timezone: Correct common_timezones dictionary.

The changes are as follows:

• Fix one day offset in all western zones.
• Correct CST from -64800 to -21600 and CDT from -68400 to -18000.
• Disambiguate PST in favor of -28000 over +28000.
• Add GMT, UTC, WET, previously excluded for being at offset 0.
• Add ACDT, AEDT, AKST, MET, MSK, NST, NZDT, PKT, which the previous
  code did not find.
• Remove numbered abbreviations -12, …, +14, which are unnecessary.
• Remove MSD and PKST, which are no longer used.

Hardcode the dict and verify it with a test, so that future
discrepancies won’t go silently unnoticed.

Signed-off-by: Anders Kaseorg <anders@zulip.com>

zerver/lib/timezone.py

@@ -1,7 +1,6 @@
-import datetime
 from functools import lru_cache
 from io import TextIOWrapper
-from typing import Any, Dict, Union
+from typing import Dict
 
 import pytz
 
@@ -21,42 +20,58 @@ def get_canonical_timezone_map() -> Dict[str, str]:
 def canonicalize_timezone(key: str) -> str:
     return get_canonical_timezone_map().get(key, key)
 
-# This method carefully trims a list of common timezones in the pytz
-# database and handles duplicate abbreviations in favor of the most
-# common/popular offset. The output of this can be directly passed as
-# tz_data to dateutil.parser. It takes about 25ms to run, so we want
-# to cache its results (while avoiding running it on process startup
-# since we only need it for Markdown rendering).
-@lru_cache(maxsize=None)
-def get_common_timezones() -> Dict[str, Union[int, Any]]:
-    tzdata = {}
-    normal = datetime.datetime(2009, 9, 1)  # Any random date is fine here.
-    for str in pytz.all_timezones:
-        tz = pytz.timezone(str)
-        timedelta = tz.utcoffset(normal)
-        if not timedelta:
-            continue
-        offset = timedelta.seconds
-        tz_name = tz.tzname(normal)
-        tzdata[tz_name] = offset
-        # Handle known duplicates/exceptions.
-        # IST: Asia/Kolkata and Europe/Dublin.
-        if tz_name == 'IST':
-            tzdata[tz_name] = 19800  # Asia/Kolkata
-        # CDT: America/AlmostAll and America/Havana.
-        if tz_name == 'CDT':
-            tzdata[tz_name] = -68400  # America/AlmostAll
-        # CST America/Belize -64800
-        # CST America/Costa_Rica -64800
-        # CST America/El_Salvador -64800
-        # CST America/Guatemala -64800
-        # CST America/Managua -64800
-        # CST America/Regina -64800
-        # CST America/Swift_Current -64800
-        # CST America/Tegucigalpa -64800
-        # CST Asia/Macau 28800
-        # CST Asia/Shanghai 28800
-        # CST Asia/Taipei 28800
-        if tz_name == 'CST':
-            tzdata[tz_name] = -64800  # America/All
-    return tzdata
+# Note: some of these abbreviations are fundamentally ambiguous (see
+# zerver/tests/test_timezone.py), so you should never rely on them as
+# anything more than a heuristic.
+common_timezones = {
+    "SST": -39600,
+    "HST": -36000,
+    "AKST": -32400,
+    "HDT": -32400,
+    "AKDT": -28800,
+    "PST": -28800,
+    "MST": -25200,
+    "PDT": -25200,
+    "CST": -21600,
+    "MDT": -21600,
+    "CDT": -18000,
+    "EST": -18000,
+    "AST": -14400,
+    "EDT": -14400,
+    "NST": -12600,
+    "ADT": -10800,
+    "NDT": -9000,
+    "GMT": 0,
+    "UTC": 0,
+    "WET": 0,
+    "BST": +3600,
+    "CET": +3600,
+    "MET": +3600,
+    "WAT": +3600,
+    "WEST": +3600,
+    "CAT": +7200,
+    "CEST": +7200,
+    "EET": +7200,
+    "MEST": +7200,
+    "SAST": +7200,
+    "EAT": +10800,
+    "EEST": +10800,
+    "IDT": +10800,
+    "MSK": +10800,
+    "PKT": +18000,
+    "IST": +19800,
+    "WIB": +25200,
+    "AWST": +28800,
+    "HKT": +28800,
+    "WITA": +28800,
+    "JST": +32400,
+    "KST": +32400,
+    "WIT": +32400,
+    "ACST": +34200,
+    "AEST": +36000,
+    "ChST": +36000,
+    "ACDT": +37800,
+    "AEDT": +39600,
+    "NZST": +43200,
+    "NZDT": +46800,
+}