From df132627e62dae6ed3208f861b50a4496253c256 Mon Sep 17 00:00:00 2001
From: Aevann <randomname42029@gmail.com>
Date: Fri, 27 Jan 2023 11:11:13 +0200
Subject: [PATCH] fix underscores inside links

---
 files/helpers/regex.py                          | 3 +++
 files/helpers/sanitize.py                       | 3 +++
 migrations/20230127-fix-underscores-in-links.sh | 1 +
 requirements.txt                                | 1 +
 4 files changed, 8 insertions(+)
 create mode 100644 migrations/20230127-fix-underscores-in-links.sh

diff --git a/files/helpers/regex.py b/files/helpers/regex.py
index 2f0fb5f98..668104665 100644
--- a/files/helpers/regex.py
+++ b/files/helpers/regex.py
@@ -1,5 +1,6 @@
 import random
 import re
+import regex
 from random import choice, choices
 from typing import List, Optional, Union
 
@@ -193,3 +194,5 @@ def command_regex_matcher(match, upper=False):
 reason_regex = re.compile('(/(post|comment)/[0-9]+)', flags=re.A)
 
 discord_username_regex = re.compile("(\s|^|>).{2,32}#[0-9]{4}(?=[^0-9]|$)", flags=re.A)
+
+underscore_in_link_regex = regex.compile("(?<=https:\/\/[^\s)]+)_", flags=regex.A)
diff --git a/files/helpers/sanitize.py b/files/helpers/sanitize.py
index 5ea345ffc..d4cb38f50 100644
--- a/files/helpers/sanitize.py
+++ b/files/helpers/sanitize.py
@@ -291,7 +291,10 @@ def sanitize(sanitized, golden=True, limit_pings=0, showmore=True, count_marseys
 	if FEATURES['MARKUP_COMMANDS']:
 		sanitized = command_regex.sub(command_regex_matcher, sanitized)
 
+	sanitized = underscore_in_link_regex.sub(r'\_', sanitized)
+
 	sanitized = markdown(sanitized)
+
 	sanitized = strikethrough_regex.sub(r'\1<del>\2</del>', sanitized)
 
 	# replacing zero width characters, overlines, fake colons
diff --git a/migrations/20230127-fix-underscores-in-links.sh b/migrations/20230127-fix-underscores-in-links.sh
new file mode 100644
index 000000000..ac1702550
--- /dev/null
+++ b/migrations/20230127-fix-underscores-in-links.sh
@@ -0,0 +1 @@
+pip3 install regex
diff --git a/requirements.txt b/requirements.txt
index db1851850..01949bc79 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -23,6 +23,7 @@ pyotp
 pywebpush
 qrcode
 redis
+regex
 requests
 SQLAlchemy
 tinycss2