From 219e53826b1764f977dd9fe60aea903f79ed66c1 Mon Sep 17 00:00:00 2001
From: Kovid Goyal <kovid@kovidgoyal.net>
Date: Thu, 9 May 2024 12:03:02 +0530
Subject: [PATCH] More efficient encoding for cmdline in the prompt marking
 escape code

---
 docs/shell-integration.rst              | 8 ++++----
 kitty/window.py                         | 3 ++-
 kitty_tests/__init__.py                 | 3 ++-
 shell-integration/bash/kitty.bash       | 4 +---
 shell-integration/zsh/kitty-integration | 4 +---
 5 files changed, 10 insertions(+), 12 deletions(-)

diff --git a/docs/shell-integration.rst b/docs/shell-integration.rst
index 4043d7373..bb4ac115a 100644
--- a/docs/shell-integration.rst
+++ b/docs/shell-integration.rst
@@ -458,10 +458,10 @@ to control its behavior, separated by semi-colons. They are::
 
 kitty also optionally supports sending the cmdline going to be executed with ``<OSC>133;C`` as::
 
-    <OSC>133;C;cmdline=cmdline as space separated hex encoded text<ST>
+    <OSC>133;C;cmdline=cmdline encoded by %q<ST>
     or
-    <OSC>133;C;cmdline_url=cmdline as UTF-8 URL escaped text<ST>
+    <OSC>133;C;cmdline_url=cmdline as UTF-8 URL %-escaped text<ST>
 
 
-Here, *space separated hex encoded text* means every unicode codepoint of the
-command line is encoded as 2-8 hex digits separated by spaces.
+Here, *encoded by %q* means the encoding produced by the %q format to printf in
+bash and similar shells.
diff --git a/kitty/window.py b/kitty/window.py
index 2eabf6306..42c54851e 100644
--- a/kitty/window.py
+++ b/kitty/window.py
@@ -115,6 +115,7 @@ from .utils import (
     sanitize_for_bracketed_paste,
     sanitize_title,
     sanitize_url_for_dispay_to_user,
+    shlex_split,
 )
 
 MatchPatternType = Union[Pattern[str], Tuple[Pattern[str], Optional[Pattern[str]]]]
@@ -220,7 +221,7 @@ def compile_match_query(exp: str, is_simple: bool = True) -> MatchPatternType:
 def decode_cmdline(x: str) -> str:
     ctype, sep, val = x.partition('=')
     if ctype == 'cmdline':
-        return ''.join(chr(int(x, 16)) for x in val.split())
+        return next(shlex_split(val, True))
     if ctype == 'cmdline_url':
         from urllib.parse import unquote
         return unquote(val)
diff --git a/kitty_tests/__init__.py b/kitty_tests/__init__.py
index 74ba5e600..40ffb17d0 100644
--- a/kitty_tests/__init__.py
+++ b/kitty_tests/__init__.py
@@ -15,6 +15,7 @@ import time
 from contextlib import contextmanager
 from functools import wraps
 from pty import CHILD, STDIN_FILENO, STDOUT_FILENO, fork
+from typing import Optional
 from unittest import TestCase
 
 from kitty.config import finalize_keys, finalize_mouse_mappings
@@ -64,7 +65,7 @@ class Callbacks:
     def color_profile_popped(self, x) -> None:
         pass
 
-    def cmd_output_marking(self, is_start: bool) -> None:
+    def cmd_output_marking(self, is_start: Optional[bool], data: str = '') -> None:
         pass
 
     def request_capabilities(self, q) -> None:
diff --git a/shell-integration/bash/kitty.bash b/shell-integration/bash/kitty.bash
index e45c96d96..9ab615429 100644
--- a/shell-integration/bash/kitty.bash
+++ b/shell-integration/bash/kitty.bash
@@ -204,9 +204,7 @@ _ksi_main() {
             builtin printf "\e]2;%s%s\a" "${_ksi_prompt[hostname_prefix]@P}" "${_ksi_prompt[last_cmd]//[[:cntrl:]]}"  # remove any control characters
         fi
         if [[ "${_ksi_prompt[mark]}" == "y" ]]; then
-            builtin printf "\e]133;C;cmdline="
-            for (( i=0; i<${#last_cmd}; i++ )); do builtin printf '%x ' "'${last_cmd:$i:1}"; done
-            builtin printf "\a"
+            builtin printf "\e]133;C;cmdline=%q\a" "$last_cmd"
         fi
     }
     if [[ "${_ksi_prompt[title]}" == "y" ||  "${_ksi_prompt[mark]}" ]]; then _ksi_prompt[ps0]+='$(_ksi_get_current_command)'; fi
diff --git a/shell-integration/zsh/kitty-integration b/shell-integration/zsh/kitty-integration
index 0905cdf1a..76459d478 100644
--- a/shell-integration/zsh/kitty-integration
+++ b/shell-integration/zsh/kitty-integration
@@ -215,9 +215,7 @@ _ksi_deferred_init() {
             # its preexec hook before us, we'll incorrectly mark its output as
             # belonging to the command (as if the user typed it into zle) rather
             # than command output.
-            builtin print -nu "$_ksi_fd" '\e]133;C;cmdline='
-            for (( i=0; i<${#1}; i++ )); do builtin print -nu "$_ksi_fd" -f '%x ' "'${1:$i:1}"; done
-            builtin print -nu "$_ksi_fd" '\a'
+            builtin print -nu "$_ksi_fd" -f '\e]133;C;cmdline=%q\a' "$1"
             (( _ksi_state = 1 ))
         }