Fix win32 input mode encoding inconsistencies

This pulls in almost all of the original PR in #2235.
I skipped a dead key case that I recall being tricky:
I didn't want to break the non win32-input mode version
of that.

I'd be happy to have that case re-evaluated in a smaller
PR where we can focus on its details.

Co-authored-by: Dominik Kreutzer <kreudom@gmail.com>

docs/changelog.md

@@ -43,6 +43,7 @@ As features stabilize some brief notes about them will accumulate here.
   integrated titlebar buttons. #3499
 * Windows: closing a window while the debug overlay was active could
   leave a lingering wezterm-gui.exe running. #3522
+* Windows: inconsistencies with win32 input mode. Thanks to @kreudom! #2235
 
 ### 20230408-112425-69ae8472
 

wezterm-input-types/src/lib.rs

@@ -1375,6 +1375,9 @@ pub struct KeyEvent {
 
     /// If triggered from a raw key event, here it is.
     pub raw: Option<RawKeyEvent>,
+
+    #[cfg(windows)]
+    pub win32_uni_char: Option<char>,
 }
 
 fn normalize_shift(key: KeyCode, modifiers: Modifiers) -> (KeyCode, Modifiers) {
@@ -1536,7 +1539,10 @@ impl KeyEvent {
         const LEFT_CTRL_PRESSED: usize = 0x08;
         const RIGHT_CTRL_PRESSED: usize = 0x04;
 
-        if self.modifiers.contains(Modifiers::SHIFT) {
+        if self
+            .modifiers
+            .intersects(Modifiers::SHIFT | Modifiers::LEFT_SHIFT | Modifiers::RIGHT_SHIFT)
+        {
             control_key_state |= SHIFT_PRESSED;
         }
 
@@ -1565,36 +1571,7 @@ impl KeyEvent {
         match &self.key {
             KeyCode::Composed(_) => None,
             KeyCode::Char(c) => {
-                let c = match *c {
+                let uni = self.win32_uni_char.unwrap_or(*c) as u32;
-                    // Delete key is transmitted as 0x0
-                    '\x7f' => '\x00',
-                    // Backspace key is transmitted as 0x8, 0x7f or 0x0
-                    '\x08' => {
-                        if self.modifiers.contains(Modifiers::CTRL) {
-                            if self.modifiers.contains(Modifiers::ALT)
-                                || self.modifiers.contains(Modifiers::SHIFT)
-                            {
-                                '\x00'
-                            } else {
-                                '\x7f'
-                            }
-                        } else {
-                            '\x08'
-                        }
-                    }
-                    _ => *c,
-                };
-
-                let c = if self.modifiers.contains(Modifiers::CTRL) {
-                    // Ensure that we rewrite the unicode value to the ASCII CTRL
-                    // equivalent value.
-                    // <https://github.com/microsoft/terminal/issues/13134>
-                    ctrl_mapping(c).unwrap_or(c)
-                } else {
-                    c
-                };
-                let uni = c as u32;
-
                 Some(format!(
                     "\u{1b}[{};{};{};{};{};{}_",
                     vkey, scan_code, uni, key_down, control_key_state, self.repeat_count
@@ -2176,7 +2153,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "o".to_string()
@@ -2188,7 +2167,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: false,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\x1b[111;1:3u".to_string()
@@ -2209,7 +2190,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\x1b[11;1~".to_string()
@@ -2221,7 +2204,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: false,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\x1b[11;1:3~".to_string()
@@ -2239,7 +2224,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\x1b[105;4u".to_string()
@@ -2251,7 +2238,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\x1b[105;4u".to_string()
@@ -2264,7 +2253,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\x1b[49;4u".to_string()
@@ -2278,7 +2269,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::K1)
             )
@@ -2293,7 +2286,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\x1b[105;6u".to_string()
@@ -2305,7 +2300,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\x1b[105;6u".to_string()
@@ -2330,6 +2327,8 @@ mod test {
                     scan_code: 0,
                     repeat_count: 1,
                 }),
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\x1b[105;6u".to_string()
@@ -2342,7 +2341,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\x1b[105;8u".to_string()
@@ -2354,7 +2355,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\x1b[105;8u".to_string()
@@ -2375,7 +2378,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\u{1b}[97:65;1u".to_string()
@@ -2387,7 +2392,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: false,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\u{1b}[97:65;1:3u".to_string()
@@ -2431,7 +2438,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 None
             )
@@ -2446,7 +2455,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: false,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 None
             )
@@ -2461,7 +2472,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 None
             )
@@ -2476,7 +2489,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: false,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 None
             )
@@ -2500,7 +2515,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 None
             )
@@ -2515,7 +2532,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 None
             )
@@ -2531,7 +2550,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 None
             )
@@ -2546,7 +2567,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 None
             )
@@ -2562,7 +2585,9 @@ mod test {
                     leds: KeyboardLedStatus::NUM_LOCK,
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::Keypad0)
             )
@@ -2577,7 +2602,9 @@ mod test {
                     leds: KeyboardLedStatus::NUM_LOCK,
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::Keypad0)
             )
@@ -2593,7 +2620,9 @@ mod test {
                     leds: KeyboardLedStatus::NUM_LOCK,
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::Keypad5)
             )
@@ -2609,7 +2638,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::Keypad5)
             )
@@ -2634,7 +2665,9 @@ mod test {
                     leds: KeyboardLedStatus::NUM_LOCK,
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::Keypad5)
             )
@@ -2649,7 +2682,9 @@ mod test {
                     leds: KeyboardLedStatus::NUM_LOCK,
                     repeat_count: 1,
                     key_is_down: false,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::Keypad5)
             )
@@ -2665,7 +2700,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::Keypad5)
             )
@@ -2681,7 +2718,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: false,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::Keypad5)
             )
@@ -2701,7 +2740,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\"".to_string()
@@ -2714,7 +2755,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "\"".to_string()
@@ -2727,7 +2770,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "!".to_string()
@@ -2740,7 +2785,9 @@ mod test {
                 leds: KeyboardLedStatus::empty(),
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             "".to_string()
@@ -2762,7 +2809,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::A)
             )
@@ -2778,7 +2827,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::A)
             )
@@ -2803,7 +2854,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::A)
             )
@@ -2819,7 +2872,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::A)
             )
@@ -2839,7 +2894,9 @@ mod test {
                 leds: KeyboardLedStatus::NUM_LOCK,
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             " ".to_string()
@@ -2852,7 +2909,9 @@ mod test {
                 leds: KeyboardLedStatus::CAPS_LOCK,
                 repeat_count: 1,
                 key_is_down: true,
-                raw: None
+                raw: None,
+                #[cfg(windows)]
+                win32_uni_char: None,
             }
             .encode_kitty(flags),
             " ".to_string()
@@ -2866,7 +2925,9 @@ mod test {
                     leds: KeyboardLedStatus::empty(),
                     repeat_count: 1,
                     key_is_down: true,
-                    raw: None
+                    raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::NumLock)
             )
@@ -2883,6 +2944,8 @@ mod test {
                     repeat_count: 1,
                     key_is_down: true,
                     raw: None,
+                    #[cfg(windows)]
+                    win32_uni_char: None,
                 },
                 Some(PhysKeyCode::CapsLock)
             )

