cpython_2/solution/solution.patch

diff --git a/Lib/sysconfig.py b/Lib/sysconfig.py
index eaee837f10e..bf04ac541e6 100644
--- a/Lib/sysconfig.py
+++ b/Lib/sysconfig.py
@@ -546,6 +546,7 @@ def get_config_vars(*args):
 
         if os.name == 'nt':
             _init_non_posix(_CONFIG_VARS)
+            _CONFIG_VARS['TZPATH'] = ''
         if os.name == 'posix':
             _init_posix(_CONFIG_VARS)
         # For backward compatibility, see issue19555
diff --git a/Lib/zoneinfo/__init__.py b/Lib/zoneinfo/__init__.py
new file mode 100644
index 00000000000..81a2d5ea97b
--- /dev/null
+++ b/Lib/zoneinfo/__init__.py
@@ -0,0 +1,29 @@
+__all__ = [
+    "ZoneInfo",
+    "reset_tzpath",
+    "TZPATH",
+    "ZoneInfoNotFoundError",
+    "InvalidTZPathWarning",
+]
+
+from . import _tzpath
+from ._common import ZoneInfoNotFoundError
+
+try:
+    from _zoneinfo import ZoneInfo
+except ImportError:  # pragma: nocover
+    from ._zoneinfo import ZoneInfo
+
+reset_tzpath = _tzpath.reset_tzpath
+InvalidTZPathWarning = _tzpath.InvalidTZPathWarning
+
+
+def __getattr__(name):
+    if name == "TZPATH":
+        return _tzpath.TZPATH
+    else:
+        raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
+
+
+def __dir__():
+    return sorted(list(globals()) + ["TZPATH"])
diff --git a/Lib/zoneinfo/_common.py b/Lib/zoneinfo/_common.py
new file mode 100644
index 00000000000..3d35d4f4b46
--- /dev/null
+++ b/Lib/zoneinfo/_common.py
@@ -0,0 +1,166 @@
+import struct
+
+
+def load_tzdata(key):
+    import importlib.resources
+
+    components = key.split("/")
+    package_name = ".".join(["tzdata.zoneinfo"] + components[:-1])
+    resource_name = components[-1]
+
+    try:
+        return importlib.resources.open_binary(package_name, resource_name)
+    except (ImportError, FileNotFoundError, UnicodeEncodeError):
+        # There are three types of exception that can be raised that all amount
+        # to "we cannot find this key":
+        #
+        # ImportError: If package_name doesn't exist (e.g. if tzdata is not
+        #   installed, or if there's an error in the folder name like
+        #   Amrica/New_York)
+        # FileNotFoundError: If resource_name doesn't exist in the package
+        #   (e.g. Europe/Krasnoy)
+        # UnicodeEncodeError: If package_name or resource_name are not UTF-8,
+        #   such as keys containing a surrogate character.
+        raise ZoneInfoNotFoundError(f"No time zone found with key {key}")
+
+
+def load_data(fobj):
+    header = _TZifHeader.from_file(fobj)
+
+    if header.version == 1:
+        time_size = 4
+        time_type = "l"
+    else:
+        # Version 2+ has 64-bit integer transition times
+        time_size = 8
+        time_type = "q"
+
+        # Version 2+ also starts with a Version 1 header and data, which
+        # we need to skip now
+        skip_bytes = (
+            header.timecnt * 5  # Transition times and types
+            + header.typecnt * 6  # Local time type records
+            + header.charcnt  # Time zone designations
+            + header.leapcnt * 8  # Leap second records
+            + header.isstdcnt  # Standard/wall indicators
+            + header.isutcnt  # UT/local indicators
+        )
+
+        fobj.seek(skip_bytes, 1)
+
+        # Now we need to read the second header, which is not the same
+        # as the first
+        header = _TZifHeader.from_file(fobj)
+
+    typecnt = header.typecnt
+    timecnt = header.timecnt
+    charcnt = header.charcnt
+
+    # The data portion starts with timecnt transitions and indices
+    if timecnt:
+        trans_list_utc = struct.unpack(
+            f">{timecnt}{time_type}", fobj.read(timecnt * time_size)
+        )
+        trans_idx = struct.unpack(f">{timecnt}B", fobj.read(timecnt))
+    else:
+        trans_list_utc = ()
+        trans_idx = ()
+
+    # Read the ttinfo struct, (utoff, isdst, abbrind)
+    if typecnt:
+        utcoff, isdst, abbrind = zip(
+            *(struct.unpack(">lbb", fobj.read(6)) for i in range(typecnt))
+        )
+    else:
+        utcoff = ()
+        isdst = ()
+        abbrind = ()
+
+    # Now read the abbreviations. They are null-terminated strings, indexed
+    # not by position in the array but by position in the unsplit
+    # abbreviation string. I suppose this makes more sense in C, which uses
+    # null to terminate the strings, but it's inconvenient here...
+    char_total = 0
+    abbr_vals = {}
+    abbr_chars = fobj.read(charcnt)
+
+    def get_abbr(idx):
+        # Gets a string starting at idx and running until the next \x00
+        #
+        # We cannot pre-populate abbr_vals by splitting on \x00 because there
+        # are some zones that use subsets of longer abbreviations, like so:
+        #
+        #  LMT\x00AHST\x00HDT\x00
+        #
+        # Where the idx to abbr mapping should be:
+        #
+        # {0: "LMT", 4: "AHST", 5: "HST", 9: "HDT"}
+        if idx not in abbr_vals:
+            span_end = abbr_chars.find(b"\x00", idx)
+            abbr_vals[idx] = abbr_chars[idx:span_end].decode()
+
+        return abbr_vals[idx]
+
+    abbr = tuple(get_abbr(idx) for idx in abbrind)
+
+    # The remainder of the file consists of leap seconds (currently unused) and
+    # the standard/wall and ut/local indicators, which are metadata we don't need.
+    # In version 2 files, we need to skip the unnecessary data to get at the TZ string:
+    if header.version >= 2:
+        # Each leap second record has size (time_size + 4)
+        skip_bytes = header.isutcnt + header.isstdcnt + header.leapcnt * 12
+        fobj.seek(skip_bytes, 1)
+
+        c = fobj.read(1)  # Should be \n
+        assert c == b"\n", c
+
+        tz_bytes = b""
+        while (c := fobj.read(1)) != b"\n":
+            tz_bytes += c
+
+        tz_str = tz_bytes
+    else:
+        tz_str = None
+
+    return trans_idx, trans_list_utc, utcoff, isdst, abbr, tz_str
+
+
+class _TZifHeader:
+    __slots__ = [
+        "version",
+        "isutcnt",
+        "isstdcnt",
+        "leapcnt",
+        "timecnt",
+        "typecnt",
+        "charcnt",
+    ]
+
+    def __init__(self, *args):
+        assert len(self.__slots__) == len(args)
+        for attr, val in zip(self.__slots__, args):
+            setattr(self, attr, val)
+
+    @classmethod
+    def from_file(cls, stream):
+        # The header starts with a 4-byte "magic" value
+        if stream.read(4) != b"TZif":
+            raise ValueError("Invalid TZif file: magic not found")
+
+        _version = stream.read(1)
+        if _version == b"\x00":
+            version = 1
+        else:
+            version = int(_version)
+        stream.read(15)
+
+        args = (version,)
+
+        # Slots are defined in the order that the bytes are arranged
+        args = args + struct.unpack(">6l", stream.read(24))
+
+        return cls(*args)
+
+
+class ZoneInfoNotFoundError(KeyError):
+    """Exception raised when a ZoneInfo key is not found."""
diff --git a/Lib/zoneinfo/_tzpath.py b/Lib/zoneinfo/_tzpath.py
new file mode 100644
index 00000000000..8cff0b171bf
--- /dev/null
+++ b/Lib/zoneinfo/_tzpath.py
@@ -0,0 +1,110 @@
+import os
+import sys
+import sysconfig
+
+
+def reset_tzpath(to=None):
+    global TZPATH
+
+    tzpaths = to
+    if tzpaths is not None:
+        if isinstance(tzpaths, (str, bytes)):
+            raise TypeError(
+                f"tzpaths must be a list or tuple, "
+                + f"not {type(tzpaths)}: {tzpaths!r}"
+            )
+        elif not all(map(os.path.isabs, tzpaths)):
+            raise ValueError(_get_invalid_paths_message(tzpaths))
+        base_tzpath = tzpaths
+    else:
+        env_var = os.environ.get("PYTHONTZPATH", None)
+        if env_var is not None:
+            base_tzpath = _parse_python_tzpath(env_var)
+        else:
+            base_tzpath = _parse_python_tzpath(
+                sysconfig.get_config_var("TZPATH")
+            )
+
+    TZPATH = tuple(base_tzpath)
+
+
+def _parse_python_tzpath(env_var):
+    if not env_var:
+        return ()
+
+    raw_tzpath = env_var.split(os.pathsep)
+    new_tzpath = tuple(filter(os.path.isabs, raw_tzpath))
+
+    # If anything has been filtered out, we will warn about it
+    if len(new_tzpath) != len(raw_tzpath):
+        import warnings
+
+        msg = _get_invalid_paths_message(raw_tzpath)
+
+        warnings.warn(
+            "Invalid paths specified in PYTHONTZPATH environment variable."
+            + msg,
+            InvalidTZPathWarning,
+        )
+
+    return new_tzpath
+
+
+def _get_invalid_paths_message(tzpaths):
+    invalid_paths = (path for path in tzpaths if not os.path.isabs(path))
+
+    prefix = "\n    "
+    indented_str = prefix + prefix.join(invalid_paths)
+
+    return (
+        "Paths should be absolute but found the following relative paths:"
+        + indented_str
+    )
+
+
+def find_tzfile(key):
+    """Retrieve the path to a TZif file from a key."""
+    _validate_tzfile_path(key)
+    for search_path in TZPATH:
+        filepath = os.path.join(search_path, key)
+        if os.path.isfile(filepath):
+            return filepath
+
+    return None
+
+
+_TEST_PATH = os.path.normpath(os.path.join("_", "_"))[:-1]
+
+
+def _validate_tzfile_path(path, _base=_TEST_PATH):
+    if os.path.isabs(path):
+        raise ValueError(
+            f"ZoneInfo keys may not be absolute paths, got: {path}"
+        )
+
+    # We only care about the kinds of path normalizations that would change the
+    # length of the key - e.g. a/../b -> a/b, or a/b/ -> a/b. On Windows,
+    # normpath will also change from a/b to a\b, but that would still preserve
+    # the length.
+    new_path = os.path.normpath(path)
+    if len(new_path) != len(path):
+        raise ValueError(
+            f"ZoneInfo keys must be normalized relative paths, got: {path}"
+        )
+
+    resolved = os.path.normpath(os.path.join(_base, new_path))
+    if not resolved.startswith(_base):
+        raise ValueError(
+            f"ZoneInfo keys must refer to subdirectories of TZPATH, got: {path}"
+        )
+
+
+del _TEST_PATH
+
+
+class InvalidTZPathWarning(RuntimeWarning):
+    """Warning raised if an invalid path is specified in PYTHONTZPATH."""
+
+
+TZPATH = ()
+reset_tzpath()
diff --git a/Lib/zoneinfo/_zoneinfo.py b/Lib/zoneinfo/_zoneinfo.py
new file mode 100644
index 00000000000..69133ae80a4
--- /dev/null
+++ b/Lib/zoneinfo/_zoneinfo.py
@@ -0,0 +1,755 @@
+import bisect
+import calendar
+import collections
+import functools
+import os
+import re
+import struct
+import sys
+import weakref
+from datetime import datetime, timedelta, timezone, tzinfo
+
+from . import _common, _tzpath
+
+EPOCH = datetime(1970, 1, 1)
+EPOCHORDINAL = datetime(1970, 1, 1).toordinal()
+
+# It is relatively expensive to construct new timedelta objects, and in most
+# cases we're looking at the same deltas, like integer numbers of hours, etc.
+# To improve speed and memory use, we'll keep a dictionary with references
+# to the ones we've already used so far.
+#
+# Loading every time zone in the 2020a version of the time zone database
+# requires 447 timedeltas, which requires approximately the amount of space
+# that ZoneInfo("America/New_York") with 236 transitions takes up, so we will
+# set the cache size to 512 so that in the common case we always get cache
+# hits, but specifically crafted ZoneInfo objects don't leak arbitrary amounts
+# of memory.
+@functools.lru_cache(maxsize=512)
+def _load_timedelta(seconds):
+    return timedelta(seconds=seconds)
+
+
+class ZoneInfo(tzinfo):
+    _strong_cache_size = 8
+    _strong_cache = collections.OrderedDict()
+    _weak_cache = weakref.WeakValueDictionary()
+    __module__ = "zoneinfo"
+
+    def __init_subclass__(cls):
+        cls._strong_cache = collections.OrderedDict()
+        cls._weak_cache = weakref.WeakValueDictionary()
+
+    def __new__(cls, key):
+        instance = cls._weak_cache.get(key, None)
+        if instance is None:
+            instance = cls._weak_cache.setdefault(key, cls._new_instance(key))
+            instance._from_cache = True
+
+        # Update the "strong" cache
+        cls._strong_cache[key] = cls._strong_cache.pop(key, instance)
+
+        if len(cls._strong_cache) > cls._strong_cache_size:
+            cls._strong_cache.popitem(last=False)
+
+        return instance
+
+    @classmethod
+    def no_cache(cls, key):
+        obj = cls._new_instance(key)
+        obj._from_cache = False
+
+        return obj
+
+    @classmethod
+    def _new_instance(cls, key):
+        obj = super().__new__(cls)
+        obj._key = key
+        obj._file_path = obj._find_tzfile(key)
+
+        if obj._file_path is not None:
+            file_obj = open(obj._file_path, "rb")
+        else:
+            file_obj = _common.load_tzdata(key)
+
+        with file_obj as f:
+            obj._load_file(f)
+
+        return obj
+
+    @classmethod
+    def from_file(cls, fobj, /, key=None):
+        obj = super().__new__(cls)
+        obj._key = key
+        obj._file_path = None
+        obj._load_file(fobj)
+        obj._file_repr = repr(fobj)
+
+        # Disable pickling for objects created from files
+        obj.__reduce__ = obj._file_reduce
+
+        return obj
+
+    @classmethod
+    def clear_cache(cls, *, only_keys=None):
+        if only_keys is not None:
+            for key in only_keys:
+                cls._weak_cache.pop(key, None)
+                cls._strong_cache.pop(key, None)
+
+        else:
+            cls._weak_cache.clear()
+            cls._strong_cache.clear()
+
+    @property
+    def key(self):
+        return self._key
+
+    def utcoffset(self, dt):
+        return self._find_trans(dt).utcoff
+
+    def dst(self, dt):
+        return self._find_trans(dt).dstoff
+
+    def tzname(self, dt):
+        return self._find_trans(dt).tzname
+
+    def fromutc(self, dt):
+        """Convert from datetime in UTC to datetime in local time"""
+
+        if not isinstance(dt, datetime):
+            raise TypeError("fromutc() requires a datetime argument")
+        if dt.tzinfo is not self:
+            raise ValueError("dt.tzinfo is not self")
+
+        timestamp = self._get_local_timestamp(dt)
+        num_trans = len(self._trans_utc)
+
+        if num_trans >= 1 and timestamp < self._trans_utc[0]:
+            tti = self._tti_before
+            fold = 0
+        elif (
+            num_trans == 0 or timestamp > self._trans_utc[-1]
+        ) and not isinstance(self._tz_after, _ttinfo):
+            tti, fold = self._tz_after.get_trans_info_fromutc(
+                timestamp, dt.year
+            )
+        elif num_trans == 0:
+            tti = self._tz_after
+            fold = 0
+        else:
+            idx = bisect.bisect_right(self._trans_utc, timestamp)
+
+            if num_trans > 1 and timestamp >= self._trans_utc[1]:
+                tti_prev, tti = self._ttinfos[idx - 2 : idx]
+            elif timestamp > self._trans_utc[-1]:
+                tti_prev = self._ttinfos[-1]
+                tti = self._tz_after
+            else:
+                tti_prev = self._tti_before
+                tti = self._ttinfos[0]
+
+            # Detect fold
+            shift = tti_prev.utcoff - tti.utcoff
+            fold = shift.total_seconds() > timestamp - self._trans_utc[idx - 1]
+        dt += tti.utcoff
+        if fold:
+            return dt.replace(fold=1)
+        else:
+            return dt
+
+    def _find_trans(self, dt):
+        if dt is None:
+            if self._fixed_offset:
+                return self._tz_after
+            else:
+                return _NO_TTINFO
+
+        ts = self._get_local_timestamp(dt)
+
+        lt = self._trans_local[dt.fold]
+
+        num_trans = len(lt)
+
+        if num_trans and ts < lt[0]:
+            return self._tti_before
+        elif not num_trans or ts > lt[-1]:
+            if isinstance(self._tz_after, _TZStr):
+                return self._tz_after.get_trans_info(ts, dt.year, dt.fold)
+            else:
+                return self._tz_after
+        else:
+            # idx is the transition that occurs after this timestamp, so we
+            # subtract off 1 to get the current ttinfo
+            idx = bisect.bisect_right(lt, ts) - 1
+            assert idx >= 0
+            return self._ttinfos[idx]
+
+    def _get_local_timestamp(self, dt):
+        return (
+            (dt.toordinal() - EPOCHORDINAL) * 86400
+            + dt.hour * 3600
+            + dt.minute * 60
+            + dt.second
+        )
+
+    def __str__(self):
+        if self._key is not None:
+            return f"{self._key}"
+        else:
+            return repr(self)
+
+    def __repr__(self):
+        if self._key is not None:
+            return f"{self.__class__.__name__}(key={self._key!r})"
+        else:
+            return f"{self.__class__.__name__}.from_file({self._file_repr})"
+
+    def __reduce__(self):
+        return (self.__class__._unpickle, (self._key, self._from_cache))
+
+    def _file_reduce(self):
+        import pickle
+
+        raise pickle.PicklingError(
+            "Cannot pickle a ZoneInfo file created from a file stream."
