On Thursday, March 13, 2008 at 22:24:36 +0100, Karel Zak wrote: > I really don't like this ADJTIME_PATH idea, environment variables > suck... Do you have any arguments, besides your own taste? I mean: I also am not a big fan of env vars in general. However ADJTIME_PATH seems quite well adapted to this specific case. > I think command line option and a global config file (e.g > /etc/hwclock) are a good idea. I already exposed some negative points of --adjfile, and did not hear them contested (well, appart the "admins dont really run hwclock by hand" statement, obviously discussable). The /etc/hwclock.conf idea seems very good, it gracefully suits the specifics of the problem, and could as a bonus carry other options. However there are arguments against it: - Create a new fixed location file to help moving another fixed location file, is... strange. And someone will surely ask how to define another place for /etc/hwclock.conf. That's guaranteed. ;-) - Supporting /etc/hwclock.conf will need to add parsing code, deal with errors, define sensible fallbacks, define and document a syntax, and all that. This will make hwclock bigger, and more complex. Is it worth to pay this price for a feature, moving the adjfile to a non-default place, that will never be used by 99.99% of users? In comparision, ADJTIME_PATH or --adjfile each cost 3 lines of code, and 2 lines of manpage. Alain. -- To unsubscribe from this list: send the line "unsubscribe util-linux-ng" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
