[core] Fix support for upcoming Python 3.12 (#8130)

This also adds the following test runners:
- `3.12-dev` on `ubuntu-latest`
- `3.12-dev` on `windows-latest`
- `pypy-3.10` on `ubuntu-latest`

Authored by: Grub4K

.github/workflows/core.yml

@@ -13,13 +13,16 @@ jobs:
       matrix:
         os: [ubuntu-latest]
         # CPython 3.11 is in quick-test
-        python-version: ['3.8', '3.9', '3.10', pypy-3.7, pypy-3.8]
+        python-version: ['3.8', '3.9', '3.10', '3.12-dev', pypy-3.7, pypy-3.8, pypy-3.10]
         run-tests-ext: [sh]
         include:
         # atleast one of each CPython/PyPy tests must be in windows
         - os: windows-latest
           python-version: '3.7'
           run-tests-ext: bat
+        - os: windows-latest
+          python-version: '3.12-dev'
+          run-tests-ext: bat
         - os: windows-latest
           python-version: pypy-3.9
           run-tests-ext: bat

devscripts/update-version.py

@@ -10,14 +10,14 @@
 import argparse
 import contextlib
 import sys
-from datetime import datetime
+from datetime import datetime, timezone
 
 from devscripts.utils import read_version, run_process, write_file
 
 
 def get_new_version(version, revision):
     if not version:
-        version = datetime.utcnow().strftime('%Y.%m.%d')
+        version = datetime.now(timezone.utc).strftime('%Y.%m.%d')
 
     if revision:
         assert revision.isdigit(), 'Revision must be a number'

yt_dlp/YoutubeDL.py

@@ -2591,7 +2591,7 @@ def _fill_common_fields(self, info_dict, final=True):
                 # Working around out-of-range timestamp values (e.g. negative ones on Windows,
                 # see http://bugs.python.org/issue1646728)
                 with contextlib.suppress(ValueError, OverflowError, OSError):
-                    upload_date = datetime.datetime.utcfromtimestamp(info_dict[ts_key])
+                    upload_date = datetime.datetime.fromtimestamp(info_dict[ts_key], datetime.timezone.utc)
                     info_dict[date_key] = upload_date.strftime('%Y%m%d')
 
         live_keys = ('is_live', 'was_live')

yt_dlp/extractor/aws.py

@@ -12,7 +12,7 @@ class AWSIE(InfoExtractor):  # XXX: Conventionally, base classes should end with
 
     def _aws_execute_api(self, aws_dict, video_id, query=None):
         query = query or {}
-        amz_date = datetime.datetime.utcnow().strftime('%Y%m%dT%H%M%SZ')
+        amz_date = datetime.datetime.now(datetime.timezone.utc).strftime('%Y%m%dT%H%M%SZ')
         date = amz_date[:8]
         headers = {
             'Accept': 'application/json',

yt_dlp/extractor/goplay.py

@@ -383,9 +383,9 @@ def __get_current_timestamp():
         months = [None, 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']
         days = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
 
-        time_now = datetime.datetime.utcnow()
+        time_now = datetime.datetime.now(datetime.timezone.utc)
         format_string = "{} {} {} %H:%M:%S UTC %Y".format(days[time_now.weekday()], months[time_now.month], time_now.day)
-        time_string = datetime.datetime.utcnow().strftime(format_string)
+        time_string = time_now.strftime(format_string)
         return time_string
 
     def __str__(self):

yt_dlp/extractor/motherless.py

@@ -151,7 +151,7 @@ def _real_extract(self, url):
                     'd': 'days',
                 }
                 kwargs = {_AGO_UNITS.get(uploaded_ago[-1]): delta}
-                upload_date = (datetime.datetime.utcnow() - datetime.timedelta(**kwargs)).strftime('%Y%m%d')
+                upload_date = (datetime.datetime.now(datetime.timezone.utc) - datetime.timedelta(**kwargs)).strftime('%Y%m%d')
 
