created for anyone who hasn't 'migrated' (IOW, fresh installs) as it is now... - if timezone-data is the only thing that uses such configuration file, then timezone-data ebuild should take care of creating it, not openrc which doesn't use it anywhere - while talking about documentation aspects, all the comments about setting timezone that used to exist in /etc/conf.d/clock are now lost