[mythtv] [mythtv-commits] Ticket #5791: Allow time zone ID's whose to differ by space/underscore

Michael T. Dean mtdean at thirdcontact.com
Wed Oct 8 23:15:41 UTC 2008


On 10/08/2008 06:49 PM, Michael T. Dean wrote:
> On 10/08/2008 06:05 PM, Simon Kenyon wrote:
>   
>> On Tue, 2008-10-07 at 19:32 +0000, MythTV wrote:
>>     
>>> #5791: Allow time zone ID's whose to differ by space/underscor
>> is this change anything to do with the fact that my FE will not start,
>> stating the following:
>>
>> 2008-10-08 22:58:42.270 Time zone settings on the master backend differ
>> from those on this system.
>> 2008-10-08 22:58:42.271 Detected time zone settings:
>>     Master: Zone ID: 'Eire', UTC Offset: '3600', Current Time:
>> '2008-10-08T22:58:42'
>>      Local: Zone ID: 'TZif2
> Is that exactly what the log file said?  If so, it seems that something 
> is wrong in the code as it should have given UTC Offset and Current time 
> for the frontend.

Nevermind.  Janne figured out where that info is coming from, and 
because there's a binary file where I had assumed a text file would be, 
the binary data I'm reading from that file is breaking the logging info 
(and the check for which zone rules are in use on your frontend 
system).  The only files we're reading data from are /etc/timezone and 
/etc/sysconfig/clock.

/etc/timezone is used on Debian-based systems (and they're the only ones 
I know of that use it).  It should be a text file that contains a single 
line with the name of the file (relative to the top directory of the 
zoneinfo database) with the name of the file used to set the time zone 
rules.  I.e. in your case, it should contain "Eire".

/etc/sysconfig/clock on a Red Hat-based system should contain a line:
ZONE=Eire

However, the code we have would prevent /etc/sysconfig/clock from being 
interpreted if it's a binary file, so as Janne indicated, it seems you 
have an invalid (from the Debian perspective) /etc/timezone.  I think 
Janne is currently working on adding code to ignore invalid /etc/timezone.

Mike


More information about the mythtv-dev mailing list