test_pytz_choibalsan failing with pytz 2024.2 (tzdata 2024b)

[Jun Omae]

See ​https://github.com/edgewall/trac/actions/runs/11005628298/job/30558646895#step:7:551

======================================================================
FAIL: test_pytz_choibalsan (trac.util.tests.datefmt.LocalTimezoneTestCase)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/jun66j5/src/tracdev/git/trac/util/tests/datefmt.py", line 2024, in test_pytz_choibalsan
    self._compare_pytz(tz, '1983-04-01T00:00')  # +2:00 (TZ change)
  File "/home/jun66j5/src/tracdev/git/trac/util/tests/datefmt.py", line 2010, in _compare_pytz
    self._compare_pytz_arithmetic(tz, dt_naive)
  File "/home/jun66j5/src/tracdev/git/trac/util/tests/datefmt.py", line 1974, in _compare_pytz_arithmetic
    self.assertEqual(dt_tz.isoformat(), dt_localtz.isoformat())
AssertionError: '1983-04-01T01:00:00+09:00' != '1983-04-01T02:00:00+10:00'
- 1983-04-01T01:00:00+09:00
?             ^        -
+ 1983-04-01T02:00:00+10:00
?             ^       +

[Jun Omae]

According to ​2024b release of tz code and data available - tz-announce - lists.iana.org:

   Changes to past timestamps

     Asia/Choibalsan is now an alias for Asia/Ulaanbaatar rather than
     being a separate Zone with differing behavior before April 2008.
     This seems better given our wildly conflicting information about
     Mongolia's time zone history.

So I'll remove datefmt tests with Asia/Choibalsan between pytz and libc.

[Jun Omae]

Removed test_pytz_choibalsan in [17841] and merged in [17842].