         comment_count = len(re.findall(r'''class\s*=\s*['"]media-comment-contents\b''', webpage))
         uploader_id = self._html_search_regex(

yt_dlp/extractor/panopto.py

@@ -1,7 +1,7 @@
 import calendar
 import json
 import functools
-from datetime import datetime
+from datetime import datetime, timezone
 from random import random
 
 from .common import InfoExtractor
@@ -243,7 +243,7 @@ def _mark_watched(self, base_url, video_id, delivery_info):
         invocation_id = delivery_info.get('InvocationId')
         stream_id = traverse_obj(delivery_info, ('Delivery', 'Streams', ..., 'PublicID'), get_all=False, expected_type=str)
         if invocation_id and stream_id and duration:
-            timestamp_str = f'/Date({calendar.timegm(datetime.utcnow().timetuple())}000)/'
+            timestamp_str = f'/Date({calendar.timegm(datetime.now(timezone.utc).timetuple())}000)/'
             data = {
                 'streamRequests': [
                     {

yt_dlp/networking/_urllib.py

@@ -429,7 +429,7 @@ def _send(self, request):
         except urllib.error.HTTPError as e:
             if isinstance(e.fp, (http.client.HTTPResponse, urllib.response.addinfourl)):
                 # Prevent file object from being closed when urllib.error.HTTPError is destroyed.
-                e._closer.file = None
+                e._closer.close_called = True
                 raise HTTPError(UrllibResponseAdapter(e.fp), redirect_loop='redirect error' in str(e)) from e
             raise  # unexpected
         except urllib.error.URLError as e:

yt_dlp/networking/exceptions.py

@@ -115,7 +115,7 @@ def __init__(self, http_error: HTTPError):
             hdrs=http_error.response.headers,
             fp=http_error.response
         )
-        self._closer.file = None  # Disable auto close
+        self._closer.close_called = True  # Disable auto close
         self._http_error = http_error
         HTTPError.__init__(self, http_error.response, redirect_loop=http_error.redirect_loop)
 

yt_dlp/utils/_utils.py

@@ -669,6 +669,7 @@ def replace_insane(char):
 
 def sanitize_path(s, force=False):
     """Sanitizes and normalizes path on Windows"""
+    # XXX: this handles drive relative paths (c:sth) incorrectly
     if sys.platform == 'win32':
         force = False
         drive_or_unc, _ = os.path.splitdrive(s)
@@ -687,7 +688,10 @@ def sanitize_path(s, force=False):
         sanitized_path.insert(0, drive_or_unc + os.path.sep)
     elif force and s and s[0] == os.path.sep:
         sanitized_path.insert(0, os.path.sep)
-    return os.path.join(*sanitized_path)
+    # TODO: Fix behavioral differences <3.12
+    # The workaround using `normpath` only superficially passes tests
+    # Ref: https://github.com/python/cpython/pull/100351
+    return os.path.normpath(os.path.join(*sanitized_path))
 
 
 def sanitize_url(url, *, scheme='http'):
@@ -1256,7 +1260,7 @@ def datetime_from_str(date_str, precision='auto', format='%Y%m%d'):
     if precision == 'auto':
         auto_precision = True
         precision = 'microsecond'
-    today = datetime_round(datetime.datetime.utcnow(), precision)
+    today = datetime_round(datetime.datetime.now(datetime.timezone.utc), precision)
     if date_str in ('now', 'today'):
         return today
     if date_str == 'yesterday':
@@ -1319,8 +1323,8 @@ def datetime_round(dt, precision='day'):
         'second': 1,
     }
     roundto = lambda x, n: ((x + n / 2) // n) * n
-    timestamp = calendar.timegm(dt.timetuple())
-    return datetime.datetime.utcfromtimestamp(roundto(timestamp, unit_seconds[precision]))
+    timestamp = roundto(calendar.timegm(dt.timetuple()), unit_seconds[precision])
+    return datetime.datetime.fromtimestamp(timestamp, datetime.timezone.utc)
 
 
 def hyphenate_date(date_str):