+        )
+
+    @classmethod
+    def _unpickle(cls, key, from_cache, /):
+        if from_cache:
+            return cls(key)
+        else:
+            return cls.no_cache(key)
+
+    def _find_tzfile(self, key):
+        return _tzpath.find_tzfile(key)
+
+    def _load_file(self, fobj):
+        # Retrieve all the data as it exists in the zoneinfo file
+        trans_idx, trans_utc, utcoff, isdst, abbr, tz_str = _common.load_data(
+            fobj
+        )
+
+        # Infer the DST offsets (needed for .dst()) from the data
+        dstoff = self._utcoff_to_dstoff(trans_idx, utcoff, isdst)
+
+        # Convert all the transition times (UTC) into "seconds since 1970-01-01 local time"
+        trans_local = self._ts_to_local(trans_idx, trans_utc, utcoff)
+
+        # Construct `_ttinfo` objects for each transition in the file
+        _ttinfo_list = [
+            _ttinfo(
+                _load_timedelta(utcoffset), _load_timedelta(dstoffset), tzname
+            )
+            for utcoffset, dstoffset, tzname in zip(utcoff, dstoff, abbr)
+        ]
+
+        self._trans_utc = trans_utc
+        self._trans_local = trans_local
+        self._ttinfos = [_ttinfo_list[idx] for idx in trans_idx]
+
+        # Find the first non-DST transition
+        for i in range(len(isdst)):
+            if not isdst[i]:
+                self._tti_before = _ttinfo_list[i]
+                break
+        else:
+            if self._ttinfos:
+                self._tti_before = self._ttinfos[0]
+            else:
+                self._tti_before = None
+
+        # Set the "fallback" time zone
+        if tz_str is not None and tz_str != b"":
+            self._tz_after = _parse_tz_str(tz_str.decode())
+        else:
+            if not self._ttinfos and not _ttinfo_list:
+                raise ValueError("No time zone information found.")
+
+            if self._ttinfos:
+                self._tz_after = self._ttinfos[-1]
+            else:
+                self._tz_after = _ttinfo_list[-1]
+
+        # Determine if this is a "fixed offset" zone, meaning that the output
+        # of the utcoffset, dst and tzname functions does not depend on the
+        # specific datetime passed.
+        #
+        # We make three simplifying assumptions here:
+        #
+        # 1. If _tz_after is not a _ttinfo, it has transitions that might
+        #    actually occur (it is possible to construct TZ strings that
+        #    specify STD and DST but no transitions ever occur, such as
+        #    AAA0BBB,0/0,J365/25).
+        # 2. If _ttinfo_list contains more than one _ttinfo object, the objects
+        #    represent different offsets.
+        # 3. _ttinfo_list contains no unused _ttinfos (in which case an
+        #    otherwise fixed-offset zone with extra _ttinfos defined may
+        #    appear to *not* be a fixed offset zone).
+        #
+        # Violations to these assumptions would be fairly exotic, and exotic
+        # zones should almost certainly not be used with datetime.time (the
+        # only thing that would be affected by this).
+        if len(_ttinfo_list) > 1 or not isinstance(self._tz_after, _ttinfo):
+            self._fixed_offset = False
+        elif not _ttinfo_list:
+            self._fixed_offset = True
+        else:
+            self._fixed_offset = _ttinfo_list[0] == self._tz_after
+
+    @staticmethod
+    def _utcoff_to_dstoff(trans_idx, utcoffsets, isdsts):
+        # Now we must transform our ttis and abbrs into `_ttinfo` objects,
+        # but there is an issue: .dst() must return a timedelta with the
+        # difference between utcoffset() and the "standard" offset, but
+        # the "base offset" and "DST offset" are not encoded in the file;
+        # we can infer what they are from the isdst flag, but it is not
+        # sufficient to to just look at the last standard offset, because
+        # occasionally countries will shift both DST offset and base offset.
+
+        typecnt = len(isdsts)
+        dstoffs = [0] * typecnt  # Provisionally assign all to 0.
+        dst_cnt = sum(isdsts)
+        dst_found = 0
+
+        for i in range(1, len(trans_idx)):
+            if dst_cnt == dst_found:
+                break
+
+            idx = trans_idx[i]
+
+            dst = isdsts[idx]
+
+            # We're only going to look at daylight saving time
+            if not dst:
+                continue
+
+            # Skip any offsets that have already been assigned
+            if dstoffs[idx] != 0:
+                continue
+
+            dstoff = 0
+            utcoff = utcoffsets[idx]
+
+            comp_idx = trans_idx[i - 1]
+
+            if not isdsts[comp_idx]:
+                dstoff = utcoff - utcoffsets[comp_idx]
+
+            if not dstoff and idx < (typecnt - 1):
+                comp_idx = trans_idx[i + 1]
+
+                # If the following transition is also DST and we couldn't
+                # find the DST offset by this point, we're going ot have to
+                # skip it and hope this transition gets assigned later
+                if isdsts[comp_idx]:
+                    continue
+
+                dstoff = utcoff - utcoffsets[comp_idx]
+
+            if dstoff:
+                dst_found += 1
+                dstoffs[idx] = dstoff
+        else:
+            # If we didn't find a valid value for a given index, we'll end up
+            # with dstoff = 0 for something where `isdst=1`. This is obviously
+            # wrong - one hour will be a much better guess than 0
+            for idx in range(typecnt):
+                if not dstoffs[idx] and isdsts[idx]:
+                    dstoffs[idx] = 3600
+
+        return dstoffs
+
+    @staticmethod
+    def _ts_to_local(trans_idx, trans_list_utc, utcoffsets):
+        """Generate number of seconds since 1970 *in the local time*.
+
+        This is necessary to easily find the transition times in local time"""
+        if not trans_list_utc:
+            return [[], []]
+
+        # Start with the timestamps and modify in-place
+        trans_list_wall = [list(trans_list_utc), list(trans_list_utc)]
+
+        if len(utcoffsets) > 1:
+            offset_0 = utcoffsets[0]
+            offset_1 = utcoffsets[trans_idx[0]]
+            if offset_1 > offset_0:
+                offset_1, offset_0 = offset_0, offset_1
+        else:
+            offset_0 = offset_1 = utcoffsets[0]
+
+        trans_list_wall[0][0] += offset_0
+        trans_list_wall[1][0] += offset_1
+
+        for i in range(1, len(trans_idx)):
+            offset_0 = utcoffsets[trans_idx[i - 1]]
+            offset_1 = utcoffsets[trans_idx[i]]
+
+            if offset_1 > offset_0:
+                offset_1, offset_0 = offset_0, offset_1
+
+            trans_list_wall[0][i] += offset_0
+            trans_list_wall[1][i] += offset_1
+
+        return trans_list_wall
+
+
+class _ttinfo:
+    __slots__ = ["utcoff", "dstoff", "tzname"]
+
+    def __init__(self, utcoff, dstoff, tzname):
+        self.utcoff = utcoff
+        self.dstoff = dstoff
+        self.tzname = tzname
+
+    def __eq__(self, other):
+        return (
+            self.utcoff == other.utcoff
+            and self.dstoff == other.dstoff
+            and self.tzname == other.tzname
+        )
+
+    def __repr__(self):  # pragma: nocover
+        return (
+            f"{self.__class__.__name__}"
+            + f"({self.utcoff}, {self.dstoff}, {self.tzname})"
+        )
+
+
+_NO_TTINFO = _ttinfo(None, None, None)
+
+
+class _TZStr:
+    __slots__ = (
+        "std",
+        "dst",
+        "start",
+        "end",
+        "get_trans_info",
+        "get_trans_info_fromutc",
+        "dst_diff",
+    )
+
+    def __init__(
+        self, std_abbr, std_offset, dst_abbr, dst_offset, start=None, end=None
+    ):
+        self.dst_diff = dst_offset - std_offset
+        std_offset = _load_timedelta(std_offset)
+        self.std = _ttinfo(
+            utcoff=std_offset, dstoff=_load_timedelta(0), tzname=std_abbr
+        )
+
+        self.start = start
+        self.end = end
+
+        dst_offset = _load_timedelta(dst_offset)
+        delta = _load_timedelta(self.dst_diff)
+        self.dst = _ttinfo(utcoff=dst_offset, dstoff=delta, tzname=dst_abbr)
+
+        # These are assertions because the constructor should only be called
+        # by functions that would fail before passing start or end
+        assert start is not None, "No transition start specified"
+        assert end is not None, "No transition end specified"
+
+        self.get_trans_info = self._get_trans_info
+        self.get_trans_info_fromutc = self._get_trans_info_fromutc
+
+    def transitions(self, year):
+        start = self.start.year_to_epoch(year)
+        end = self.end.year_to_epoch(year)
+        return start, end
+
+    def _get_trans_info(self, ts, year, fold):
+        """Get the information about the current transition - tti"""
+        start, end = self.transitions(year)
+
+        # With fold = 0, the period (denominated in local time) with the
+        # smaller offset starts at the end of the gap and ends at the end of
+        # the fold; with fold = 1, it runs from the start of the gap to the
+        # beginning of the fold.
+        #
+        # So in order to determine the DST boundaries we need to know both
+        # the fold and whether DST is positive or negative (rare), and it
+        # turns out that this boils down to fold XOR is_positive.
+        if fold == (self.dst_diff >= 0):
+            end -= self.dst_diff
+        else:
+            start += self.dst_diff
+
+        if start < end:
+            isdst = start <= ts < end
+        else:
+            isdst = not (end <= ts < start)
+
+        return self.dst if isdst else self.std
+
+    def _get_trans_info_fromutc(self, ts, year):
+        start, end = self.transitions(year)
+        start -= self.std.utcoff.total_seconds()
+        end -= self.dst.utcoff.total_seconds()
+
+        if start < end:
+            isdst = start <= ts < end
+        else:
+            isdst = not (end <= ts < start)
+
+        # For positive DST, the ambiguous period is one dst_diff after the end
+        # of DST; for negative DST, the ambiguous period is one dst_diff before
+        # the start of DST.
+        if self.dst_diff > 0:
+            ambig_start = end
+            ambig_end = end + self.dst_diff
+        else:
+            ambig_start = start
+            ambig_end = start - self.dst_diff
+
+        fold = ambig_start <= ts < ambig_end
+
+        return (self.dst if isdst else self.std, fold)
+
+
+def _post_epoch_days_before_year(year):
+    """Get the number of days between 1970-01-01 and YEAR-01-01"""
+    y = year - 1
+    return y * 365 + y // 4 - y // 100 + y // 400 - EPOCHORDINAL
+
+
+class _DayOffset:
+    __slots__ = ["d", "julian", "hour", "minute", "second"]
+
+    def __init__(self, d, julian, hour=2, minute=0, second=0):
+        if not (0 + julian) <= d <= 365:
+            min_day = 0 + julian
+            raise ValueError(f"d must be in [{min_day}, 365], not: {d}")
+
+        self.d = d
+        self.julian = julian
+        self.hour = hour
+        self.minute = minute
+        self.second = second
+
+    def year_to_epoch(self, year):
+        days_before_year = _post_epoch_days_before_year(year)
+
+        d = self.d
+        if self.julian and d >= 59 and calendar.isleap(year):
+            d += 1
+
+        epoch = (days_before_year + d) * 86400
+        epoch += self.hour * 3600 + self.minute * 60 + self.second
+
+        return epoch
+
+
+class _CalendarOffset:
+    __slots__ = ["m", "w", "d", "hour", "minute", "second"]
+
+    _DAYS_BEFORE_MONTH = (
+        -1,
+        0,
+        31,
+        59,
+        90,
+        120,
+        151,
+        181,
+        212,
+        243,
+        273,
+        304,
+        334,
+    )
+
+    def __init__(self, m, w, d, hour=2, minute=0, second=0):
+        if not 0 < m <= 12:
+            raise ValueError("m must be in (0, 12]")
+
+        if not 0 < w <= 5:
+            raise ValueError("w must be in (0, 5]")
+
+        if not 0 <= d <= 6:
+            raise ValueError("d must be in [0, 6]")
+
+        self.m = m
+        self.w = w
+        self.d = d
+        self.hour = hour
+        self.minute = minute
+        self.second = second
+
+    @classmethod
+    def _ymd2ord(cls, year, month, day):
+        return (
+            _post_epoch_days_before_year(year)
+            + cls._DAYS_BEFORE_MONTH[month]
+            + (month > 2 and calendar.isleap(year))
+            + day
+        )
+
+    # TODO: These are not actually epoch dates as they are expressed in local time
+    def year_to_epoch(self, year):
+        """Calculates the datetime of the occurrence from the year"""
+        # We know year and month, we need to convert w, d into day of month
+        #
+        # Week 1 is the first week in which day `d` (where 0 = Sunday) appears.
+        # Week 5 represents the last occurrence of day `d`, so we need to know
+        # the range of the month.
+        first_day, days_in_month = calendar.monthrange(year, self.m)
+
+        # This equation seems magical, so I'll break it down:
+        # 1. calendar says 0 = Monday, POSIX says 0 = Sunday
+        #    so we need first_day + 1 to get 1 = Monday -> 7 = Sunday,
+        #    which is still equivalent because this math is mod 7
+        # 2. Get first day - desired day mod 7: -1 % 7 = 6, so we don't need
+        #    to do anything to adjust negative numbers.
+        # 3. Add 1 because month days are a 1-based index.
+        month_day = (self.d - (first_day + 1)) % 7 + 1
+
+        # Now use a 0-based index version of `w` to calculate the w-th
+        # occurrence of `d`
+        month_day += (self.w - 1) * 7
+
+        # month_day will only be > days_in_month if w was 5, and `w` means
+        # "last occurrence of `d`", so now we just check if we over-shot the
+        # end of the month and if so knock off 1 week.
+        if month_day > days_in_month:
+            month_day -= 7
+
+        ordinal = self._ymd2ord(year, self.m, month_day)
+        epoch = ordinal * 86400
+        epoch += self.hour * 3600 + self.minute * 60 + self.second
+        return epoch
+
+
+def _parse_tz_str(tz_str):
+    # The tz string has the format:
+    #
+    # std[offset[dst[offset],start[/time],end[/time]]]
+    #
+    # std and dst must be 3 or more characters long and must not contain
+    # a leading colon, embedded digits, commas, nor a plus or minus signs;
+    # The spaces between "std" and "offset" are only for display and are
+    # not actually present in the string.
+    #
+    # The format of the offset is ``[+|-]hh[:mm[:ss]]``
+
+    offset_str, *start_end_str = tz_str.split(",", 1)
+
+    # fmt: off
+    parser_re = re.compile(
+        r"(?P<std>[^<0-9:.+-]+|<[a-zA-Z0-9+\-]+>)" +
+        r"((?P<stdoff>[+-]?\d{1,2}(:\d{2}(:\d{2})?)?)" +
+            r"((?P<dst>[^0-9:.+-]+|<[a-zA-Z0-9+\-]+>)" +
+                r"((?P<dstoff>[+-]?\d{1,2}(:\d{2}(:\d{2})?)?))?" +
+            r")?" + # dst
+        r")?$" # stdoff
+    )
+    # fmt: on
+
+    m = parser_re.match(offset_str)
+
+    if m is None:
+        raise ValueError(f"{tz_str} is not a valid TZ string")
+
+    std_abbr = m.group("std")
+    dst_abbr = m.group("dst")
+    dst_offset = None
+
+    std_abbr = std_abbr.strip("<>")
+
+    if dst_abbr:
+        dst_abbr = dst_abbr.strip("<>")
+
+    if std_offset := m.group("stdoff"):
+        try:
+            std_offset = _parse_tz_delta(std_offset)
+        except ValueError as e:
+            raise ValueError(f"Invalid STD offset in {tz_str}") from e
+    else:
+        std_offset = 0
+
+    if dst_abbr is not None:
+        if dst_offset := m.group("dstoff"):
+            try:
+                dst_offset = _parse_tz_delta(dst_offset)
+            except ValueError as e:
+                raise ValueError(f"Invalid DST offset in {tz_str}") from e
+        else:
+            dst_offset = std_offset + 3600
+
+        if not start_end_str:
+            raise ValueError(f"Missing transition rules: {tz_str}")
+
+        start_end_strs = start_end_str[0].split(",", 1)
+        try:
+            start, end = (_parse_dst_start_end(x) for x in start_end_strs)
+        except ValueError as e:
+            raise ValueError(f"Invalid TZ string: {tz_str}") from e
+
+        return _TZStr(std_abbr, std_offset, dst_abbr, dst_offset, start, end)
+    elif start_end_str:
+        raise ValueError(f"Transition rule present without DST: {tz_str}")
+    else:
+        # This is a static ttinfo, don't return _TZStr
+        return _ttinfo(
+            _load_timedelta(std_offset), _load_timedelta(0), std_abbr
+        )
+
+
+def _parse_dst_start_end(dststr):
+    date, *time = dststr.split("/")
+    if date[0] == "M":
+        n_is_julian = False
+        m = re.match(r"M(\d{1,2})\.(\d).(\d)$", date)
+        if m is None:
+            raise ValueError(f"Invalid dst start/end date: {dststr}")
+        date_offset = tuple(map(int, m.groups()))
+        offset = _CalendarOffset(*date_offset)
+    else:
+        if date[0] == "J":
+            n_is_julian = True
+            date = date[1:]
+        else:
+            n_is_julian = False
+
+        doy = int(date)
+        offset = _DayOffset(doy, n_is_julian)
+
+    if time:
+        time_components = list(map(int, time[0].split(":")))
+        n_components = len(time_components)
+        if n_components < 3:
+            time_components.extend([0] * (3 - n_components))
+        offset.hour, offset.minute, offset.second = time_components
+
+    return offset
+
+
+def _parse_tz_delta(tz_delta):
+    match = re.match(
+        r"(?P<sign>[+-])?(?P<h>\d{1,2})(:(?P<m>\d{2})(:(?P<s>\d{2}))?)?",
+        tz_delta,
+    )
+    # Anything passed to this function should already have hit an equivalent
+    # regular expression to find the section to parse.
+    assert match is not None, tz_delta
+
+    h, m, s = (
+        int(v) if v is not None else 0
+        for v in map(match.group, ("h", "m", "s"))
+    )
+
+    total = h * 3600 + m * 60 + s
+
+    if not -86400 < total < 86400:
+        raise ValueError(
+            "Offset must be strictly between -24h and +24h:" + tz_delta
+        )
+
+    # Yes, +5 maps to an offset of -5h
+    if match.group("sign") != "-":
+        total *= -1
+
+    return total
diff --git a/Makefile.pre.in b/Makefile.pre.in
index d545a9efb3c..dbfd805f1a0 100644
--- a/Makefile.pre.in
+++ b/Makefile.pre.in
@@ -197,6 +197,9 @@ OPENSSL_INCLUDES=@OPENSSL_INCLUDES@
 OPENSSL_LIBS=@OPENSSL_LIBS@
 OPENSSL_LDFLAGS=@OPENSSL_LDFLAGS@
 
+# Default zoneinfo.TZPATH. Added here to expose it in sysconfig.get_config_var
+TZPATH=@TZPATH@
+
 # Modes for directories, executables and data files created by the
 # install process.  Default to user-only-writable for all file types.
 DIRMODE=	755
diff --git a/Modules/Setup b/Modules/Setup
index 87e73bac78f..02cfb67518d 100644
--- a/Modules/Setup
+++ b/Modules/Setup
@@ -181,6 +181,7 @@ _symtable symtablemodule.c
 #_elementtree -I$(srcdir)/Modules/expat -DHAVE_EXPAT_CONFIG_H -DUSE_PYEXPAT_CAPI _elementtree.c	# elementtree accelerator
 #_pickle _pickle.c	# pickle accelerator
 #_datetime _datetimemodule.c	# datetime accelerator
+#_zoneinfo _zoneinfo.c	# zoneinfo accelerator
 #_bisect _bisectmodule.c	# Bisection algorithms
 #_heapq _heapqmodule.c	# Heap queue algorithm
 #_asyncio _asynciomodule.c  # Fast asyncio Future
diff --git a/Modules/_zoneinfo.c b/Modules/_zoneinfo.c
new file mode 100644
index 00000000000..9f5e64d8486
--- /dev/null
+++ b/Modules/_zoneinfo.c
@@ -0,0 +1,2695 @@
+#include "Python.h"
+#include "structmember.h"
+
+#include <ctype.h>
+#include <stddef.h>
+#include <stdint.h>
+
+#include "datetime.h"
+
+// Imports
+static PyObject *io_open = NULL;
+static PyObject *_tzpath_find_tzfile = NULL;
+static PyObject *_common_mod = NULL;
+
+typedef struct TransitionRuleType TransitionRuleType;
+typedef struct StrongCacheNode StrongCacheNode;
+
+typedef struct {
+    PyObject *utcoff;
+    PyObject *dstoff;
+    PyObject *tzname;
+    long utcoff_seconds;
+} _ttinfo;
+
+typedef struct {
+    _ttinfo std;
+    _ttinfo dst;
+    int dst_diff;
+    TransitionRuleType *start;
+    TransitionRuleType *end;
+    unsigned char std_only;
+} _tzrule;
+
+typedef struct {
+    PyDateTime_TZInfo base;
+    PyObject *key;
+    PyObject *file_repr;
+    PyObject *weakreflist;
+    unsigned int num_transitions;
+    unsigned int num_ttinfos;
+    int64_t *trans_list_utc;
+    int64_t *trans_list_wall[2];
+    _ttinfo **trans_ttinfos;  // References to the ttinfo for each transition
+    _ttinfo *ttinfo_before;
+    _tzrule tzrule_after;
+    _ttinfo *_ttinfos;  // Unique array of ttinfos for ease of deallocation
+    unsigned char fixed_offset;
+    unsigned char source;
+} PyZoneInfo_ZoneInfo;
+
+struct TransitionRuleType {
+    int64_t (*year_to_timestamp)(TransitionRuleType *, int);
+};
+
+typedef struct {
+    TransitionRuleType base;
+    uint8_t month;
+    uint8_t week;
+    uint8_t day;
+    int8_t hour;
+    int8_t minute;
+    int8_t second;
+} CalendarRule;
+
+typedef struct {
+    TransitionRuleType base;
+    uint8_t julian;
+    unsigned int day;
+    int8_t hour;
+    int8_t minute;
+    int8_t second;
+} DayRule;
+
+struct StrongCacheNode {
+    StrongCacheNode *next;
+    StrongCacheNode *prev;
+    PyObject *key;
+    PyObject *zone;
+};
+
+static PyTypeObject PyZoneInfo_ZoneInfoType;
+
+// Globals
+static PyObject *TIMEDELTA_CACHE = NULL;
+static PyObject *ZONEINFO_WEAK_CACHE = NULL;
+static StrongCacheNode *ZONEINFO_STRONG_CACHE = NULL;
+static size_t ZONEINFO_STRONG_CACHE_MAX_SIZE = 8;
+
+static _ttinfo NO_TTINFO = {NULL, NULL, NULL, 0};
+
+// Constants
+static const int EPOCHORDINAL = 719163;
+static int DAYS_IN_MONTH[] = {
+    -1, 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31,
+};
+
+static int DAYS_BEFORE_MONTH[] = {
+    -1, 0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334,
+};
+
+static const int SOURCE_NOCACHE = 0;
+static const int SOURCE_CACHE = 1;
+static const int SOURCE_FILE = 2;
+
+// Forward declarations
+static int
+load_data(PyZoneInfo_ZoneInfo *self, PyObject *file_obj);
+static void
+utcoff_to_dstoff(size_t *trans_idx, long *utcoffs, long *dstoffs,
+                 unsigned char *isdsts, size_t num_transitions,
+                 size_t num_ttinfos);
+static int
+ts_to_local(size_t *trans_idx, int64_t *trans_utc, long *utcoff,
+            int64_t *trans_local[2], size_t num_ttinfos,
+            size_t num_transitions);
+
+static int
+parse_tz_str(PyObject *tz_str_obj, _tzrule *out);
+
+static ssize_t
+parse_abbr(const char *const p, PyObject **abbr);
+static ssize_t
+parse_tz_delta(const char *const p, long *total_seconds);
+static ssize_t
+parse_transition_time(const char *const p, int8_t *hour, int8_t *minute,
+                      int8_t *second);
+static ssize_t
+parse_transition_rule(const char *const p, TransitionRuleType **out);
+
+static _ttinfo *
+find_tzrule_ttinfo(_tzrule *rule, int64_t ts, unsigned char fold, int year);
+static _ttinfo *
+find_tzrule_ttinfo_fromutc(_tzrule *rule, int64_t ts, int year,
+                           unsigned char *fold);
+
+static int
+build_ttinfo(long utcoffset, long dstoffset, PyObject *tzname, _ttinfo *out);
+static void
+xdecref_ttinfo(_ttinfo *ttinfo);
+static int
+ttinfo_eq(const _ttinfo *const tti0, const _ttinfo *const tti1);
+
+static int
+build_tzrule(PyObject *std_abbr, PyObject *dst_abbr, long std_offset,
+             long dst_offset, TransitionRuleType *start,
+             TransitionRuleType *end, _tzrule *out);
+static void
+free_tzrule(_tzrule *tzrule);
+
+static PyObject *
+load_timedelta(long seconds);
+
+static int
+get_local_timestamp(PyObject *dt, int64_t *local_ts);
+static _ttinfo *
+find_ttinfo(PyZoneInfo_ZoneInfo *self, PyObject *dt);
+
+static int
+ymd_to_ord(int y, int m, int d);
+static int
+is_leap_year(int year);
+
+static size_t
+_bisect(const int64_t value, const int64_t *arr, size_t size);
+
+static void
+eject_from_strong_cache(const PyTypeObject *const type, PyObject *key);
+static void
+clear_strong_cache(const PyTypeObject *const type);
+static void
+update_strong_cache(const PyTypeObject *const type, PyObject *key,
+                    PyObject *zone);
+static PyObject *
+zone_from_strong_cache(const PyTypeObject *const type, PyObject *key);
+
+static PyObject *
+zoneinfo_new_instance(PyTypeObject *type, PyObject *key)
+{
+    PyObject *file_obj = NULL;
+    PyObject *file_path = NULL;
+
+    file_path = PyObject_CallFunctionObjArgs(_tzpath_find_tzfile, key, NULL);
+    if (file_path == NULL) {
+        return NULL;
+    }
+    else if (file_path == Py_None) {
+        file_obj = PyObject_CallMethod(_common_mod, "load_tzdata", "O", key);
+        if (file_obj == NULL) {
+            Py_DECREF(file_path);
+            return NULL;
+        }
+    }
+
+    PyObject *self = (PyObject *)(type->tp_alloc(type, 0));
+    if (self == NULL) {
+        goto error;
+    }
+
+    if (file_obj == NULL) {
+        file_obj = PyObject_CallFunction(io_open, "Os", file_path, "rb");
+        if (file_obj == NULL) {
+            goto error;
+        }
+    }
+
+    if (load_data((PyZoneInfo_ZoneInfo *)self, file_obj)) {
+        goto error;
+    }
+
+    PyObject *rv = PyObject_CallMethod(file_obj, "close", NULL);
+    Py_DECREF(file_obj);
+    file_obj = NULL;
+    if (rv == NULL) {
+        goto error;
+    }
+    Py_DECREF(rv);
+
+    ((PyZoneInfo_ZoneInfo *)self)->key = key;
+    Py_INCREF(key);
+
+    goto cleanup;
+error:
+    Py_XDECREF(self);
+    self = NULL;
+cleanup:
+    if (file_obj != NULL) {
+        PyObject *tmp = PyObject_CallMethod(file_obj, "close", NULL);
+        Py_DECREF(tmp);
+        Py_DECREF(file_obj);
+    }
+    Py_DECREF(file_path);
+    return self;
+}
+
+static PyObject *
+get_weak_cache(PyTypeObject *type)
+{
+    if (type == &PyZoneInfo_ZoneInfoType) {
+        return ZONEINFO_WEAK_CACHE;
+    }
+    else {
+        PyObject *cache =
+            PyObject_GetAttrString((PyObject *)type, "_weak_cache");
+        // We are assuming that the type lives at least as long as the function
+        // that calls get_weak_cache, and that it holds a reference to the
+        // cache, so we'll return a "borrowed reference".
