Version 3.27.0rc0

.github/workflows/ci-cd.yaml

@@ -145,10 +145,16 @@ jobs:
           sudo apt-get -y install python3-dbus build-essential libpoppler-cpp-dev pkg-config python3-dev tesseract-ocr
           pip install --upgrade pdftotext pytesseract
 
-      - name: Install all other dependencies
+      - name: Install all other dependencies (GIL)
+        if: "${{ !matrix.disable-gil }}"
         run: |
           pip install --upgrade coveralls -r requirements.txt -r tests/requirements_pytest.txt
 
+      - name: Install all other dependencies (free-threaded)
+        if: "${{ matrix.disable-gil }}"
+        run: |
+          pip install --upgrade coveralls -r requirements-free-threaded.txt -r tests/requirements_pytest.txt
+
 #      - name: Install all other dependencies (py12)
 #        if: matrix.python-version == '3.12'
 #        run: |

CHANGELOG.rst

@@ -39,11 +39,13 @@ Unreleased
 
 Added
 -----
-* New Sub-directive in ``pypdf`` Filter: Added ``extraction_mode`` sub-directive.
-* Python 3.13 Testing: **webchanges** now is tested on Python 3.13 before releasing. However, the `aioxmpp
+* Python 3.13: **webchanges** now is tested on Python 3.13 before releasing. However, the `aioxmpp
   <https://pypi.org/project/aioxmpp/>`__ library required by the ``xmpp`` reporter will not install in Python 3.13 (at
   least on Windows), and the development of the `library <https://codeberg.org/jssfr/aioxmpp>`__ has been
   halted.
+* Python 3.13t (free-threaded, GIL-free) remains unsupported due to lack of support by dependencies such
+  as ``lxml``.
+* New Sub-directive in ``pypdf`` Filter: Added ``extraction_mode`` sub-directive.
 
 Internals
 ---------

requirements-free-threaded.txt

@@ -0,0 +1,10 @@
+colorama; sys_platform == "win32"
+cssselect
+h2
+html2text
+httpx
+markdown2
+msgpack
+platformdirs
+pyyaml
+tzdata; sys_platform == "win32"

requirements.txt

@@ -3,7 +3,7 @@ cssselect
 h2
 html2text
 httpx
-lxml>=5.3.0  # prior versions don't build in 3.13
+lxml >= 5.3.0  # prior versions don't build in 3.13
 markdown2
 msgpack
 platformdirs

webchanges/filters.py

@@ -8,6 +8,7 @@
 import csv
 import hashlib
 import html
+import importlib.util
 import io
 import itertools
 import logging
@@ -25,8 +26,14 @@
 
 import html2text
 import yaml
-from lxml import etree  # noqa: S410 insecure use of XML modules, prefer "defusedxml". TODO
-from lxml.cssselect import CSSSelector  # noqa: S410 insecure use of XML ... "defusedxml". TODO
+
+try:
+    from lxml import etree  # noqa: S410 insecure use of XML modules, prefer "defusedxml". TODO
+    from lxml.cssselect import CSSSelector  # noqa: S410 insecure use of XML ... "defusedxml". TODO
+except ImportError as e:
+    from xml import etree
+
+    CSSSelector = str(e)
 
 from webchanges import __project_name__
 from webchanges.util import TrackSubClasses
@@ -412,7 +419,8 @@ def filter(self, data: str | bytes, mime_type: str, subfilter: dict[str, Any]) -
         if isinstance(bs4, str):
             self.raise_import_error('BeautifulSoup', self.__kind__, bs4)
 
-        soup = bs4.BeautifulSoup(data, features='lxml')
+        bs4_features = 'lxml' if importlib.util.find_spec('lxml') is not None else 'html'
+        soup = bs4.BeautifulSoup(data, features=bs4_features)
 
         if isinstance(jsbeautifier, str):
             logger.warning(
@@ -551,7 +559,8 @@ def filter(self, data: str | bytes, mime_type: str, subfilter: dict[str, Any]) -
             if isinstance(bs4, str):
                 self.raise_import_error('BeautifulSoup', self.__kind__, bs4)
 
-            bs4_parser: str = options.pop('parser', 'lxml')
+            default_bs4_parser = 'lxml' if importlib.util.find_spec('lxml') is not None else 'html'
+            bs4_parser: str = options.pop('parser', default_bs4_parser)
             try:
                 soup = bs4.BeautifulSoup(data, bs4_parser)
             except bs4.FeatureNotFound: