| 1 |
----- Original Message ----- |
| 2 |
From: "Mark Knecht" <markknecht@×××××.com> |
| 3 |
To: <gentoo-amd64@l.g.o> |
| 4 |
Sent: Wednesday, January 16, 2008 8:37 PM |
| 5 |
Subject: Re: [gentoo-amd64] Problem with latest timezone update? |
| 6 |
|
| 7 |
|
| 8 |
> On Jan 16, 2008 4:56 PM, Drake Donahue <donahue95@×××××××.net> wrote: |
| 9 |
>> |
| 10 |
>> ----- Original Message ----- |
| 11 |
>> From: "Steev Klimaszewski" <steev@g.o> |
| 12 |
>> To: <gentoo-amd64@l.g.o> |
| 13 |
>> Sent: Wednesday, January 16, 2008 6:01 PM |
| 14 |
>> Subject: Re: [gentoo-amd64] Problem with latest timezone update? |
| 15 |
>> |
| 16 |
>> |
| 17 |
>> <snip> |
| 18 |
>> |
| 19 |
>> > Except that neither etc-update nor dispatch-conf touch |
| 20 |
>> > the /etc/localtime file... |
| 21 |
>> > |
| 22 |
>> |
| 23 |
>> True, but... |
| 24 |
>> The files that are (maybe) affected directly by etc-update and/or |
| 25 |
>> dispatch-conf are /etc/conf.d/clock and /etc/init.d/clock. ('maybe' is |
| 26 |
>> used |
| 27 |
>> because: an emerge update must have affected one or both files; |
| 28 |
>> etc-update |
| 29 |
>> and/or dispatch-conf must have been invoked by the user; the user must |
| 30 |
>> have |
| 31 |
>> chosen action that resulted in a change to one or both files. |
| 32 |
>> /etc/conf.d/clock is the configuration file for /etc/init.d/clock. |
| 33 |
>> When /etc/init.d/clock runs (normally at boot), /etc/conf.d/clock is |
| 34 |
>> read. |
| 35 |
>> If CLOCK="UTC" is not set in /etc/conf.d/clock, |
| 36 |
> |
| 37 |
> UTC was set... |
| 38 |
|
| 39 |
Now CLOCK="LOCAL" ? |
| 40 |
|
| 41 |
> |
| 42 |
>> the option --localtime is |
| 43 |
>> used and TBLURB="Local Time" is set. |
| 44 |
>> The /etc/localtime file is a copy of one of the binary files in |
| 45 |
>> /usr/share/zoneinfo made by the system installer |
| 46 |
> |
| 47 |
> installer == Mark, me the guy who built the system, correct? |
| 48 |
|
| 49 |
yes. |
| 50 |
|
| 51 |
> |
| 52 |
>> initially; it is subject to |
| 53 |
>> update by repeating the manual copy process anytime after system install. |
| 54 |
> |
| 55 |
> Which is what I did today to fix this problem. |
| 56 |
> |
| 57 |
>> Thus etc-update and/or dispatch-conf can't change localtime; but can |
| 58 |
>> change |
| 59 |
>> whether localtime runs or not. |
| 60 |
> |
| 61 |
> Humm...seems like what you say is true but doesn't explain how emerge |
| 62 |
> -DuN system changed the file. It was clearly changed since I could |
| 63 |
> look inside with vi and compare to the Los_Angeles file and see that |
| 64 |
> they were clearly different... |
| 65 |
> |
| 66 |
|
| 67 |
Probably the previous /etc/localtime file was a copy of PST8PDT. |
| 68 |
IIRC the geographic file names are relatively recent in origin. |
| 69 |
If the old localtime had PST8PDT and the new had Los_Angeles in the few |
| 70 |
readable characters the difference is explained. |
| 71 |
|
| 72 |
Alternatively: |
| 73 |
The start and end of US daylight savings time changed effective 2007-2008. |
| 74 |
This resulted in an update to PST8PDT and its cousins ( to Los_Angeles and |
| 75 |
its geographic cousins also, if they existed before the change). The |
| 76 |
timezone-data ebuild also has a series of more recent bugfixes. |
| 77 |
|
| 78 |
As Nicholas explains the /sys-libs/timezone-data update ebuild should have |
| 79 |
updated /usr/share/zoneinfo files to new versions. |
| 80 |
The /sys-libs/timezone-data update ebuild should then have copied the new |
| 81 |
version of the file named in /etc/conf/clock's TIMEZONE= statement from |
| 82 |
/usr/share/zoneinfo to /etc/localtime. |
| 83 |
|
| 84 |
If it did not succeed in updating /etc/localtime because TIMEZONE= was |
| 85 |
blank, invalid, or not set, the old /etc/localtime would have contained |
| 86 |
readable characters: ' Local time zone must be set -- see zic manual page'. |
| 87 |
|
| 88 |
|
| 89 |
> Thanks, |
| 90 |
> Mark |
| 91 |
> -- |
| 92 |
> gentoo-amd64@l.g.o mailing list |
| 93 |
> |
| 94 |
|
| 95 |
-- |
| 96 |
gentoo-amd64@l.g.o mailing list |
Archives