+        Py_XDECREF(cache);
+        return cache;
+    }
+}
+
+static PyObject *
+zoneinfo_new(PyTypeObject *type, PyObject *args, PyObject *kw)
+{
+    PyObject *key = NULL;
+    static char *kwlist[] = {"key", NULL};
+    if (PyArg_ParseTupleAndKeywords(args, kw, "O", kwlist, &key) == 0) {
+        return NULL;
+    }
+
+    PyObject *instance = zone_from_strong_cache(type, key);
+    if (instance != NULL) {
+        return instance;
+    }
+
+    PyObject *weak_cache = get_weak_cache(type);
+    instance = PyObject_CallMethod(weak_cache, "get", "O", key, Py_None);
+    if (instance == NULL) {
+        return NULL;
+    }
+
+    if (instance == Py_None) {
+        Py_DECREF(instance);
+        PyObject *tmp = zoneinfo_new_instance(type, key);
+        if (tmp == NULL) {
+            return NULL;
+        }
+
+        instance =
+            PyObject_CallMethod(weak_cache, "setdefault", "OO", key, tmp);
+        ((PyZoneInfo_ZoneInfo *)instance)->source = SOURCE_CACHE;
+
+        Py_DECREF(tmp);
+
+        if (instance == NULL) {
+            return NULL;
+        }
+    }
+
+    update_strong_cache(type, key, instance);
+    return instance;
+}
+
+static void
+zoneinfo_dealloc(PyObject *obj_self)
+{
+    PyZoneInfo_ZoneInfo *self = (PyZoneInfo_ZoneInfo *)obj_self;
+
+    if (self->weakreflist != NULL) {
+        PyObject_ClearWeakRefs(obj_self);
+    }
+
+    if (self->trans_list_utc != NULL) {
+        PyMem_Free(self->trans_list_utc);
+    }
+
+    for (size_t i = 0; i < 2; i++) {
+        if (self->trans_list_wall[i] != NULL) {
+            PyMem_Free(self->trans_list_wall[i]);
+        }
+    }
+
+    if (self->_ttinfos != NULL) {
+        for (size_t i = 0; i < self->num_ttinfos; ++i) {
+            xdecref_ttinfo(&(self->_ttinfos[i]));
+        }
+        PyMem_Free(self->_ttinfos);
+    }
+
+    if (self->trans_ttinfos != NULL) {
+        PyMem_Free(self->trans_ttinfos);
+    }
+
+    free_tzrule(&(self->tzrule_after));
+
+    Py_XDECREF(self->key);
+    Py_XDECREF(self->file_repr);
+
+    Py_TYPE(self)->tp_free((PyObject *)self);
+}
+
+static PyObject *
+zoneinfo_from_file(PyTypeObject *type, PyObject *args, PyObject *kwargs)
+{
+    PyObject *file_obj = NULL;
+    PyObject *file_repr = NULL;
+    PyObject *key = Py_None;
+    PyZoneInfo_ZoneInfo *self = NULL;
+
+    static char *kwlist[] = {"", "key", NULL};
+    if (!PyArg_ParseTupleAndKeywords(args, kwargs, "O|O", kwlist, &file_obj,
+                                     &key)) {
+        return NULL;
+    }
+
+    PyObject *obj_self = (PyObject *)(type->tp_alloc(type, 0));
+    self = (PyZoneInfo_ZoneInfo *)obj_self;
+    if (self == NULL) {
+        return NULL;
+    }
+
+    file_repr = PyUnicode_FromFormat("%R", file_obj);
+    if (file_repr == NULL) {
+        goto error;
+    }
+
+    if (load_data(self, file_obj)) {
+        goto error;
+    }
+
+    self->source = SOURCE_FILE;
+    self->file_repr = file_repr;
+    self->key = key;
+    Py_INCREF(key);
+
+    return obj_self;
+error:
+    Py_XDECREF(file_repr);
+    Py_XDECREF(self);
+    return NULL;
+}
+
+static PyObject *
+zoneinfo_no_cache(PyTypeObject *cls, PyObject *args, PyObject *kwargs)
+{
+    static char *kwlist[] = {"key", NULL};
+    PyObject *key = NULL;
+    if (!PyArg_ParseTupleAndKeywords(args, kwargs, "O", kwlist, &key)) {
+        return NULL;
+    }
+
+    PyObject *out = zoneinfo_new_instance(cls, key);
+    if (out != NULL) {
+        ((PyZoneInfo_ZoneInfo *)out)->source = SOURCE_NOCACHE;
+    }
+
+    return out;
+}
+
+static PyObject *
+zoneinfo_clear_cache(PyObject *cls, PyObject *args, PyObject *kwargs)
+{
+    PyObject *only_keys = NULL;
+    static char *kwlist[] = {"only_keys", NULL};
+
+    if (!(PyArg_ParseTupleAndKeywords(args, kwargs, "|$O", kwlist,
+                                      &only_keys))) {
+        return NULL;
+    }
+
+    PyTypeObject *type = (PyTypeObject *)cls;
+    PyObject *weak_cache = get_weak_cache(type);
+
+    if (only_keys == NULL || only_keys == Py_None) {
+        PyObject *rv = PyObject_CallMethod(weak_cache, "clear", NULL);
+        if (rv != NULL) {
+            Py_DECREF(rv);
+        }
+
+        clear_strong_cache(type);
+        ZONEINFO_STRONG_CACHE = NULL;
+    }
+    else {
+        PyObject *item = NULL;
+        PyObject *pop = PyUnicode_FromString("pop");
+        if (pop == NULL) {
+            return NULL;
+        }
+
+        PyObject *iter = PyObject_GetIter(only_keys);
+        if (iter == NULL) {
+            Py_DECREF(pop);
+            return NULL;
+        }
+
+        while ((item = PyIter_Next(iter))) {
+            // Remove from strong cache
+            eject_from_strong_cache(type, item);
+
+            // Remove from weak cache
+            PyObject *tmp = PyObject_CallMethodObjArgs(weak_cache, pop, item,
+                                                       Py_None, NULL);
+
+            Py_DECREF(item);
+            if (tmp == NULL) {
+                break;
+            }
+            Py_DECREF(tmp);
+        }
+        Py_DECREF(iter);
+        Py_DECREF(pop);
+    }
+
+    if (PyErr_Occurred()) {
+        return NULL;
+    }
+
+    Py_RETURN_NONE;
+}
+
+static PyObject *
+zoneinfo_utcoffset(PyObject *self, PyObject *dt)
+{
+    _ttinfo *tti = find_ttinfo((PyZoneInfo_ZoneInfo *)self, dt);
+    if (tti == NULL) {
+        return NULL;
+    }
+    Py_INCREF(tti->utcoff);
+    return tti->utcoff;
+}
+
+static PyObject *
+zoneinfo_dst(PyObject *self, PyObject *dt)
+{
+    _ttinfo *tti = find_ttinfo((PyZoneInfo_ZoneInfo *)self, dt);
+    if (tti == NULL) {
+        return NULL;
+    }
+    Py_INCREF(tti->dstoff);
+    return tti->dstoff;
+}
+
+static PyObject *
+zoneinfo_tzname(PyObject *self, PyObject *dt)
+{
+    _ttinfo *tti = find_ttinfo((PyZoneInfo_ZoneInfo *)self, dt);
+    if (tti == NULL) {
+        return NULL;
+    }
+    Py_INCREF(tti->tzname);
+    return tti->tzname;
+}
+
+#define HASTZINFO(p) (((_PyDateTime_BaseTZInfo *)(p))->hastzinfo)
+#define GET_DT_TZINFO(p) \
+    (HASTZINFO(p) ? ((PyDateTime_DateTime *)(p))->tzinfo : Py_None)
+
+static PyObject *
+zoneinfo_fromutc(PyObject *obj_self, PyObject *dt)
+{
+    if (!PyDateTime_Check(dt)) {
+        PyErr_SetString(PyExc_TypeError,
+                        "fromutc: argument must be a datetime");
+        return NULL;
+    }
+    if (GET_DT_TZINFO(dt) != obj_self) {
+        PyErr_SetString(PyExc_ValueError,
+                        "fromutc: dt.tzinfo "
+                        "is not self");
+        return NULL;
+    }
+
+    PyZoneInfo_ZoneInfo *self = (PyZoneInfo_ZoneInfo *)obj_self;
+
+    int64_t timestamp;
+    if (get_local_timestamp(dt, &timestamp)) {
+        return NULL;
+    }
+    size_t num_trans = self->num_transitions;
+
+    _ttinfo *tti = NULL;
+    unsigned char fold = 0;
+
+    if (num_trans >= 1 && timestamp < self->trans_list_utc[0]) {
+        tti = self->ttinfo_before;
+    }
+    else if (num_trans == 0 ||
+             timestamp > self->trans_list_utc[num_trans - 1]) {
+        tti = find_tzrule_ttinfo_fromutc(&(self->tzrule_after), timestamp,
+                                         PyDateTime_GET_YEAR(dt), &fold);
+
+        // Immediately after the last manual transition, the fold/gap is
+        // between self->trans_ttinfos[num_transitions - 1] and whatever
+        // ttinfo applies immediately after the last transition, not between
+        // the STD and DST rules in the tzrule_after, so we may need to
+        // adjust the fold value.
+        if (num_trans) {
+            _ttinfo *tti_prev = NULL;
+            if (num_trans == 1) {
+                tti_prev = self->ttinfo_before;
+            }
+            else {
+                tti_prev = self->trans_ttinfos[num_trans - 2];
+            }
+            int64_t diff = tti_prev->utcoff_seconds - tti->utcoff_seconds;
+            if (diff > 0 &&
+                timestamp < (self->trans_list_utc[num_trans - 1] + diff)) {
+                fold = 1;
+            }
+        }
+    }
+    else {
+        size_t idx = _bisect(timestamp, self->trans_list_utc, num_trans);
+        _ttinfo *tti_prev = NULL;
+
+        if (idx >= 2) {
+            tti_prev = self->trans_ttinfos[idx - 2];
+            tti = self->trans_ttinfos[idx - 1];
+        }
+        else {
+            tti_prev = self->ttinfo_before;
+            tti = self->trans_ttinfos[0];
+        }
+
+        // Detect fold
+        int64_t shift =
+            (int64_t)(tti_prev->utcoff_seconds - tti->utcoff_seconds);
+        if (shift > (timestamp - self->trans_list_utc[idx - 1])) {
+            fold = 1;
+        }
+    }
+
+    PyObject *tmp = PyNumber_Add(dt, tti->utcoff);
+    if (tmp == NULL) {
+        return NULL;
+    }
+
+    if (fold) {
+        if (PyDateTime_CheckExact(tmp)) {
+            ((PyDateTime_DateTime *)tmp)->fold = 1;
+            dt = tmp;
+        }
+        else {
+            PyObject *replace = PyObject_GetAttrString(tmp, "replace");
+            PyObject *args = PyTuple_New(0);
+            PyObject *kwargs = PyDict_New();
+
+            Py_DECREF(tmp);
+            if (args == NULL || kwargs == NULL || replace == NULL) {
+                Py_XDECREF(args);
+                Py_XDECREF(kwargs);
+                Py_XDECREF(replace);
+                return NULL;
+            }
+
+            dt = NULL;
+            if (!PyDict_SetItemString(kwargs, "fold", _PyLong_One)) {
+                dt = PyObject_Call(replace, args, kwargs);
+            }
+
+            Py_DECREF(args);
+            Py_DECREF(kwargs);
+            Py_DECREF(replace);
+
+            if (dt == NULL) {
+                return NULL;
+            }
+        }
+    }
+    else {
+        dt = tmp;
+    }
+    return dt;
+}
+
+static PyObject *
+zoneinfo_repr(PyZoneInfo_ZoneInfo *self)
+{
+    PyObject *rv = NULL;
+    const char *type_name = Py_TYPE((PyObject *)self)->tp_name;
+    if (!(self->key == Py_None)) {
+        rv = PyUnicode_FromFormat("%s(key=%R)", type_name, self->key);
+    }
+    else {
+        assert(PyUnicode_Check(self->file_repr));
+        rv = PyUnicode_FromFormat("%s.from_file(%U)", type_name,
+                                  self->file_repr);
+    }
+
+    return rv;
+}
+
+static PyObject *
+zoneinfo_str(PyZoneInfo_ZoneInfo *self)
+{
+    if (!(self->key == Py_None)) {
+        Py_INCREF(self->key);
+        return self->key;
+    }
+    else {
+        return zoneinfo_repr(self);
+    }
+}
+
+/* Pickles the ZoneInfo object by key and source.
+ *
+ * ZoneInfo objects are pickled by reference to the TZif file that they came
+ * from, which means that the exact transitions may be different or the file
+ * may not un-pickle if the data has changed on disk in the interim.
+ *
+ * It is necessary to include a bit indicating whether or not the object
+ * was constructed from the cache, because from-cache objects will hit the
+ * unpickling process's cache, whereas no-cache objects will bypass it.
+ *
+ * Objects constructed from ZoneInfo.from_file cannot be pickled.
+ */
+static PyObject *
+zoneinfo_reduce(PyObject *obj_self, PyObject *unused)
+{
+    PyZoneInfo_ZoneInfo *self = (PyZoneInfo_ZoneInfo *)obj_self;
+    if (self->source == SOURCE_FILE) {
+        // Objects constructed from files cannot be pickled.
+        PyObject *pickle = PyImport_ImportModule("pickle");
+        if (pickle == NULL) {
+            return NULL;
+        }
+
+        PyObject *pickle_error =
+            PyObject_GetAttrString(pickle, "PicklingError");
+        Py_DECREF(pickle);
+        if (pickle_error == NULL) {
+            return NULL;
+        }
+
+        PyErr_Format(pickle_error,
+                     "Cannot pickle a ZoneInfo file from a file stream.");
+        Py_DECREF(pickle_error);
+        return NULL;
+    }
+
+    unsigned char from_cache = self->source == SOURCE_CACHE ? 1 : 0;
+    PyObject *constructor = PyObject_GetAttrString(obj_self, "_unpickle");
+
+    if (constructor == NULL) {
+        return NULL;
+    }
+
+    PyObject *rv = Py_BuildValue("O(OB)", constructor, self->key, from_cache);
+    Py_DECREF(constructor);
+    return rv;
+}
+
+static PyObject *
+zoneinfo__unpickle(PyTypeObject *cls, PyObject *args)
+{
+    PyObject *key;
+    unsigned char from_cache;
+    if (!PyArg_ParseTuple(args, "OB", &key, &from_cache)) {
+        return NULL;
+    }
+
+    if (from_cache) {
+        PyObject *val_args = Py_BuildValue("(O)", key);
+        if (val_args == NULL) {
+            return NULL;
+        }
+
+        PyObject *rv = zoneinfo_new(cls, val_args, NULL);
+
+        Py_DECREF(val_args);
+        return rv;
+    }
+    else {
+        return zoneinfo_new_instance(cls, key);
+    }
+}
+
+/* It is relatively expensive to construct new timedelta objects, and in most
+ * cases we're looking at a relatively small number of timedeltas, such as
+ * integer number of hours, etc. We will keep a cache so that we construct
+ * a minimal number of these.
+ *
+ * Possibly this should be replaced with an LRU cache so that it's not possible
+ * for the memory usage to explode from this, but in order for this to be a
+ * serious problem, one would need to deliberately craft a malicious time zone
+ * file with many distinct offsets. As of tzdb 2019c, loading every single zone
+ * fills the cache with ~450 timedeltas for a total size of ~12kB.
+ *
+ * This returns a new reference to the timedelta.
+ */
+static PyObject *
+load_timedelta(long seconds)
+{
+    PyObject *rv = NULL;
+    PyObject *pyoffset = PyLong_FromLong(seconds);
+    if (pyoffset == NULL) {
+        return NULL;
+    }
+    int contains = PyDict_Contains(TIMEDELTA_CACHE, pyoffset);
+    if (contains == -1) {
+        goto error;
+    }
+
+    if (!contains) {
+        PyObject *tmp = PyDateTimeAPI->Delta_FromDelta(
+            0, seconds, 0, 1, PyDateTimeAPI->DeltaType);
+
+        if (tmp == NULL) {
+            goto error;
+        }
+
+        rv = PyDict_SetDefault(TIMEDELTA_CACHE, pyoffset, tmp);
+        Py_DECREF(tmp);
+    }
+    else {
+        rv = PyDict_GetItem(TIMEDELTA_CACHE, pyoffset);
+    }
+
+    Py_DECREF(pyoffset);
+    Py_INCREF(rv);
+    return rv;
+error:
+    Py_DECREF(pyoffset);
+    return NULL;
+}
+
+/* Constructor for _ttinfo object - this starts by initializing the _ttinfo
+ * to { NULL, NULL, NULL }, so that Py_XDECREF will work on partially
+ * initialized _ttinfo objects.
+ */
+static int
+build_ttinfo(long utcoffset, long dstoffset, PyObject *tzname, _ttinfo *out)
+{
+    out->utcoff = NULL;
+    out->dstoff = NULL;
+    out->tzname = NULL;
+
+    out->utcoff_seconds = utcoffset;
+    out->utcoff = load_timedelta(utcoffset);
+    if (out->utcoff == NULL) {
+        return -1;
+    }
+
+    out->dstoff = load_timedelta(dstoffset);
+    if (out->dstoff == NULL) {
+        return -1;
+    }
+
+    out->tzname = tzname;
+    Py_INCREF(tzname);
+
+    return 0;
+}
+
+/* Decrease reference count on any non-NULL members of a _ttinfo  */
+static void
+xdecref_ttinfo(_ttinfo *ttinfo)
+{
+    if (ttinfo != NULL) {
+        Py_XDECREF(ttinfo->utcoff);
+        Py_XDECREF(ttinfo->dstoff);
+        Py_XDECREF(ttinfo->tzname);
+    }
+}
+
+/* Equality function for _ttinfo. */
+static int
+ttinfo_eq(const _ttinfo *const tti0, const _ttinfo *const tti1)
+{
+    int rv;
+    if ((rv = PyObject_RichCompareBool(tti0->utcoff, tti1->utcoff, Py_EQ)) <
+        1) {
+        goto end;
+    }
+
+    if ((rv = PyObject_RichCompareBool(tti0->dstoff, tti1->dstoff, Py_EQ)) <
+        1) {
+        goto end;
+    }
+
+    if ((rv = PyObject_RichCompareBool(tti0->tzname, tti1->tzname, Py_EQ)) <
+        1) {
+        goto end;
+    }
+end:
+    return rv;
+}
+
+/* Given a file-like object, this populates a ZoneInfo object
+ *
+ * The current version calls into a Python function to read the data from
+ * file into Python objects, and this translates those Python objects into
+ * C values and calculates derived values (e.g. dstoff) in C.
+ *
+ * This returns 0 on success and -1 on failure.
+ *
+ * The function will never return while `self` is partially initialized —
+ * the object only needs to be freed / deallocated if this succeeds.
