From ff967e5636f25474c814ed9f97036292d5c6fc10 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Afonso=20Janu=C3=A1rio?= <afonso-januario@hotmail.com>
Date: Tue, 8 Sep 2026 10:08:17 +0100
Subject: [PATCH] Avoid the chardet.universaldetector deprecation warning

chardet 7 moved UniversalDetector to chardet.detector and turned the
old chardet.universaldetector import path into a stub that raises a
DeprecationWarning on import. Projects that run their test suites with
warnings turned into errors (like Flexget, per the linked issue) hit a
hard failure the moment html5lib falls back to chardet for encoding
detection.

Try the new import path first and fall back to the old one, so this
keeps working across both old and new chardet releases without
triggering the warning on the new ones. Fixes #601.
---
 AUTHORS.rst              |  1 +
 CHANGES.rst              |  3 +++
 html5lib/_inputstream.py | 12 +++++++++---
 3 files changed, 13 insertions(+), 3 deletions(-)

diff --git a/AUTHORS.rst b/AUTHORS.rst
index 90401390..998be2f4 100644
--- a/AUTHORS.rst
+++ b/AUTHORS.rst
@@ -63,4 +63,5 @@ Patches and suggestions
 - Ville Skyttä
 - Hugo van Kemenade
 - Mark Vasilkov
+- Afonso Januário
 
diff --git a/CHANGES.rst b/CHANGES.rst
index 47dcda3a..ca514ece 100644
--- a/CHANGES.rst
+++ b/CHANGES.rst
@@ -20,6 +20,9 @@ Bug fixes:
 
 * The sanitizer now permits ``<summary>`` tags. It used to allow ``<details>``
   already. (#423)
+* Encoding detection no longer triggers a ``DeprecationWarning`` under
+  chardet 7+, which moved ``UniversalDetector`` out of
+  ``chardet.universaldetector``. (#601)
 
 1.1
 ~~~
diff --git a/html5lib/_inputstream.py b/html5lib/_inputstream.py
index a93b5a4e..a024f4b6 100644
--- a/html5lib/_inputstream.py
+++ b/html5lib/_inputstream.py
@@ -483,10 +483,16 @@ def determineEncoding(self, chardet=True):
         # Guess with chardet, if available
         if chardet:
             try:
-                from chardet.universaldetector import UniversalDetector
+                # chardet 7+ moved UniversalDetector here; the older
+                # chardet.universaldetector path still works but now raises
+                # a DeprecationWarning.
+                from chardet.detector import UniversalDetector
             except ImportError:
-                pass
-            else:
+                try:
+                    from chardet.universaldetector import UniversalDetector
+                except ImportError:
+                    UniversalDetector = None
+            if UniversalDetector is not None:
                 buffers = []
                 detector = UniversalDetector()
                 while not detector.done: