[mythtv] [mythtv-commits] Ticket #5896: Allow NIS-format /etc/timezone

[Michael T. Dean]

On 01/03/2009 12:05 PM, Otto Kolsi wrote:
> Michael T. Dean wrote:
>   
>> Do you have an /etc/timezone file?  If so, what are its contents?  We 
>>     
> No, there's no /etc/timezone file.
>   
>> You would need to use the myth_time_zone.pl script (the myth_timezone.pl 
>> test script was "obsoleted" when I found out that time zone is actually 
>> 2 words, not one, and changed the protocol command to QUERY_TIME_ZONE).  
>>     
> Okay, it took awhile to find that is is attached to #5319. Here's the 
> output:
>
> $ ./myth_time_zone.pl
> Current Timezone Information
>        id: Europe/Helsinki
>    offset: 7200
>      time: 2009-01-03T18:46:12
>   
>> Or, if you have any remote backends, you can find out what time zone is 
>> in use in the remote backend's log file (assuming -v important is 
>>     
> I don't have remote backend but a remote frontend and in it's log there is:
>
> "Backend is running in Europe/Helsinki time zone."
>
>   
>> You don't have an /etc/timezone that's a link to/copy of a file in 
>> /usr/share/zoneinfo, do you?  I.e. does:
>>     
> No, it is not a link or copy, it does not exist at all.
>
> Actually, what seems to be happening in MythWeb is that in setTimezone() 
> the first if() condition is not met. So the QUERY_TIME_ZONE command is 
> actually never sent to backend (at least within that function). Maybe 
> the problem lies within MythWeb and not in the master backend?
>
> I tried quickly do some debugging with MythWeb and in setTimezone()
> $_SESSION['backend']['timezone']['value'] seems to have value "U" which 
> doesn't sound right..

Yeah.  My assumption was that the backend had a bad value due to some
unhandled configuration case; however, it seems like the value is good
in the backend.  My first guess, then--without having looked at any code
at all--is that this is a separate, unrelated issue in the MythWeb code
(as the backend has the proper time zone ID and the protocol portion
seems to work without issue--as the perl script uses the protocol to
retrieve the zone ID).

I'll try to take a look at it in a week or two when I get more time, but
I have a feeling that if it's really in the MythWeb code, it's likely a
very subtle issue that Rob Smith will be far more likely to find than I
would be.

Thanks for the additional info.

Mike