+ */
+static int
+load_data(PyZoneInfo_ZoneInfo *self, PyObject *file_obj)
+{
+    PyObject *data_tuple = NULL;
+
+    long *utcoff = NULL;
+    long *dstoff = NULL;
+    size_t *trans_idx = NULL;
+    unsigned char *isdst = NULL;
+
+    self->trans_list_utc = NULL;
+    self->trans_list_wall[0] = NULL;
+    self->trans_list_wall[1] = NULL;
+    self->trans_ttinfos = NULL;
+    self->_ttinfos = NULL;
+    self->file_repr = NULL;
+
+    size_t ttinfos_allocated = 0;
+
+    data_tuple = PyObject_CallMethod(_common_mod, "load_data", "O", file_obj);
+
+    if (data_tuple == NULL) {
+        goto error;
+    }
+
+    if (!PyTuple_CheckExact(data_tuple)) {
+        PyErr_Format(PyExc_TypeError, "Invalid data result type: %r",
+                     data_tuple);
+        goto error;
+    }
+
+    // Unpack the data tuple
+    PyObject *trans_idx_list = PyTuple_GetItem(data_tuple, 0);
+    if (trans_idx_list == NULL) {
+        goto error;
+    }
+
+    PyObject *trans_utc = PyTuple_GetItem(data_tuple, 1);
+    if (trans_utc == NULL) {
+        goto error;
+    }
+
+    PyObject *utcoff_list = PyTuple_GetItem(data_tuple, 2);
+    if (utcoff_list == NULL) {
+        goto error;
+    }
+
+    PyObject *isdst_list = PyTuple_GetItem(data_tuple, 3);
+    if (isdst_list == NULL) {
+        goto error;
+    }
+
+    PyObject *abbr = PyTuple_GetItem(data_tuple, 4);
+    if (abbr == NULL) {
+        goto error;
+    }
+
+    PyObject *tz_str = PyTuple_GetItem(data_tuple, 5);
+    if (tz_str == NULL) {
+        goto error;
+    }
+
+    // Load the relevant sizes
+    Py_ssize_t num_transitions = PyTuple_Size(trans_utc);
+    if (num_transitions == -1) {
+        goto error;
+    }
+
+    Py_ssize_t num_ttinfos = PyTuple_Size(utcoff_list);
+    if (num_ttinfos == -1) {
+        goto error;
+    }
+
+    self->num_transitions = (size_t)num_transitions;
+    self->num_ttinfos = (size_t)num_ttinfos;
+
+    // Load the transition indices and list
+    self->trans_list_utc =
+        PyMem_Malloc(self->num_transitions * sizeof(int64_t));
+    trans_idx = PyMem_Malloc(self->num_transitions * sizeof(Py_ssize_t));
+
+    for (Py_ssize_t i = 0; i < self->num_transitions; ++i) {
+        PyObject *num = PyTuple_GetItem(trans_utc, i);
+        if (num == NULL) {
+            goto error;
+        }
+        self->trans_list_utc[i] = PyLong_AsLongLong(num);
+        if (self->trans_list_utc[i] == -1 && PyErr_Occurred()) {
+            goto error;
+        }
+
+        num = PyTuple_GetItem(trans_idx_list, i);
+        if (num == NULL) {
+            goto error;
+        }
+
+        Py_ssize_t cur_trans_idx = PyLong_AsSsize_t(num);
+        if (cur_trans_idx == -1) {
+            goto error;
+        }
+
+        trans_idx[i] = (size_t)cur_trans_idx;
+        if (trans_idx[i] > self->num_ttinfos) {
+            PyErr_Format(
+                PyExc_ValueError,
+                "Invalid transition index found while reading TZif: %zd",
+                cur_trans_idx);
+
+            goto error;
+        }
+    }
+
+    // Load UTC offsets and isdst (size num_ttinfos)
+    utcoff = PyMem_Malloc(self->num_ttinfos * sizeof(long));
+    isdst = PyMem_Malloc(self->num_ttinfos * sizeof(unsigned char));
+
+    if (utcoff == NULL || isdst == NULL) {
+        goto error;
+    }
+    for (Py_ssize_t i = 0; i < self->num_ttinfos; ++i) {
+        PyObject *num = PyTuple_GetItem(utcoff_list, i);
+        if (num == NULL) {
+            goto error;
+        }
+
+        utcoff[i] = PyLong_AsLong(num);
+        if (utcoff[i] == -1 && PyErr_Occurred()) {
+            goto error;
+        }
+
+        num = PyTuple_GetItem(isdst_list, i);
+        if (num == NULL) {
+            goto error;
+        }
+
+        int isdst_with_error = PyObject_IsTrue(num);
+        if (isdst_with_error == -1) {
+            goto error;
+        }
+        else {
+            isdst[i] = (unsigned char)isdst_with_error;
+        }
+    }
+
+    dstoff = PyMem_Calloc(self->num_ttinfos, sizeof(long));
+    if (dstoff == NULL) {
+        goto error;
+    }
+
+    // Derive dstoff and trans_list_wall from the information we've loaded
+    utcoff_to_dstoff(trans_idx, utcoff, dstoff, isdst, self->num_transitions,
+                     self->num_ttinfos);
+
+    if (ts_to_local(trans_idx, self->trans_list_utc, utcoff,
+                    self->trans_list_wall, self->num_ttinfos,
+                    self->num_transitions)) {
+        goto error;
+    }
+
+    // Build _ttinfo objects from utcoff, dstoff and abbr
+    self->_ttinfos = PyMem_Malloc(self->num_ttinfos * sizeof(_ttinfo));
+    for (size_t i = 0; i < self->num_ttinfos; ++i) {
+        PyObject *tzname = PyTuple_GetItem(abbr, i);
+        if (tzname == NULL) {
+            goto error;
+        }
+
+        ttinfos_allocated++;
+        if (build_ttinfo(utcoff[i], dstoff[i], tzname, &(self->_ttinfos[i]))) {
+            goto error;
+        }
+    }
+
+    // Build our mapping from transition to the ttinfo that applies
+    self->trans_ttinfos =
+        PyMem_Calloc(self->num_transitions, sizeof(_ttinfo *));
+    for (size_t i = 0; i < self->num_transitions; ++i) {
+        size_t ttinfo_idx = trans_idx[i];
+        assert(ttinfo_idx < self->num_ttinfos);
+        self->trans_ttinfos[i] = &(self->_ttinfos[ttinfo_idx]);
+    }
+
+    // Set ttinfo_before to the first non-DST transition
+    for (size_t i = 0; i < self->num_ttinfos; ++i) {
+        if (!isdst[i]) {
+            self->ttinfo_before = &(self->_ttinfos[i]);
+            break;
+        }
+    }
+
+    // If there are only DST ttinfos, pick the first one, if there are no
+    // ttinfos at all, set ttinfo_before to NULL
+    if (self->ttinfo_before == NULL && self->num_ttinfos > 0) {
+        self->ttinfo_before = &(self->_ttinfos[0]);
+    }
+
+    if (tz_str != Py_None && PyObject_IsTrue(tz_str)) {
+        if (parse_tz_str(tz_str, &(self->tzrule_after))) {
+            goto error;
+        }
+    }
+    else {
+        if (!self->num_ttinfos) {
+            PyErr_Format(PyExc_ValueError, "No time zone information found.");
+            goto error;
+        }
+
+        size_t idx;
+        if (!self->num_transitions) {
+            idx = self->num_ttinfos - 1;
+        }
+        else {
+            idx = trans_idx[self->num_transitions - 1];
+        }
+
+        _ttinfo *tti = &(self->_ttinfos[idx]);
+        build_tzrule(tti->tzname, NULL, tti->utcoff_seconds, 0, NULL, NULL,
+                     &(self->tzrule_after));
+
+        // We've abused the build_tzrule constructor to construct an STD-only
+        // rule mimicking whatever ttinfo we've picked up, but it's possible
+        // that the one we've picked up is a DST zone, so we need to make sure
+        // that the dstoff is set correctly in that case.
+        if (PyObject_IsTrue(tti->dstoff)) {
+            _ttinfo *tti_after = &(self->tzrule_after.std);
+            Py_DECREF(tti_after->dstoff);
+            tti_after->dstoff = tti->dstoff;
+            Py_INCREF(tti_after->dstoff);
+        }
+    }
+
+    // Determine if this is a "fixed offset" zone, meaning that the output of
+    // the utcoffset, dst and tzname functions does not depend on the specific
+    // datetime passed.
+    //
+    // We make three simplifying assumptions here:
+    //
+    // 1. If tzrule_after is not std_only, it has transitions that might occur
+    //    (it is possible to construct TZ strings that specify STD and DST but
+    //    no transitions ever occur, such as AAA0BBB,0/0,J365/25).
+    // 2. If self->_ttinfos contains more than one _ttinfo object, the objects
+    //    represent different offsets.
+    // 3. self->ttinfos contains no unused _ttinfos (in which case an otherwise
+    //    fixed-offset zone with extra _ttinfos defined may appear to *not* be
+    //    a fixed offset zone).
+    //
+    // Violations to these assumptions would be fairly exotic, and exotic
+    // zones should almost certainly not be used with datetime.time (the
+    // only thing that would be affected by this).
+    if (self->num_ttinfos > 1 || !self->tzrule_after.std_only) {
+        self->fixed_offset = 0;
+    }
+    else if (self->num_ttinfos == 0) {
+        self->fixed_offset = 1;
+    }
+    else {
+        int constant_offset =
+            ttinfo_eq(&(self->_ttinfos[0]), &self->tzrule_after.std);
+        if (constant_offset < 0) {
+            goto error;
+        }
+        else {
+            self->fixed_offset = constant_offset;
+        }
+    }
+
+    int rv = 0;
+    goto cleanup;
+error:
+    // These resources only need to be freed if we have failed, if we succeed
+    // in initializing a PyZoneInfo_ZoneInfo object, we can rely on its dealloc
+    // method to free the relevant resources.
+    if (self->trans_list_utc != NULL) {
+        PyMem_Free(self->trans_list_utc);
+        self->trans_list_utc = NULL;
+    }
+
+    for (size_t i = 0; i < 2; ++i) {
+        if (self->trans_list_wall[i] != NULL) {
+            PyMem_Free(self->trans_list_wall[i]);
+            self->trans_list_wall[i] = NULL;
+        }
+    }
+
+    if (self->_ttinfos != NULL) {
+        for (size_t i = 0; i < ttinfos_allocated; ++i) {
+            xdecref_ttinfo(&(self->_ttinfos[i]));
+        }
+        PyMem_Free(self->_ttinfos);
+        self->_ttinfos = NULL;
+    }
+
+    if (self->trans_ttinfos != NULL) {
+        PyMem_Free(self->trans_ttinfos);
+        self->trans_ttinfos = NULL;
+    }
+
+    rv = -1;
+cleanup:
+    Py_XDECREF(data_tuple);
+
+    if (utcoff != NULL) {
+        PyMem_Free(utcoff);
+    }
+
+    if (dstoff != NULL) {
+        PyMem_Free(dstoff);
+    }
+
+    if (isdst != NULL) {
+        PyMem_Free(isdst);
+    }
+
+    if (trans_idx != NULL) {
+        PyMem_Free(trans_idx);
+    }
+
+    return rv;
+}
+
+/* Function to calculate the local timestamp of a transition from the year. */
+int64_t
+calendarrule_year_to_timestamp(TransitionRuleType *base_self, int year)
+{
+    CalendarRule *self = (CalendarRule *)base_self;
+
+    // We want (year, month, day of month); we have year and month, but we
+    // need to turn (week, day-of-week) into day-of-month
+    //
+    // Week 1 is the first week in which day `day` (where 0 = Sunday) appears.
+    // Week 5 represents the last occurrence of day `day`, so we need to know
+    // the first weekday of the month and the number of days in the month.
+    int8_t first_day = (ymd_to_ord(year, self->month, 1) + 6) % 7;
+    uint8_t days_in_month = DAYS_IN_MONTH[self->month];
+    if (self->month == 2 && is_leap_year(year)) {
+        days_in_month += 1;
+    }
+
+    // This equation seems magical, so I'll break it down:
+    // 1. calendar says 0 = Monday, POSIX says 0 = Sunday so we need first_day
+    //    + 1 to get 1 = Monday -> 7 = Sunday, which is still equivalent
+    //    because this math is mod 7
+    // 2. Get first day - desired day mod 7 (adjusting by 7 for negative
+    //    numbers so that -1 % 7 = 6).
+    // 3. Add 1 because month days are a 1-based index.
+    int8_t month_day = ((int8_t)(self->day) - (first_day + 1)) % 7;
+    if (month_day < 0) {
+        month_day += 7;
+    }
+    month_day += 1;
+
+    // Now use a 0-based index version of `week` to calculate the w-th
+    // occurrence of `day`
+    month_day += ((int8_t)(self->week) - 1) * 7;
+
+    // month_day will only be > days_in_month if w was 5, and `w` means "last
+    // occurrence of `d`", so now we just check if we over-shot the end of the
+    // month and if so knock off 1 week.
+    if (month_day > days_in_month) {
+        month_day -= 7;
+    }
+
+    int64_t ordinal = ymd_to_ord(year, self->month, month_day) - EPOCHORDINAL;
+    return ((ordinal * 86400) + (int64_t)(self->hour * 3600) +
+            (int64_t)(self->minute * 60) + (int64_t)(self->second));
+}
+
+/* Constructor for CalendarRule. */
+int
+calendarrule_new(uint8_t month, uint8_t week, uint8_t day, int8_t hour,
+                 int8_t minute, int8_t second, CalendarRule *out)
+{
+    // These bounds come from the POSIX standard, which describes an Mm.n.d
+    // rule as:
+    //
+    //   The d'th day (0 <= d <= 6) of week n of month m of the year (1 <= n <=
+    //   5, 1 <= m <= 12, where week 5 means "the last d day in month m" which
+    //   may occur in either the fourth or the fifth week). Week 1 is the first
+    //   week in which the d'th day occurs. Day zero is Sunday.
+    if (month <= 0 || month > 12) {
+        PyErr_Format(PyExc_ValueError, "Month must be in (0, 12]");
+        return -1;
+    }
+
+    if (week <= 0 || week > 5) {
+        PyErr_Format(PyExc_ValueError, "Week must be in (0, 5]");
+        return -1;
+    }
+
+    // day is an unsigned integer, so day < 0 should always return false, but
+    // if day's type changes to a signed integer *without* changing this value,
+    // it may create a bug. Considering that the compiler should be able to
+    // optimize out the first comparison if day is an unsigned integer anyway,
+    // we will leave this comparison in place and disable the compiler warning.
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wtype-limits"
+    if (day < 0 || day > 6) {
+#pragma GCC diagnostic pop
+        PyErr_Format(PyExc_ValueError, "Day must be in [0, 6]");
+        return -1;
+    }
+
+    TransitionRuleType base = {&calendarrule_year_to_timestamp};
+
+    CalendarRule new_offset = {
+        .base = base,
+        .month = month,
+        .week = week,
+        .day = day,
+        .hour = hour,
+        .minute = minute,
+        .second = second,
+    };
+
+    *out = new_offset;
+    return 0;
+}
+
+/* Function to calculate the local timestamp of a transition from the year.
+ *
+ * This translates the day of the year into a local timestamp — either a
+ * 1-based Julian day, not including leap days, or the 0-based year-day,
+ * including leap days.
+ * */
+int64_t
+dayrule_year_to_timestamp(TransitionRuleType *base_self, int year)
+{
+    // The function signature requires a TransitionRuleType pointer, but this
+    // function is only applicable to DayRule* objects.
+    DayRule *self = (DayRule *)base_self;
+
+    // ymd_to_ord calculates the number of days since 0001-01-01, but we want
+    // to know the number of days since 1970-01-01, so we must subtract off
+    // the equivalent of ymd_to_ord(1970, 1, 1).
+    //
+    // We subtract off an additional 1 day to account for January 1st (we want
+    // the number of full days *before* the date of the transition - partial
+    // days are accounted for in the hour, minute and second portions.
+    int64_t days_before_year = ymd_to_ord(year, 1, 1) - EPOCHORDINAL - 1;
+
+    // The Julian day specification skips over February 29th in leap years,
+    // from the POSIX standard:
+    //
+    //   Leap days shall not be counted. That is, in all years-including leap
+    //   years-February 28 is day 59 and March 1 is day 60. It is impossible to
+    //   refer explicitly to the occasional February 29.
+    //
+    // This is actually more useful than you'd think — if you want a rule that
+    // always transitions on a given calendar day (other than February 29th),
+    // you would use a Julian day, e.g. J91 always refers to April 1st and J365
+    // always refers to December 31st.
+    unsigned int day = self->day;
+    if (self->julian && day >= 59 && is_leap_year(year)) {
+        day += 1;
+    }
+
+    return ((days_before_year + day) * 86400) + (self->hour * 3600) +
+           (self->minute * 60) + self->second;
+}
+
+/* Constructor for DayRule. */
+static int
+dayrule_new(uint8_t julian, unsigned int day, int8_t hour, int8_t minute,
+            int8_t second, DayRule *out)
+{
+    // The POSIX standard specifies that Julian days must be in the range (1 <=
+    // n <= 365) and that non-Julian (they call it "0-based Julian") days must
+    // be in the range (0 <= n <= 365).
+    if (day < julian || day > 365) {
+        PyErr_Format(PyExc_ValueError, "day must be in [%u, 365], not: %u",
+                     julian, day);
+        return -1;
+    }
+
+    TransitionRuleType base = {
+        &dayrule_year_to_timestamp,
+    };
+
+    DayRule tmp = {
+        .base = base,
+        .julian = julian,
+        .day = day,
+        .hour = hour,
+        .minute = minute,
+        .second = second,
+    };
+
+    *out = tmp;
+
+    return 0;
+}
+
+/* Calculate the start and end rules for a _tzrule in the given year. */
+static void
+tzrule_transitions(_tzrule *rule, int year, int64_t *start, int64_t *end)
+{
+    assert(rule->start != NULL);
+    assert(rule->end != NULL);
+    *start = rule->start->year_to_timestamp(rule->start, year);
+    *end = rule->end->year_to_timestamp(rule->end, year);
+}
+
+/* Calculate the _ttinfo that applies at a given local time from a _tzrule.
+ *
+ * This takes a local timestamp and fold for disambiguation purposes; the year
+ * could technically be calculated from the timestamp, but given that the
+ * callers of this function already have the year information accessible from
+ * the datetime struct, it is taken as an additional parameter to reduce
+ * unncessary calculation.
+ * */
+static _ttinfo *
+find_tzrule_ttinfo(_tzrule *rule, int64_t ts, unsigned char fold, int year)
+{
+    if (rule->std_only) {
+        return &(rule->std);
+    }
+
+    int64_t start, end;
+    uint8_t isdst;
+
+    tzrule_transitions(rule, year, &start, &end);
+
+    // With fold = 0, the period (denominated in local time) with the smaller
+    // offset starts at the end of the gap and ends at the end of the fold;
+    // with fold = 1, it runs from the start of the gap to the beginning of the
+    // fold.
+    //
+    // So in order to determine the DST boundaries we need to know both the
+    // fold and whether DST is positive or negative (rare), and it turns out
+    // that this boils down to fold XOR is_positive.
+    if (fold == (rule->dst_diff >= 0)) {
+        end -= rule->dst_diff;
+    }
+    else {
+        start += rule->dst_diff;
+    }
+
+    if (start < end) {
+        isdst = (ts >= start) && (ts < end);
+    }
+    else {
+        isdst = (ts < end) || (ts >= start);
+    }
+
+    if (isdst) {
+        return &(rule->dst);
+    }
+    else {
+        return &(rule->std);
+    }
+}
+
+/* Calculate the ttinfo and fold that applies for a _tzrule at an epoch time.
+ *
+ * This function can determine the _ttinfo that applies at a given epoch time,
+ * (analogous to trans_list_utc), and whether or not the datetime is in a fold.
+ * This is to be used in the .fromutc() function.
+ *
+ * The year is technically a redundant parameter, because it can be calculated
+ * from the timestamp, but all callers of this function should have the year
+ * in the datetime struct anyway, so taking it as a parameter saves unnecessary
+ * calculation.
+ **/
+static _ttinfo *
+find_tzrule_ttinfo_fromutc(_tzrule *rule, int64_t ts, int year,
+                           unsigned char *fold)
+{
+    if (rule->std_only) {
+        *fold = 0;
+        return &(rule->std);
+    }
+
+    int64_t start, end;
+    uint8_t isdst;
+    tzrule_transitions(rule, year, &start, &end);
+    start -= rule->std.utcoff_seconds;
+    end -= rule->dst.utcoff_seconds;
+
+    if (start < end) {
+        isdst = (ts >= start) && (ts < end);
+    }
+    else {
+        isdst = (ts < end) || (ts >= start);
+    }
+
+    // For positive DST, the ambiguous period is one dst_diff after the end of
+    // DST; for negative DST, the ambiguous period is one dst_diff before the
+    // start of DST.
+    int64_t ambig_start, ambig_end;
+    if (rule->dst_diff > 0) {
+        ambig_start = end;
+        ambig_end = end + rule->dst_diff;
+    }
+    else {
+        ambig_start = start;
+        ambig_end = start - rule->dst_diff;
+    }
+
+    *fold = (ts >= ambig_start) && (ts < ambig_end);
+
+    if (isdst) {
+        return &(rule->dst);
+    }
+    else {
+        return &(rule->std);
+    }
+}
+
+/* Parse a TZ string in the format specified by the POSIX standard:
+ *
+ *  std offset[dst[offset],start[/time],end[/time]]
+ *
+ *  std and dst must be 3 or more characters long and must not contain a
+ *  leading colon, embedded digits, commas, nor a plus or minus signs; The
+ *  spaces between "std" and "offset" are only for display and are not actually
+ *  present in the string.
+ *
+ *  The format of the offset is ``[+|-]hh[:mm[:ss]]``
+ *
+ * See the POSIX.1 spec: IEE Std 1003.1-2018 §8.3:
+ *
+ * https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap08.html
+ */
+static int
+parse_tz_str(PyObject *tz_str_obj, _tzrule *out)
+{
+    PyObject *std_abbr = NULL;
+    PyObject *dst_abbr = NULL;
+    TransitionRuleType *start = NULL;
+    TransitionRuleType *end = NULL;
+    long std_offset, dst_offset;
+
+    char *tz_str = PyBytes_AsString(tz_str_obj);
+    if (tz_str == NULL) {
+        return -1;
+    }
+    char *p = tz_str;
+
+    // Read the `std` abbreviation, which must be at least 3 characters long.
+    ssize_t num_chars = parse_abbr(p, &std_abbr);
+    if (num_chars < 1) {
+        PyErr_Format(PyExc_ValueError, "Invalid STD format in %R", tz_str_obj);
+        goto error;
+    }
+
+    p += num_chars;
+
+    // Now read the STD offset, which is required
+    num_chars = parse_tz_delta(p, &std_offset);
+    if (num_chars < 0) {
+        PyErr_Format(PyExc_ValueError, "Invalid STD offset in %R", tz_str_obj);
+        goto error;
+    }
+    p += num_chars;
+
+    // If the string ends here, there is no DST, otherwise we must parse the
+    // DST abbreviation and start and end dates and times.
+    if (*p == '\0') {
+        goto complete;
+    }
+
+    num_chars = parse_abbr(p, &dst_abbr);
+    if (num_chars < 1) {
+        PyErr_Format(PyExc_ValueError, "Invalid DST format in %R", tz_str_obj);
+        goto error;
+    }
+    p += num_chars;
+
+    if (*p == ',') {
+        // From the POSIX standard:
+        //
+        // If no offset follows dst, the alternative time is assumed to be one
+        // hour ahead of standard time.
+        dst_offset = std_offset + 3600;
+    }
+    else {
+        num_chars = parse_tz_delta(p, &dst_offset);
+        if (num_chars < 0) {
+            PyErr_Format(PyExc_ValueError, "Invalid DST offset in %R",
+                         tz_str_obj);
+            goto error;
+        }
+
+        p += num_chars;
+    }
+
+    TransitionRuleType **transitions[2] = {&start, &end};
+    for (size_t i = 0; i < 2; ++i) {
+        if (*p != ',') {
+            PyErr_Format(PyExc_ValueError,
+                         "Missing transition rules in TZ string: %R",
+                         tz_str_obj);
+            goto error;
+        }
+        p++;
+
+        num_chars = parse_transition_rule(p, transitions[i]);
+        if (num_chars < 0) {
+            PyErr_Format(PyExc_ValueError,
+                         "Malformed transition rule in TZ string: %R",
+                         tz_str_obj);
+            goto error;
+        }
+        p += num_chars;
+    }
+
+    if (*p != '\0') {
+        PyErr_Format(PyExc_ValueError,
+                     "Extraneous characters at end of TZ string: %R",
+                     tz_str_obj);
+        goto error;
+    }
+
+complete:
+    build_tzrule(std_abbr, dst_abbr, std_offset, dst_offset, start, end, out);
+    Py_DECREF(std_abbr);
+    Py_XDECREF(dst_abbr);
+
+    return 0;
+error:
+    Py_XDECREF(std_abbr);
+    if (dst_abbr != NULL && dst_abbr != Py_None) {
+        Py_DECREF(dst_abbr);
+    }
+
+    if (start != NULL) {
+        PyMem_Free(start);
+    }
+
+    if (end != NULL) {
+        PyMem_Free(end);
+    }
+
+    return -1;
+}
+
+static ssize_t
+parse_uint(const char *const p)
+{
+    if (!isdigit(*p)) {
+        return -1;
+    }
+
+    return (*p) - '0';
+}
+
+/* Parse the STD and DST abbreviations from a TZ string. */
+static ssize_t
+parse_abbr(const char *const p, PyObject **abbr)
+{
+    const char *ptr = p;
+    char buff = *ptr;
+    const char *str_start;
+    const char *str_end;
+
+    if (*ptr == '<') {
+        ptr++;
+        str_start = ptr;
+        while ((buff = *ptr) != '>') {
+            // From the POSIX standard:
+            //
+            //   In the quoted form, the first character shall be the less-than
+            //   ( '<' ) character and the last character shall be the
+            //   greater-than ( '>' ) character. All characters between these
+            //   quoting characters shall be alphanumeric characters from the
+            //   portable character set in the current locale, the plus-sign (
+            //   '+' ) character, or the minus-sign ( '-' ) character. The std
+            //   and dst fields in this case shall not include the quoting
+            //   characters.
+            if (!isalpha(buff) && !isdigit(buff) && buff != '+' &&
+                buff != '-') {
+                return -1;
+            }
+            ptr++;
+        }
+        str_end = ptr;
+        ptr++;
+    }
+    else {
+        str_start = p;
+        // From the POSIX standard:
+        //
+        //   In the unquoted form, all characters in these fields shall be
+        //   alphabetic characters from the portable character set in the
+        //   current locale.
+        while (isalpha(*ptr)) {
+            ptr++;
+        }
+        str_end = ptr;
+    }
+
+    *abbr = PyUnicode_FromStringAndSize(str_start, str_end - str_start);
+    if (abbr == NULL) {
+        return -1;
+    }
+
+    return ptr - p;
+}
+
+/* Parse a UTC offset from a TZ str. */
+static ssize_t
+parse_tz_delta(const char *const p, long *total_seconds)
+{
+    // From the POSIX spec:
+    //
+    //   Indicates the value added to the local time to arrive at Coordinated
+    //   Universal Time. The offset has the form:
+    //
+    //   hh[:mm[:ss]]
+    //
+    //   One or more digits may be used; the value is always interpreted as a
+    //   decimal number.
+    //
+    // The POSIX spec says that the values for `hour` must be between 0 and 24
+    // hours, but RFC 8536 §3.3.1 specifies that the hours part of the
+    // transition times may be signed and range from -167 to 167.
+    long sign = -1;
+    long hours = 0;
+    long minutes = 0;
+    long seconds = 0;
+
+    const char *ptr = p;
+    char buff = *ptr;
+    if (buff == '-' || buff == '+') {
+        // Negative numbers correspond to *positive* offsets, from the spec:
+        //
+        //   If preceded by a '-', the timezone shall be east of the Prime
+        //   Meridian; otherwise, it shall be west (which may be indicated by
+        //   an optional preceding '+' ).
+        if (buff == '-') {
+            sign = 1;
+        }
+
+        ptr++;
+    }
+
+    // The hour can be 1 or 2 numeric characters
+    for (size_t i = 0; i < 2; ++i) {
+        buff = *ptr;
+        if (!isdigit(buff)) {
+            if (i == 0) {
+                return -1;
+            }
+            else {
+                break;
+            }
+        }
+
+        hours *= 10;
+        hours += buff - '0';
+        ptr++;
+    }
+
+    if (hours > 24 || hours < 0) {
+        return -1;
+    }
+
+    // Minutes and seconds always of the format ":dd"
+    long *outputs[2] = {&minutes, &seconds};
+    for (size_t i = 0; i < 2; ++i) {
+        if (*ptr != ':') {
+            goto complete;
+        }
+        ptr++;
+
+        for (size_t j = 0; j < 2; ++j) {
+            buff = *ptr;
+            if (!isdigit(buff)) {
+                return -1;
+            }
+            *(outputs[i]) *= 10;
+            *(outputs[i]) += buff - '0';
+            ptr++;
+        }
+    }
+
+complete:
+    *total_seconds = sign * ((hours * 3600) + (minutes * 60) + seconds);
+
+    return ptr - p;
+}
+
+/* Parse the date portion of a transition rule. */
+static ssize_t
+parse_transition_rule(const char *const p, TransitionRuleType **out)
+{
+    // The full transition rule indicates when to change back and forth between
+    // STD and DST, and has the form:
+    //
+    //   date[/time],date[/time]
+    //
+    // This function parses an individual date[/time] section, and returns
+    // the number of characters that contributed to the transition rule. This
+    // does not include the ',' at the end of the first rule.
+    //
+    // The POSIX spec states that if *time* is not given, the default is 02:00.
+    const char *ptr = p;
+    int8_t hour = 2;
+    int8_t minute = 0;
+    int8_t second = 0;
+
+    // Rules come in one of three flavors:
+    //
+    //   1. Jn: Julian day n, with no leap days.
+    //   2. n: Day of year (0-based, with leap days)
+    //   3. Mm.n.d: Specifying by month, week and day-of-week.
+
+    if (*ptr == 'M') {
+        uint8_t month, week, day;
+        ptr++;
+        ssize_t tmp = parse_uint(ptr);
+        if (tmp < 0) {
+            return -1;
+        }
+        month = (uint8_t)tmp;
+        ptr++;
+        if (*ptr != '.') {
+            tmp = parse_uint(ptr);
+            if (tmp < 0) {
+                return -1;
+            }
+
+            month *= 10;
+            month += (uint8_t)tmp;
+            ptr++;
+        }
+
+        uint8_t *values[2] = {&week, &day};
+        for (size_t i = 0; i < 2; ++i) {
+            if (*ptr != '.') {
+                return -1;
+            }
+            ptr++;
+
+            tmp = parse_uint(ptr);
+            if (tmp < 0) {
+                return -1;
+            }
+            ptr++;
+
+            *(values[i]) = tmp;
+        }
+
+        if (*ptr == '/') {
+            ptr++;
+            ssize_t num_chars =
+                parse_transition_time(ptr, &hour, &minute, &second);
+            if (num_chars < 0) {
+                return -1;
+            }
+            ptr += num_chars;
+        }
+
+        CalendarRule *rv = PyMem_Calloc(1, sizeof(CalendarRule));
+        if (rv == NULL) {
+            return -1;
+        }
+
+        if (calendarrule_new(month, week, day, hour, minute, second, rv)) {
+            PyMem_Free(rv);
+            return -1;
+        }
+
+        *out = (TransitionRuleType *)rv;
+    }
+    else {
+        uint8_t julian = 0;
+        unsigned int day = 0;
+        if (*ptr == 'J') {
+            julian = 1;
+            ptr++;
+        }
+
+        for (size_t i = 0; i < 3; ++i) {
+            if (!isdigit(*ptr)) {
+                if (i == 0) {
+                    return -1;
+                }
+                break;
+            }
+            day *= 10;
+            day += (*ptr) - '0';
+            ptr++;
+        }
+
+        if (*ptr == '/') {
+            ptr++;
+            ssize_t num_chars =
+                parse_transition_time(ptr, &hour, &minute, &second);
+            if (num_chars < 0) {
+                return -1;
+            }
+            ptr += num_chars;
+        }
+
+        DayRule *rv = PyMem_Calloc(1, sizeof(DayRule));
+        if (rv == NULL) {
+            return -1;
+        }
+
+        if (dayrule_new(julian, day, hour, minute, second, rv)) {
+            PyMem_Free(rv);
+            return -1;
+        }
+        *out = (TransitionRuleType *)rv;
+    }
+
+    return ptr - p;
+}
+
+/* Parse the time portion of a transition rule (e.g. following an /) */
+static ssize_t
+parse_transition_time(const char *const p, int8_t *hour, int8_t *minute,
+                      int8_t *second)
+{
+    // From the spec:
+    //
+    //   The time has the same format as offset except that no leading sign
+    //   ( '-' or '+' ) is allowed.
+    //
+    // The format for the offset is:
+    //
+    //   h[h][:mm[:ss]]
+    //
+    // RFC 8536 also allows transition times to be signed and to range from
+    // -167 to +167, but the current version only supports [0, 99].
+    //
+    // TODO: Support the full range of transition hours.
+    int8_t *components[3] = {hour, minute, second};
+    const char *ptr = p;
+    int8_t sign = 1;
+
+    if (*ptr == '-' || *ptr == '+') {
+        if (*ptr == '-') {
+            sign = -1;
+        }
+        ptr++;
+    }
+
+    for (size_t i = 0; i < 3; ++i) {
+        if (i > 0) {
+            if (*ptr != ':') {
+                break;
+            }
+            ptr++;
+        }
+
+        uint8_t buff = 0;
+        for (size_t j = 0; j < 2; j++) {
+            if (!isdigit(*ptr)) {
+                if (i == 0 && j > 0) {
+                    break;
+                }
+                return -1;
+            }
+
+            buff *= 10;
+            buff += (*ptr) - '0';
+            ptr++;
+        }
+
+        *(components[i]) = sign * buff;
+    }
+
+    return ptr - p;
+}
+
+/* Constructor for a _tzrule.
+ *
+ * If `dst_abbr` is NULL, this will construct an "STD-only" _tzrule, in which
+ * case `dst_offset` will be ignored and `start` and `end` are expected to be
+ * NULL as well.
+ *
+ * Returns 0 on success.
+ */
+static int
+build_tzrule(PyObject *std_abbr, PyObject *dst_abbr, long std_offset,
+             long dst_offset, TransitionRuleType *start,
+             TransitionRuleType *end, _tzrule *out)
+{
+    _tzrule rv = {0};
+
+    rv.start = start;
+    rv.end = end;
+
+    if (build_ttinfo(std_offset, 0, std_abbr, &rv.std)) {
+        goto error;
+    }
+
+    if (dst_abbr != NULL) {
+        rv.dst_diff = dst_offset - std_offset;
+        if (build_ttinfo(dst_offset, rv.dst_diff, dst_abbr, &rv.dst)) {
+            goto error;
+        }
+    }
+    else {
+        rv.std_only = 1;
+    }
+
+    *out = rv;
+
+    return 0;
+error:
+    xdecref_ttinfo(&rv.std);
+    xdecref_ttinfo(&rv.dst);
+    return -1;
+}
+
+/* Destructor for _tzrule. */
+static void
+free_tzrule(_tzrule *tzrule)
+{
+    xdecref_ttinfo(&(tzrule->std));
+    if (!tzrule->std_only) {
+        xdecref_ttinfo(&(tzrule->dst));
+    }
+
+    if (tzrule->start != NULL) {
+        PyMem_Free(tzrule->start);
+    }
+
+    if (tzrule->end != NULL) {
+        PyMem_Free(tzrule->end);
+    }
+}
+
+/* Calculate DST offsets from transitions and UTC offsets
+ *
+ * This is necessary because each C `ttinfo` only contains the UTC offset,
+ * time zone abbreviation and an isdst boolean - it does not include the
+ * amount of the DST offset, but we need the amount for the dst() function.
+ *
+ * Thus function uses heuristics to infer what the offset should be, so it
+ * is not guaranteed that this will work for all zones. If we cannot assign
+ * a value for a given DST offset, we'll assume it's 1H rather than 0H, so
+ * bool(dt.dst()) will always match ttinfo.isdst.
+ */
+static void
+utcoff_to_dstoff(size_t *trans_idx, long *utcoffs, long *dstoffs,
+                 unsigned char *isdsts, size_t num_transitions,
+                 size_t num_ttinfos)
+{
+    size_t dst_count = 0;
+    size_t dst_found = 0;
+    for (size_t i = 0; i < num_ttinfos; ++i) {
+        dst_count++;
+    }
+
+    for (size_t i = 1; i < num_transitions; ++i) {
+        if (dst_count == dst_found) {
+            break;
+        }
+
+        size_t idx = trans_idx[i];
+        size_t comp_idx = trans_idx[i - 1];
+
+        // Only look at DST offsets that have nto been assigned already
+        if (!isdsts[idx] || dstoffs[idx] != 0) {
+            continue;
+        }
+
+        long dstoff = 0;
+        long utcoff = utcoffs[idx];
+
+        if (!isdsts[comp_idx]) {
+            dstoff = utcoff - utcoffs[comp_idx];
+        }
+
+        if (!dstoff && idx < (num_ttinfos - 1)) {
+            comp_idx = trans_idx[i + 1];
+
+            // If the following transition is also DST and we couldn't find
+            // the DST offset by this point, we're going to have to skip it
+            // and hope this transition gets assigned later
+            if (isdsts[comp_idx]) {
+                continue;
+            }
+
+            dstoff = utcoff - utcoffs[comp_idx];
+        }
+
+        if (dstoff) {
+            dst_found++;
+            dstoffs[idx] = dstoff;
+        }
+    }
+
+    if (dst_found < dst_count) {
+        // If there are time zones we didn't find a value for, we'll end up
+        // with dstoff = 0 for something where isdst=1. This is obviously
+        // wrong — one hour will be a much better guess than 0.
+        for (size_t idx = 0; idx < num_ttinfos; ++idx) {
+            if (isdsts[idx] && !dstoffs[idx]) {
+                dstoffs[idx] = 3600;
+            }
+        }
+    }
+}
+
+#define _swap(x, y, buffer) \
+    buffer = x;             \
+    x = y;                  \
+    y = buffer;
+
+/* Calculate transitions in local time from UTC time and offsets.
+ *
+ * We want to know when each transition occurs, denominated in the number of
+ * nominal wall-time seconds between 1970-01-01T00:00:00 and the transition in
+ * *local time* (note: this is *not* equivalent to the output of
+ * datetime.timestamp, which is the total number of seconds actual elapsed
+ * since 1970-01-01T00:00:00Z in UTC).
+ *
+ * This is an ambiguous question because "local time" can be ambiguous — but it
+ * is disambiguated by the `fold` parameter, so we allocate two arrays:
+ *
+ *  trans_local[0]: The wall-time transitions for fold=0
+ *  trans_local[1]: The wall-time transitions for fold=1
+ *
+ * This returns 0 on success and a negative number of failure. The trans_local
+ * arrays must be freed if they are not NULL.
+ */
+static int
+ts_to_local(size_t *trans_idx, int64_t *trans_utc, long *utcoff,
+            int64_t *trans_local[2], size_t num_ttinfos,
+            size_t num_transitions)
+{
+    if (num_transitions == 0) {
+        return 0;
+    }
+
+    // Copy the UTC transitions into each array to be modified in place later
+    for (size_t i = 0; i < 2; ++i) {
+        trans_local[i] = PyMem_Malloc(num_transitions * sizeof(int64_t));
+        if (trans_local[i] == NULL) {
+            return -1;
+        }
+
+        memcpy(trans_local[i], trans_utc, num_transitions * sizeof(int64_t));
+    }
+
+    int64_t offset_0, offset_1, buff;
+    if (num_ttinfos > 1) {
+        offset_0 = utcoff[0];
+        offset_1 = utcoff[trans_idx[0]];
+
+        if (offset_1 > offset_0) {
+            _swap(offset_0, offset_1, buff);
+        }
+    }
+    else {
+        offset_0 = utcoff[0];
+        offset_1 = utcoff[0];
+    }
+
+    trans_local[0][0] += offset_0;
+    trans_local[1][0] += offset_1;
+
+    for (size_t i = 1; i < num_transitions; ++i) {
+        offset_0 = utcoff[trans_idx[i - 1]];
+        offset_1 = utcoff[trans_idx[i]];
+
+        if (offset_1 > offset_0) {
+            _swap(offset_1, offset_0, buff);
+        }
+
+        trans_local[0][i] += offset_0;
+        trans_local[1][i] += offset_1;
+    }
+
+    return 0;
+}
+
+/* Simple bisect_right binary search implementation */
+static size_t
+_bisect(const int64_t value, const int64_t *arr, size_t size)
+{
+    size_t lo = 0;
+    size_t hi = size;
+    size_t m;
+
+    while (lo < hi) {
+        m = (lo + hi) / 2;
+        if (arr[m] > value) {
+            hi = m;
+        }
+        else {
+            lo = m + 1;
+        }
+    }
+
+    return hi;
+}
+
+/* Find the ttinfo rules that apply at a given local datetime. */
+static _ttinfo *
+find_ttinfo(PyZoneInfo_ZoneInfo *self, PyObject *dt)
+{
+    // datetime.time has a .tzinfo attribute that passes None as the dt
+    // argument; it only really has meaning for fixed-offset zones.
+    if (dt == Py_None) {
+        if (self->fixed_offset) {
+            return &(self->tzrule_after.std);
+        }
+        else {
+            return &NO_TTINFO;
+        }
+    }
+
+    int64_t ts;
+    if (get_local_timestamp(dt, &ts)) {
+        return NULL;
+    }
+
+    unsigned char fold = PyDateTime_DATE_GET_FOLD(dt);
+    assert(fold < 2);
+    int64_t *local_transitions = self->trans_list_wall[fold];
+    size_t num_trans = self->num_transitions;
+
+    if (num_trans && ts < local_transitions[0]) {
+        return self->ttinfo_before;
+    }
+    else if (!num_trans || ts > local_transitions[self->num_transitions - 1]) {
+        return find_tzrule_ttinfo(&(self->tzrule_after), ts, fold,
+                                  PyDateTime_GET_YEAR(dt));
+    }
+    else {
+        size_t idx = _bisect(ts, local_transitions, self->num_transitions) - 1;
+        assert(idx < self->num_transitions);
+        return self->trans_ttinfos[idx];
+    }
+}
+
+static int
+is_leap_year(int year)
+{
+    const unsigned int ayear = (unsigned int)year;
+    return ayear % 4 == 0 && (ayear % 100 != 0 || ayear % 400 == 0);
+}
+
+/* Calculates ordinal datetime from year, month and day. */
+static int
+ymd_to_ord(int y, int m, int d)
+{
+    y -= 1;
+    int days_before_year = (y * 365) + (y / 4) - (y / 100) + (y / 400);
+    int yearday = DAYS_BEFORE_MONTH[m];
+    if (m > 2 && is_leap_year(y + 1)) {
+        yearday += 1;
+    }
+
+    return days_before_year + yearday + d;
+}
+
+/* Calculate the number of seconds since 1970-01-01 in local time.
+ *
+ * This gets a datetime in the same "units" as self->trans_list_wall so that we
+ * can easily determine which transitions a datetime falls between. See the
+ * comment above ts_to_local for more information.
+ * */
+static int
+get_local_timestamp(PyObject *dt, int64_t *local_ts)
+{
+    assert(local_ts != NULL);
+
+    int hour, minute, second;
+    int ord;
+    if (PyDateTime_CheckExact(dt)) {
+        int y = PyDateTime_GET_YEAR(dt);
+        int m = PyDateTime_GET_MONTH(dt);
+        int d = PyDateTime_GET_DAY(dt);
+        hour = PyDateTime_DATE_GET_HOUR(dt);
+        minute = PyDateTime_DATE_GET_MINUTE(dt);
+        second = PyDateTime_DATE_GET_SECOND(dt);
+
+        ord = ymd_to_ord(y, m, d);
+    }
+    else {
+        PyObject *num = PyObject_CallMethod(dt, "toordinal", NULL);
+        if (num == NULL) {
+            return -1;
+        }
+
+        ord = PyLong_AsLong(num);
+        Py_DECREF(num);
+        if (ord == -1 && PyErr_Occurred()) {
+            return -1;
+        }
+
+        num = PyObject_GetAttrString(dt, "hour");
+        if (num == NULL) {
+            return -1;
+        }
+        hour = PyLong_AsLong(num);
+        Py_DECREF(num);
+        if (hour == -1) {
+            return -1;
+        }
+
+        num = PyObject_GetAttrString(dt, "minute");
+        if (num == NULL) {
+            return -1;
+        }
+        minute = PyLong_AsLong(num);
+        Py_DECREF(num);
+        if (minute == -1) {
+            return -1;
+        }
+
+        num = PyObject_GetAttrString(dt, "second");
+        if (num == NULL) {
+            return -1;
+        }
+        second = PyLong_AsLong(num);
+        Py_DECREF(num);
+        if (second == -1) {
+            return -1;
+        }
+    }
+
+    *local_ts = (int64_t)(ord - EPOCHORDINAL) * 86400 +
+                (int64_t)(hour * 3600 + minute * 60 + second);
+
+    return 0;
+}
+
+/////
+// Functions for cache handling
+
+/* Constructor for StrongCacheNode */
+static StrongCacheNode *
+strong_cache_node_new(PyObject *key, PyObject *zone)
+{
+    StrongCacheNode *node = PyMem_Malloc(sizeof(StrongCacheNode));
+    if (node == NULL) {
+        return NULL;
+    }
+
+    Py_INCREF(key);
+    Py_INCREF(zone);
+
+    node->next = NULL;
+    node->prev = NULL;
+    node->key = key;
+    node->zone = zone;
+
+    return node;
+}
+
+/* Destructor for StrongCacheNode */
+void
+strong_cache_node_free(StrongCacheNode *node)
+{
+    Py_XDECREF(node->key);
+    Py_XDECREF(node->zone);
+
+    PyMem_Free(node);
+}
+
+/* Frees all nodes at or after a specified root in the strong cache.
+ *
+ * This can be used on the root node to free the entire cache or it can be used
+ * to clear all nodes that have been expired (which, if everything is going
+ * right, will actually only be 1 node at a time).
+ */
+void
+strong_cache_free(StrongCacheNode *root)
+{
+    StrongCacheNode *node = root;
+    StrongCacheNode *next_node;
+    while (node != NULL) {
+        next_node = node->next;
+        strong_cache_node_free(node);
+
+        node = next_node;
+    }
+}
+
+/* Removes a node from the cache and update its neighbors.
+ *
+ * This is used both when ejecting a node from the cache and when moving it to
+ * the front of the cache.
+ */
+static void
+remove_from_strong_cache(StrongCacheNode *node)
+{
+    if (ZONEINFO_STRONG_CACHE == node) {
+        ZONEINFO_STRONG_CACHE = node->next;
+    }
+
+    if (node->prev != NULL) {
+        node->prev->next = node->next;
+    }
+
+    if (node->next != NULL) {
+        node->next->prev = node->prev;
+    }
+
+    node->next = NULL;
+    node->prev = NULL;
+}
+
+/* Retrieves the node associated with a key, if it exists.
+ *
+ * This traverses the strong cache until it finds a matching key and returns a
+ * pointer to the relevant node if found. Returns NULL if no node is found.
+ *
+ * root may be NULL, indicating an empty cache.
+ */
+static StrongCacheNode *
+find_in_strong_cache(const StrongCacheNode *const root, PyObject *const key)
+{
+    const StrongCacheNode *node = root;
+    while (node != NULL) {
+        if (PyObject_RichCompareBool(key, node->key, Py_EQ)) {
+            return (StrongCacheNode *)node;
+        }
+
+        node = node->next;
+    }
+
+    return NULL;
+}
+
+/* Ejects a given key from the class's strong cache, if applicable.
+ *
+ * This function is used to enable the per-key functionality in clear_cache.
+ */
+static void
+eject_from_strong_cache(const PyTypeObject *const type, PyObject *key)
+{
+    if (type != &PyZoneInfo_ZoneInfoType) {
+        return;
+    }
+
+    StrongCacheNode *node = find_in_strong_cache(ZONEINFO_STRONG_CACHE, key);
+    if (node != NULL) {
+        remove_from_strong_cache(node);
+
+        strong_cache_node_free(node);
+    }
+}
+
+/* Moves a node to the front of the LRU cache.
+ *
+ * The strong cache is an LRU cache, so whenever a given node is accessed, if
+ * it is not at the front of the cache, it needs to be moved there.
+ */
+static void
+move_strong_cache_node_to_front(StrongCacheNode **root, StrongCacheNode *node)
+{
+    StrongCacheNode *root_p = *root;
+    if (root_p == node) {
+        return;
+    }
+
+    remove_from_strong_cache(node);
+
+    node->prev = NULL;
+    node->next = root_p;
+
+    if (root_p != NULL) {
+        root_p->prev = node;
+    }
+
+    *root = node;
+}
+
+/* Retrieves a ZoneInfo from the strong cache if it's present.
+ *
+ * This function finds the ZoneInfo by key and if found will move the node to
+ * the front of the LRU cache and return a new reference to it. It returns NULL
+ * if the key is not in the cache.
+ *
+ * The strong cache is currently only implemented for the base class, so this
+ * always returns a cache miss for subclasses.
+ */
+static PyObject *
+zone_from_strong_cache(const PyTypeObject *const type, PyObject *const key)
+{
+    if (type != &PyZoneInfo_ZoneInfoType) {
+        return NULL;  // Strong cache currently only implemented for base class
+    }
+
+    StrongCacheNode *node = find_in_strong_cache(ZONEINFO_STRONG_CACHE, key);
+
+    if (node != NULL) {
+        move_strong_cache_node_to_front(&ZONEINFO_STRONG_CACHE, node);
+        Py_INCREF(node->zone);
+        return node->zone;
+    }
+
+    return NULL;  // Cache miss
+}
+
+/* Inserts a new key into the strong LRU cache.
