From 78d2283458b5c4a300f381171f624cd853deac01 Mon Sep 17 00:00:00 2001
From: Andrew Godwin <andrew@aeracode.org>
Date: Tue, 20 Dec 2022 13:10:35 +0000
Subject: [PATCH] Move linkifying to all http-prefixed links

---
 core/html.py            | 22 ++++++++++++++++++++--
 tests/core/test_html.py | 12 ++++++++++++
 2 files changed, 32 insertions(+), 2 deletions(-)

diff --git a/core/html.py b/core/html.py
index 5af476e..fc5cde7 100644
--- a/core/html.py
+++ b/core/html.py
@@ -1,7 +1,21 @@
+import re
+from functools import partial
+
 import bleach
 from bleach.linkifier import LinkifyFilter
 from django.utils.safestring import mark_safe
 
+url_regex = re.compile(
+    r"""\(*  # Match any opening parentheses.
+    \b(?<![@.])(?:https?://(?:(?:\w+:)?\w+@)?)  # http://
+    ([\w-]+\.)+(?:[\w-]+)(?:\:[0-9]+)?(?!\.\w)\b   # xx.yy.tld(:##)?
+    (?:[/?][^\s\{{\}}\|\\\^\[\]`<>"]*)?
+        # /path/zz (excluding "unsafe" chars from RFC 1738,
+        # except for # and ~, which happen in practice)
+    """,
+    re.IGNORECASE | re.VERBOSE | re.UNICODE,
+)
+
 
 def allow_a(tag: str, name: str, value: str):
     if name in ["href", "title", "class"]:
@@ -26,7 +40,7 @@ def sanitize_html(post_html: str) -> str:
             "p": ["class"],
             "span": ["class"],
         },
-        filters=[LinkifyFilter],
+        filters=[partial(LinkifyFilter, url_re=url_regex)],
         strip=True,
     )
     return mark_safe(cleaner.clean(post_html))
@@ -36,7 +50,11 @@ def strip_html(post_html: str) -> str:
     """
     Strips all tags from the text, then linkifies it.
     """
-    cleaner = bleach.Cleaner(tags=[], strip=True, filters=[LinkifyFilter])
+    cleaner = bleach.Cleaner(
+        tags=[],
+        strip=True,
+        filters=[partial(LinkifyFilter, url_re=url_regex)],
+    )
     return mark_safe(cleaner.clean(post_html))
 
 
diff --git a/tests/core/test_html.py b/tests/core/test_html.py
index 5d798ac..ff28305 100644
--- a/tests/core/test_html.py
+++ b/tests/core/test_html.py
@@ -19,3 +19,15 @@ def test_sanitize_post():
 
     assert sanitize_html("<p>Hello!</p>") == "<p>Hello!</p>"
     assert sanitize_html("<p>It&#39;s great</p>") == "<p>It&#39;s great</p>"
+
+    # Note that we only want to linkify things with protocol prefixes to prevent
+    # too many false positives.
+    assert sanitize_html("<p>test.com</p>") == "<p>test.com</p>"
+    assert (
+        sanitize_html("<p>https://test.com</p>")
+        == '<p><a href="https://test.com" rel="nofollow">https://test.com</a></p>'
+    )
+    assert (
+        sanitize_html("<p>@someone@subdomain.some-domain.com</p>")
+        == "<p>@someone@subdomain.some-domain.com</p>"
+    )