Обсуждение: system timezone regression failure
Building 8.4 on a client's system, I get a regression failure apparently due to some difference between the system's timezone DB and what out regression tests expect, as shown below. I'm wondering if we should not disable the timestamptz regression test when we configure with the system timezone db. I guess i could patch the SRPM by providing an alternative result file - not sure if it's worth it though. cheers andrew *** /usr/local/src/rpms/BUILD/postgresql-8.4.0/src/test/regress/expected/timestamptz.out 2009-03-21 21:12:32.000000000 -0400 --- /usr/local/src/rpms/BUILD/postgresql-8.4.0/src/test/regress/results/timestamptz.out 2009-07-28 12:33:15.000000000 -0400 *************** *** 153,159 **** SELECT '20500710 173201 Europe/Helsinki'::timestamptz; -- DST timestamptz ------------------------------ ! Sun Jul 10 07:32:01 2050 PDT (1 row) SELECT '20500110 173201 Europe/Helsinki'::timestamptz; -- non-DST --- 153,159 ---- SELECT '20500710 173201 Europe/Helsinki'::timestamptz; -- DST timestamptz ------------------------------ ! Sun Jul 10 07:32:01 2050 PST (1 row) SELECT '20500110 173201 Europe/Helsinki'::timestamptz; -- non-DST *************** *** 165,171 **** SELECT '205000-07-10 17:32:01 Europe/Helsinki'::timestamptz; -- DST timestamptz -------------------------------- ! Thu Jul 10 07:32:01 205000 PDT (1 row) SELECT '205000-01-10 17:32:01 Europe/Helsinki'::timestamptz; -- non-DST --- 165,171 ---- SELECT '205000-07-10 17:32:01 Europe/Helsinki'::timestamptz; -- DST timestamptz -------------------------------- ! Thu Jul 10 07:32:01 205000 PST (1 row) SELECT '205000-01-10 17:32:01 Europe/Helsinki'::timestamptz; -- non-DST ======================================================================
Andrew Dunstan <andrew@dunslane.net> writes: > Building 8.4 on a client's system, I get a regression failure apparently > due to some difference between the system's timezone DB and what out > regression tests expect, as shown below. Those regression tests were *intentionally* set up to fail if someone's TZ support is not Y2038 clean. This is not a bug. Advise your client to get some less-obsolete timezone data files; or don't depend on the system TZ database. (The only reason why you should do so is if it's being kept up to date, eh?) regards, tom lane
Tom Lane wrote: > Andrew Dunstan <andrew@dunslane.net> writes: > >> Building 8.4 on a client's system, I get a regression failure apparently >> due to some difference between the system's timezone DB and what out >> regression tests expect, as shown below. >> > > Those regression tests were *intentionally* set up to fail if someone's > TZ support is not Y2038 clean. This is not a bug. Advise your client > to get some less-obsolete timezone data files; or don't depend on the > system TZ database. (The only reason why you should do so is if it's > being kept up to date, eh?) > > > Oh, you're right, I misread the diffs. The client is getting the machines updated. cheers andrew
Andrew Dunstan wrote: > > > Tom Lane wrote: >> Andrew Dunstan <andrew@dunslane.net> writes: >> >>> Building 8.4 on a client's system, I get a regression failure >>> apparently due to some difference between the system's timezone DB >>> and what out regression tests expect, as shown below. >>> >> >> Those regression tests were *intentionally* set up to fail if someone's >> TZ support is not Y2038 clean. This is not a bug. Advise your client >> to get some less-obsolete timezone data files; or don't depend on the >> system TZ database. (The only reason why you should do so is if it's >> being kept up to date, eh?) >> >> >> > > Oh, you're right, I misread the diffs. The client is getting the > machines updated. Well, this is interesting: andrew@jimbo:~> rpm -q -i timezone Name : timezone Relocations: (not relocatable) Version : 2.4 Vendor: SUSE LINUX Products GmbH, Nuernberg, Germany Release : 31.61 Build Date: Thu 02 Apr 2009 02:31:18 PM EDT Install Date: Wed 29 Jul 2009 10:52:55 AM EDT Build Host: cherubini.suse.de Group : System/Base Source RPM: glibc-2.4-31.61.src.rpm Size : 712873 License: BSD 3-Clause; GPL v2 or later; LGPL v2.1 or later Signature : DSA/SHA1, Fri 03 Apr 2009 09:20:05 AM EDT, Key ID a84edae89c800aca Packager : http://bugs.opensuse.org URL : http://www.gnu.org/software/libc/libc.html Summary : Timezone descriptions Description : These are configuration files that describe available time zones. You can select an appropriate time zone for your system with YaST. Distribution: SUSE Linux Enterprise 10 (X86-64) And I still get these errors. Looks like at least SUSE is not keeping up. I'll build without the system timezones. cheers andrew
Andrew Dunstan <andrew@dunslane.net> writes: > And I still get these errors. Looks like at least SUSE is not keeping > up. I'll build without the system timezones. Or just live with it. The purpose of the regression tests is to let you know there's a problem, not to dictate what you do about it. In this case you need to consider whether the application cares about post-Y2038 date arithmetic, and whether that's more important than having regularly updated timezone details. regards, tom lane