Re: [COMMITTERS] pgsql: Sync tzload() and tzparse() APIs with IANA release tzcode2016c.

On Tue, Mar 29, 2016 at 1:36 PM, Christian Ullrich <chris@chrullrich.net> wrote:
> * Tom Lane wrote:
>
>> Michael Paquier <michael.paquier@gmail.com> writes:
>>>
>>> Buildfarm-not-being-happy-status: woodloose, mastodon, thrips, jacana.
>>>
>>> http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=woodlouse&dt=2016-03-29%2000%3A42%3A08
>>> The origin of the problem is that, which prevents all the subsequent
>>> queries to fail:
>>>    SET TimeZone to 'UTC';
>>> + ERROR:  invalid value for parameter "TimeZone": "UTC"
>>
>>
>> Yeah.  I've been staring at that for awhile, but it's not clear where
>> the problem is.  There are a bunch of other SET TIME ZONE commands in
>> the regression tests, how is it that this trivial case fails on the
>> Windows critters?
>
>
> I think this is the reason, from the check log on woodlouse (jacana says the
> same in make style):
>
> Generating timezone files...release\zic\zic: Can't create
> C:/buildfarm/buildenv/HEAD/pgsql.build/tmp_install/share/timezone/US/Pacific-New:
> No such file or directory

Yes, I have bumped into that when running the build. And I think that
the error is in zic.c, in dolink() when performing a Link operation
because this parent directory is not created because of that:       if (link_errno == ENOENT || link_errno == ENOTSUP)
    {           if (!mkdirs(toname))               exit(EXIT_FAILURE);           retry_if_link_supported = true;
}
I think that we'd want here to check as well on EISDIR or EACCES...
Haven't checked yet though. I'll update this thread with hopefully a
patch.
-- 
Michael