From 306f9e57355c39ef8f5d87ce547cbd347ca9b80d Mon Sep 17 00:00:00 2001
From: Kovid Goyal <kovid@kovidgoyal.net>
Date: Thu, 29 Dec 2022 17:04:02 +0530
Subject: [PATCH] Prettify URLs before displaying them to the user

---
 kitty/boss.py            |  8 +++-----
 kitty/utils.py           | 15 +++++++++++++++
 kitty_tests/datatypes.py |  4 +++-
 3 files changed, 21 insertions(+), 6 deletions(-)

diff --git a/kitty/boss.py b/kitty/boss.py
index b87b694f2..6017a24d6 100644
--- a/kitty/boss.py
+++ b/kitty/boss.py
@@ -8,7 +8,7 @@ import os
 import re
 import sys
 from contextlib import suppress
-from functools import lru_cache, partial
+from functools import partial
 from gettext import gettext as _
 from time import monotonic, sleep
 from typing import (
@@ -64,7 +64,7 @@ from .utils import (
     cleanup_ssh_control_masters, func_name, get_editor, get_new_os_window_size,
     is_path_in_temp_dir, less_version, log_error, macos_version, open_url,
     parse_address_spec, parse_uri_list, platform_window_id, remove_socket_file,
-    safe_print, single_instance, startup_notification_handler, which,
+    safe_print, single_instance, startup_notification_handler, which, sanitize_url_for_dispay_to_user
 )
 from .window import CommandOutput, CwdRequest, Window
 
@@ -2603,7 +2603,5 @@ class Boss:
         pass
     mouse_discard_event = discard_event
 
-    @lru_cache(maxsize=64)
     def sanitize_url_for_dispay_to_user(self, url: str) -> str:
-        # TODO: Use punycode, remove percent encoding, etc.
-        return url
+        return sanitize_url_for_dispay_to_user(url)
diff --git a/kitty/utils.py b/kitty/utils.py
index 48e14dff8..dc568be54 100644
--- a/kitty/utils.py
+++ b/kitty/utils.py
@@ -1088,3 +1088,18 @@ def sanitize_for_bracketed_paste(text: bytes) -> bytes:
             break
         text = new_text
     return text
+
+
+@lru_cache(maxsize=64)
+def sanitize_url_for_dispay_to_user(url: str) -> str:
+    from urllib.parse import urlparse, urlunparse, unquote
+    try:
+        purl = urlparse(url)
+        if purl.netloc:
+            purl = purl._replace(netloc=purl.netloc.encode('idna').decode('ascii'))
+        if purl.path:
+            purl = purl._replace(path=unquote(purl.path))
+        url = urlunparse(purl)
+    except Exception:
+        url = 'Unpareseable URL: ' + url
+    return url
diff --git a/kitty_tests/datatypes.py b/kitty_tests/datatypes.py
index 3061883c4..9b6f12e81 100644
--- a/kitty_tests/datatypes.py
+++ b/kitty_tests/datatypes.py
@@ -12,7 +12,7 @@ from kitty.fast_data_types import (
     wcwidth, expand_ansi_c_escapes
 )
 from kitty.rgb import to_color
-from kitty.utils import is_path_in_temp_dir, sanitize_title
+from kitty.utils import is_path_in_temp_dir, sanitize_title, sanitize_url_for_dispay_to_user
 
 from . import BaseTest, filled_cursor, filled_history_buf, filled_line_buf
 
@@ -436,6 +436,8 @@ class TestDataTypes(BaseTest):
                 self.assertTrue(is_path_in_temp_dir(os.path.join(prefix, path)))
         for path in ('/home/xy/d.png', '/tmp/../home/x.jpg'):
             self.assertFalse(is_path_in_temp_dir(os.path.join(path)))
+        self.ae(sanitize_url_for_dispay_to_user(
+            'h://a\u0430b.com/El%20Ni%C3%B1o/'), 'h://xn--ab-7kc.com/El Niño/')
 
     def test_color_profile(self):
         c = ColorProfile()