+ *
+ * This function is only to be used after a cache miss — it creates a new node
+ * at the front of the cache and ejects any stale entries (keeping the size of
+ * the cache to at most ZONEINFO_STRONG_CACHE_MAX_SIZE).
+ */
+static void
+update_strong_cache(const PyTypeObject *const type, PyObject *key,
+                    PyObject *zone)
+{
+    if (type != &PyZoneInfo_ZoneInfoType) {
+        return;
+    }
+
+    StrongCacheNode *new_node = strong_cache_node_new(key, zone);
+
+    move_strong_cache_node_to_front(&ZONEINFO_STRONG_CACHE, new_node);
+
+    StrongCacheNode *node = new_node->next;
+    for (size_t i = 1; i < ZONEINFO_STRONG_CACHE_MAX_SIZE; ++i) {
+        if (node == NULL) {
+            return;
+        }
+        node = node->next;
+    }
+
+    // Everything beyond this point needs to be freed
+    if (node != NULL) {
+        if (node->prev != NULL) {
+            node->prev->next = NULL;
+        }
+        strong_cache_free(node);
+    }
+}
+
+/* Clears all entries into a type's strong cache.
+ *
+ * Because the strong cache is not implemented for subclasses, this is a no-op
+ * for everything except the base class.
+ */
+void
+clear_strong_cache(const PyTypeObject *const type)
+{
+    if (type != &PyZoneInfo_ZoneInfoType) {
+        return;
+    }
+
+    strong_cache_free(ZONEINFO_STRONG_CACHE);
+}
+
+static PyObject *
+new_weak_cache()
+{
+    PyObject *weakref_module = PyImport_ImportModule("weakref");
+    if (weakref_module == NULL) {
+        return NULL;
+    }
+
+    PyObject *weak_cache =
+        PyObject_CallMethod(weakref_module, "WeakValueDictionary", "");
+    Py_DECREF(weakref_module);
+    return weak_cache;
+}
+
+static int
+initialize_caches()
+{
+    if (TIMEDELTA_CACHE == NULL) {
+        TIMEDELTA_CACHE = PyDict_New();
+    }
+    else {
+        Py_INCREF(TIMEDELTA_CACHE);
+    }
+
+    if (TIMEDELTA_CACHE == NULL) {
+        return -1;
+    }
+
+    if (ZONEINFO_WEAK_CACHE == NULL) {
+        ZONEINFO_WEAK_CACHE = new_weak_cache();
+    }
+    else {
+        Py_INCREF(ZONEINFO_WEAK_CACHE);
+    }
+
+    if (ZONEINFO_WEAK_CACHE == NULL) {
+        return -1;
+    }
+
+    return 0;
+}
+
+static PyObject *
+zoneinfo_init_subclass(PyTypeObject *cls, PyObject *args, PyObject **kwargs)
+{
+    PyObject *weak_cache = new_weak_cache();
+    if (weak_cache == NULL) {
+        return NULL;
+    }
+
+    PyObject_SetAttrString((PyObject *)cls, "_weak_cache", weak_cache);
+    Py_RETURN_NONE;
+}
+
+/////
+// Specify the ZoneInfo type
+static PyMethodDef zoneinfo_methods[] = {
+    {"clear_cache", (PyCFunction)(void (*)(void))zoneinfo_clear_cache,
+     METH_VARARGS | METH_KEYWORDS | METH_CLASS,
+     PyDoc_STR("Clear the ZoneInfo cache.")},
+    {"no_cache", (PyCFunction)(void (*)(void))zoneinfo_no_cache,
+     METH_VARARGS | METH_KEYWORDS | METH_CLASS,
+     PyDoc_STR("Get a new instance of ZoneInfo, bypassing the cache.")},
+    {"from_file", (PyCFunction)(void (*)(void))zoneinfo_from_file,
+     METH_VARARGS | METH_KEYWORDS | METH_CLASS,
+     PyDoc_STR("Create a ZoneInfo file from a file object.")},
+    {"utcoffset", (PyCFunction)zoneinfo_utcoffset, METH_O,
+     PyDoc_STR("Retrieve a timedelta representing the UTC offset in a zone at "
+               "the given datetime.")},
+    {"dst", (PyCFunction)zoneinfo_dst, METH_O,
+     PyDoc_STR("Retrieve a timedelta representing the amount of DST applied "
+               "in a zone at the given datetime.")},
+    {"tzname", (PyCFunction)zoneinfo_tzname, METH_O,
+     PyDoc_STR("Retrieve a string containing the abbreviation for the time "
+               "zone that applies in a zone at a given datetime.")},
+    {"fromutc", (PyCFunction)zoneinfo_fromutc, METH_O,
+     PyDoc_STR("Given a datetime with local time in UTC, retrieve an adjusted "
+               "datetime in local time.")},
+    {"__reduce__", (PyCFunction)zoneinfo_reduce, METH_NOARGS,
+     PyDoc_STR("Function for serialization with the pickle protocol.")},
+    {"_unpickle", (PyCFunction)zoneinfo__unpickle, METH_VARARGS | METH_CLASS,
+     PyDoc_STR("Private method used in unpickling.")},
+    {"__init_subclass__", (PyCFunction)(void (*)(void))zoneinfo_init_subclass,
+     METH_VARARGS | METH_KEYWORDS,
+     PyDoc_STR("Function to initialize subclasses.")},
+    {NULL} /* Sentinel */
+};
+
+static PyMemberDef zoneinfo_members[] = {
+    {.name = "key",
+     .offset = offsetof(PyZoneInfo_ZoneInfo, key),
+     .type = T_OBJECT_EX,
+     .flags = READONLY,
+     .doc = NULL},
+    {NULL}, /* Sentinel */
+};
+
+static PyTypeObject PyZoneInfo_ZoneInfoType = {
+    PyVarObject_HEAD_INIT(NULL, 0)  //
+        .tp_name = "zoneinfo.ZoneInfo",
+    .tp_basicsize = sizeof(PyZoneInfo_ZoneInfo),
+    .tp_weaklistoffset = offsetof(PyZoneInfo_ZoneInfo, weakreflist),
+    .tp_repr = (reprfunc)zoneinfo_repr,
+    .tp_str = (reprfunc)zoneinfo_str,
+    .tp_getattro = PyObject_GenericGetAttr,
+    .tp_flags = (Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE),
+    /* .tp_doc = zoneinfo_doc, */
+    .tp_methods = zoneinfo_methods,
+    .tp_members = zoneinfo_members,
+    .tp_new = zoneinfo_new,
+    .tp_dealloc = zoneinfo_dealloc,
+};
+
+/////
+// Specify the _zoneinfo module
+static PyMethodDef module_methods[] = {{NULL, NULL}};
+static void
+module_free()
+{
+    Py_XDECREF(_tzpath_find_tzfile);
+    _tzpath_find_tzfile = NULL;
+
+    Py_XDECREF(_common_mod);
+    _common_mod = NULL;
+
+    Py_XDECREF(io_open);
+    io_open = NULL;
+
+    xdecref_ttinfo(&NO_TTINFO);
+
+    Py_XDECREF(TIMEDELTA_CACHE);
+    if (!Py_REFCNT(TIMEDELTA_CACHE)) {
+        TIMEDELTA_CACHE = NULL;
+    }
+
+    Py_XDECREF(ZONEINFO_WEAK_CACHE);
+    if (!Py_REFCNT(ZONEINFO_WEAK_CACHE)) {
+        ZONEINFO_WEAK_CACHE = NULL;
+    }
+
+    strong_cache_free(ZONEINFO_STRONG_CACHE);
+    ZONEINFO_STRONG_CACHE = NULL;
+}
+
+static int
+zoneinfomodule_exec(PyObject *m)
+{
+    PyDateTime_IMPORT;
+    PyZoneInfo_ZoneInfoType.tp_base = PyDateTimeAPI->TZInfoType;
+    if (PyType_Ready(&PyZoneInfo_ZoneInfoType) < 0) {
+        goto error;
+    }
+
+    Py_INCREF(&PyZoneInfo_ZoneInfoType);
+    PyModule_AddObject(m, "ZoneInfo", (PyObject *)&PyZoneInfo_ZoneInfoType);
+
+    /* Populate imports */
+    PyObject *_tzpath_module = PyImport_ImportModule("zoneinfo._tzpath");
+    if (_tzpath_module == NULL) {
+        goto error;
+    }
+
+    _tzpath_find_tzfile =
+        PyObject_GetAttrString(_tzpath_module, "find_tzfile");
+    Py_DECREF(_tzpath_module);
+    if (_tzpath_find_tzfile == NULL) {
+        goto error;
+    }
+
+    PyObject *io_module = PyImport_ImportModule("io");
+    if (io_module == NULL) {
+        goto error;
+    }
+
+    io_open = PyObject_GetAttrString(io_module, "open");
+    Py_DECREF(io_module);
+    if (io_open == NULL) {
+        goto error;
+    }
+
+    _common_mod = PyImport_ImportModule("zoneinfo._common");
+    if (_common_mod == NULL) {
+        goto error;
+    }
+
+    if (NO_TTINFO.utcoff == NULL) {
+        NO_TTINFO.utcoff = Py_None;
+        NO_TTINFO.dstoff = Py_None;
+        NO_TTINFO.tzname = Py_None;
+
+        for (size_t i = 0; i < 3; ++i) {
+            Py_INCREF(Py_None);
+        }
+    }
+
+    if (initialize_caches()) {
+        goto error;
+    }
+
+    return 0;
+
+error:
+    return -1;
+}
+
+static PyModuleDef_Slot zoneinfomodule_slots[] = {
+    {Py_mod_exec, zoneinfomodule_exec}, {0, NULL}};
+
+static struct PyModuleDef zoneinfomodule = {
+    PyModuleDef_HEAD_INIT,
+    .m_name = "_zoneinfo",
+    .m_doc = "C implementation of the zoneinfo module",
+    .m_size = 0,
+    .m_methods = module_methods,
+    .m_slots = zoneinfomodule_slots,
+    .m_free = (freefunc)module_free};
+
+PyMODINIT_FUNC
+PyInit__zoneinfo(void)
+{
+    return PyModuleDef_Init(&zoneinfomodule);
+}
diff --git a/configure b/configure
index 64bcde6bfdf..56d66d02352 100755
--- a/configure
+++ b/configure
@@ -658,6 +658,7 @@ LIBFFI_INCLUDEDIR
 PKG_CONFIG_LIBDIR
 PKG_CONFIG_PATH
 PKG_CONFIG
+TZPATH
 SHLIBS
 CFLAGSFORSHARED
 LINKFORSHARED
@@ -819,6 +820,7 @@ with_assertions
 enable_optimizations
 with_lto
 with_hash_algorithm
+with_tzpath
 with_address_sanitizer
 with_memory_sanitizer
 with_undefined_behavior_sanitizer
@@ -1525,6 +1527,9 @@ Optional Packages:
   --with-hash-algorithm=[fnv|siphash24]
                           select hash algorithm for use in Python/pyhash.c
                           (default is SipHash24)
+  --with-tzpath=<list of absolute paths separated by pathsep>
+                           Select the default time zone search path for zoneinfo.TZPATH
+
   --with-address-sanitizer
                           enable AddressSanitizer memory error detector,
                           'asan' (default is no)
@@ -10154,6 +10159,47 @@ $as_echo "default" >&6; }
 fi
 
 
+validate_tzpath() {
+    # Checks that each element of hte path is an absolute path
+    if test -z "$1"; then
+        # Empty string is allowed: it indicates no system TZPATH
+        return 0
+    fi
+
+    # Bad paths are those that don't start with /
+        if ( echo $1 | grep -qE '(^|:)([^/]|$)' ); then
+        as_fn_error $? "--with-tzpath must contain only absolute paths, not $1" "$LINENO" 5
+        return 1;
+    fi
+}
+
+TZPATH="/usr/share/zoneinfo:/usr/lib/zoneinfo:/usr/share/lib/zoneinfo:/etc/zoneinfo"
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for --with-tzpath" >&5
+$as_echo_n "checking for --with-tzpath... " >&6; }
+
+# Check whether --with-tzpath was given.
+if test "${with_tzpath+set}" = set; then :
+  withval=$with_tzpath;
+case "$withval" in
+    yes)
+        as_fn_error $? "--with-tzpath requires a value" "$LINENO" 5
+        ;;
+    *)
+        validate_tzpath "$withval"
+        TZPATH="$withval"
+        { $as_echo "$as_me:${as_lineno-$LINENO}: result: \"$withval\"" >&5
+$as_echo "\"$withval\"" >&6; }
+        ;;
+esac
+
+else
+  validate_tzpath "$TZPATH"
+ { $as_echo "$as_me:${as_lineno-$LINENO}: result: \"$TZPATH\"" >&5
+$as_echo "\"$TZPATH\"" >&6; }
+fi
+
+
+
 { $as_echo "$as_me:${as_lineno-$LINENO}: checking for --with-address-sanitizer" >&5
 $as_echo_n "checking for --with-address-sanitizer... " >&6; }
 
diff --git a/configure.ac b/configure.ac
index 21c47b56358..497d7c191d5 100644
--- a/configure.ac
+++ b/configure.ac
@@ -2946,6 +2946,42 @@ esac
 ],
 [AC_MSG_RESULT(default)])
 
+validate_tzpath() {
+    # Checks that each element of hte path is an absolute path
+    if test -z "$1"; then
+        # Empty string is allowed: it indicates no system TZPATH
+        return 0
+    fi
+
+    # Bad paths are those that don't start with /
+    dnl quadrigraphs "@<:@" and "@:>@" produce "[" and "]" in the output
+    if ( echo $1 | grep -qE '(^|:)(@<:@^/@:>@|$)' ); then
+        AC_MSG_ERROR([--with-tzpath must contain only absolute paths, not $1])
+        return 1;
+    fi
+}
+
+TZPATH="/usr/share/zoneinfo:/usr/lib/zoneinfo:/usr/share/lib/zoneinfo:/etc/zoneinfo"
+AC_MSG_CHECKING(for --with-tzpath)
+AC_ARG_WITH(tzpath,
+            AS_HELP_STRING([--with-tzpath=<list of absolute paths separated by pathsep>]
+                           [Select the default time zone search path for zoneinfo.TZPATH]),
+[
+case "$withval" in
+    yes)
+        AC_MSG_ERROR([--with-tzpath requires a value])
+        ;;
+    *)
+        validate_tzpath "$withval"
+        TZPATH="$withval"
+        AC_MSG_RESULT("$withval")
+        ;;
+esac
+],
+[validate_tzpath "$TZPATH"
+ AC_MSG_RESULT("$TZPATH")])
+AC_SUBST(TZPATH)
+
 AC_MSG_CHECKING(for --with-address-sanitizer)
 AC_ARG_WITH(address_sanitizer,
             AS_HELP_STRING([--with-address-sanitizer],
diff --git a/setup.py b/setup.py
index 794ba2f7662..68fc3120cc3 100644
--- a/setup.py
+++ b/setup.py
@@ -304,6 +304,17 @@ def find_library_file(compiler, libname, std_dirs, paths):
     else:
         assert False, "Internal error: Path not found in std_dirs or paths"
 
+def validate_tzpath():
+    base_tzpath = sysconfig.get_config_var('TZPATH')
+    if not base_tzpath:
+        return
+
+    tzpaths = base_tzpath.split(os.pathsep)
+    bad_paths = [tzpath for tzpath in tzpaths if not os.path.isabs(tzpath)]
+    if bad_paths:
+        raise ValueError('TZPATH must contain only absolute paths, '
+                         + f'found:\n{tzpaths!r}\nwith invalid paths:\n'
+                         + f'{bad_paths!r}')
 
 def find_module_file(module, dirlist):
     """Find a module in a set of possible folders. If it is not found
@@ -816,6 +827,8 @@ class PyBuildExt(build_ext):
         # uses modf().
         self.add(Extension('_datetime', ['_datetimemodule.c'],
                            libraries=['m']))
+        # zoneinfo module
+        self.add(Extension('_zoneinfo', ['_zoneinfo.c'])),
         # random number generator implemented in C
         self.add(Extension("_random", ["_randommodule.c"],
                            extra_compile_args=['-DPy_BUILD_CORE_MODULE']))
@@ -2495,6 +2508,7 @@ def main():
         ProcessPoolExecutor = None
 
     sys.modules['concurrent.futures.process'] = DummyProcess
+    validate_tzpath()
 
     # turn off warnings when deprecated modules are imported
     import warnings