Re: The timezone oddities - Mailing list pgsql-general

From Tom Lane
Subject Re: The timezone oddities
Date
Msg-id 31574.1391553878@sss.pgh.pa.us
Whole thread Raw
In response to Re: The timezone oddities  (Adrian Klaver <adrian.klaver@gmail.com>)
Responses Re: The timezone oddities  (Rob Sargent <robjsargent@gmail.com>)
List pgsql-general
Adrian Klaver <adrian.klaver@gmail.com> writes:
> On 02/04/2014 12:31 PM, Rob Sargent wrote:
>> Perhaps building from source does make a guess at TZ. I am not residing
>> in the Navaho national territory, but is that just Mountain time?

> Yes:
> http://en.wikipedia.org/wiki/List_of_tz_database_time_zones
> The reason for specificity is that the Navajo Nation observes DST, while
> the rest of Arizona does not.

Not quite.  The zoneinfo database has a number of names for Mountain Time
with DST, eg America/Denver.  A quick look says that "Navajo" is just an
alias for that one, so no probe of the system's timezone behavior is going
to be able to tell the difference.  IIRC, given two zone names that both
match the system's behavior equally well, initdb chooses the shorter one.
That heuristic produces good results in many places, but not so much here.

Note that this is exactly the same choice of zone you'd have gotten from
pre-9.2 if it wasn't told a timezone setting to use.  We only moved this
examination of the system's behavior from every-postmaster-start to
initdb.

            regards, tom lane


pgsql-general by date:

Previous
From: AlexK
Date:
Subject: Re: Is it reasonable to store double[] arrays of 30K elements
Next
From: Rob Sargent
Date:
Subject: Re: The timezone oddities