linkcheck: Fix conversion from UTC time to the UNIX epoch (#11649)

Co-authored-by: Adam Turner <9087854+AA-Turner@users.noreply.github.com>

CHANGES

@@ -6,6 +6,8 @@ Bugs fixed
 
 * #11618: Fix a regression in the MoveModuleTargets transform,
   introduced in #10478 (#9662).
+* #11649: linkcheck: Fix conversions from UTC to UNIX time
+  for timezones west of London.
 
 Release 7.2.3 (released Aug 23, 2023)
 =====================================
@@ -24,7 +26,7 @@ Bugs fixed
   when ``autodoc_preserve_defaults`` is ``True``.
 * Restore support string methods on path objects.
   This is deprecated and will be removed in Sphinx 8.
-  Use :py:func`os.fspath` to convert :py:class:`~pathlib.Path` objects to strings,
+  Use :py:func:`os.fspath` to convert :py:class:`~pathlib.Path` objects to strings,
   or :py:class:`~pathlib.Path`'s methods to work with path objects.
 
 Release 7.2.2 (released Aug 17, 2023)

sphinx/builders/linkcheck.py

@@ -7,7 +7,6 @@
 import re
 import socket
 import time
-from email.utils import parsedate_tz
 from html.parser import HTMLParser
 from os import path
 from queue import PriorityQueue, Queue
@@ -29,6 +28,7 @@
     red,
     turquoise,
 )
+from sphinx.util.http_date import rfc1123_to_epoch
 from sphinx.util.nodes import get_node_line
 
 if TYPE_CHECKING:
@@ -488,11 +488,8 @@ def limit_rate(self, response_url: str, retry_after: str) -> float | None:
             except ValueError:
                 try:
                     # An HTTP-date: time of next attempt.
-                    parsed = parsedate_tz(retry_after)
-                    assert parsed is not None
-                    # the 10th element is the GMT offset in seconds
-                    next_check = time.mktime(parsed[:9]) - (parsed[9] or 0)
-                except (AssertionError, TypeError, ValueError):
+                    next_check = rfc1123_to_epoch(retry_after)
+                except (ValueError, TypeError):
                     # TypeError: Invalid date format.
                     # ValueError: Invalid date, e.g. Oct 52th.
                     pass

sphinx/util/http_date.py

@@ -4,7 +4,12 @@
 """
 
 import time
-from email.utils import formatdate, parsedate
+import warnings
+from email.utils import formatdate, parsedate_tz
+
+from sphinx.deprecation import RemovedInSphinx90Warning
+
+_GMT_OFFSET = float(time.localtime().tm_gmtoff)
 
 
 def epoch_to_rfc1123(epoch: float) -> str:
@@ -14,7 +19,21 @@ def epoch_to_rfc1123(epoch: float) -> str:
 
 def rfc1123_to_epoch(rfc1123: str) -> float:
     """Return epoch offset from HTTP-date string."""
-    t = parsedate(rfc1123)
-    if t:
-        return time.mktime(t)
-    raise ValueError
+    t = parsedate_tz(rfc1123)
+    if t is None:
+        raise ValueError
+    if not rfc1123.endswith(" GMT"):
+        warnings.warn(
+            "HTTP-date string does not meet RFC 7231 requirements "
+            f"(must end with 'GMT'): {rfc1123!r}",
+            RemovedInSphinx90Warning, stacklevel=3,
+        )
+    epoch_secs = time.mktime(time.struct_time(t[:9])) + _GMT_OFFSET
+    if (gmt_offset := t[9]) != 0:
+        warnings.warn(
+            "HTTP-date string does not meet RFC 7231 requirements "
+            f"(must be GMT time): {rfc1123!r}",
+            RemovedInSphinx90Warning, stacklevel=3,
+        )
+        return epoch_secs - (gmt_offset or 0)
+    return epoch_secs

tests/test_build_linkcheck.py

@@ -5,6 +5,7 @@
 import http.server
 import json
 import re
+import sys
 import textwrap
 import time
 import wsgiref.handlers
@@ -16,6 +17,7 @@
 import pytest
 from urllib3.poolmanager import PoolManager
 
+import sphinx.util.http_date
 from sphinx.builders.linkcheck import (
     CheckRequest,
     Hyperlink,
@@ -772,11 +774,22 @@ def test_too_many_requests_retry_after_int_delay(app, capsys, status):
     )
 
 
+@pytest.mark.parametrize('tz', [None, 'GMT', 'GMT+3', 'GMT-3'])
 @pytest.mark.sphinx('linkcheck', testroot='linkcheck-localserver', freshenv=True)
-def test_too_many_requests_retry_after_HTTP_date(app, capsys):
+def test_too_many_requests_retry_after_HTTP_date(tz, app, monkeypatch, capsys):
     retry_after = wsgiref.handlers.format_date_time(time.time())
-    with http_server(make_retry_after_handler([(429, retry_after), (200, None)])):
-        app.build()
+
+    with monkeypatch.context() as m:
+        if tz is not None:
+            m.setenv('TZ', tz)
+            if sys.platform != "win32":
+                time.tzset()
+            m.setattr(sphinx.util.http_date, '_GMT_OFFSET',
+                      float(time.localtime().tm_gmtoff))
+
+        with http_server(make_retry_after_handler([(429, retry_after), (200, None)])):
+            app.build()
+
     content = (app.outdir / 'output.json').read_text(encoding='utf8')
     assert json.loads(content) == {
         "filename": "index.rst",