window/src/os/windows/window.rs

@@ -1964,6 +1964,7 @@ unsafe fn ime_composition(
                 repeat_count: 1,
                 key_is_down: true,
                 raw: None,
+                win32_uni_char: None,
             };
             inner
                 .events
@@ -2349,8 +2350,11 @@ unsafe fn key(hwnd: HWND, msg: UINT, wparam: WPARAM, lparam: LPARAM) -> Option<L
         return Some(0);
     }
 
+    let keys = {
         let mut keys = [0u8; 256];
         GetKeyboardState(keys.as_mut_ptr());
+        keys
+    };
 
     let mut modifiers = Modifiers::default();
     if keys[VK_SHIFT as usize] & 0x80 != 0 {
@@ -2410,18 +2414,6 @@ unsafe fn key(hwnd: HWND, msg: UINT, wparam: WPARAM, lparam: LPARAM) -> Option<L
         modifiers |= Modifiers::SUPER;
     }
 
-    // If control is pressed, clear that out and remember it in our
-    // own set of modifiers.
-    // We used to also remove shift from this set, but it impacts
-    // handling of eg: ctrl+shift+' (which is equivalent to ctrl+" in a US English
-    // layout.
-    // The shift normalization is now handled by the normalize_shift() method.
-    if modifiers.contains(Modifiers::CTRL) {
-        keys[VK_CONTROL as usize] = 0;
-        keys[VK_LCONTROL as usize] = 0;
-        keys[VK_RCONTROL as usize] = 0;
-    }
-
     let mut leds = KeyboardLedStatus::empty();
     if keys[VK_CAPITAL as usize] & 1 != 0 {
         leds |= KeyboardLedStatus::CAPS_LOCK;
@@ -2446,11 +2438,14 @@ unsafe fn key(hwnd: HWND, msg: UINT, wparam: WPARAM, lparam: LPARAM) -> Option<L
         handled: handled_raw.clone(),
     };
 
-    let key = if msg == WM_IME_CHAR || msg == WM_CHAR {
+    let (key, win32_uni_char) = if msg == WM_IME_CHAR || msg == WM_CHAR {
         // If we were sent a character by the IME, some other apps,
         // or by ourselves via TranslateMessage, then take that
         // value as-is.
-        Some(KeyCode::Char(std::char::from_u32_unchecked(wparam as u32)))
+        (
+            Some(KeyCode::Char(std::char::from_u32_unchecked(wparam as u32))),
+            None,
+        )
     } else {
         // Otherwise we're dealing with a raw key message.
         // ToUnicode has frustrating statefulness so we take care to
@@ -2475,7 +2470,8 @@ unsafe fn key(hwnd: HWND, msg: UINT, wparam: WPARAM, lparam: LPARAM) -> Option<L
             // If this event is only modifiers then don't ask the system
             // for further resolution, as we don't want ToUnicode to
             // perturb its inscrutable global state.
-            phys_code.map(|p| p.to_key_code())
+            // Modifier-only keypresses are reported as NUL when using win32 input mode.
+            (phys_code.map(|p| p.to_key_code()), Some('\x00'))
         } else {
             // If we think this might be a dead key, process it for ourselves.
             // Our KeyboardLayoutInfo struct probed the layout for the key
@@ -2514,7 +2510,11 @@ unsafe fn key(hwnd: HWND, msg: UINT, wparam: WPARAM, lparam: LPARAM) -> Option<L
                             leds,
                             repeat_count: 1,
                             key_is_down: !releasing,
-                            raw: None,
+                            win32_uni_char: Some(c),
+                            raw: Some(RawKeyEvent {
+                                scan_code: 0,
+                                ..raw_key_event.clone()
+                            }),
                         }
                         .normalize_shift()
                         .resurface_positional_modifier_key()
@@ -2580,7 +2580,7 @@ unsafe fn key(hwnd: HWND, msg: UINT, wparam: WPARAM, lparam: LPARAM) -> Option<L
             };
 
             if dead.is_some() {
-                dead
+                (dead, None)
             } else {
                 // We get here for the various UP (but not DOWN as we shortcircuit
                 // those above) messages.
@@ -2588,6 +2588,8 @@ unsafe fn key(hwnd: HWND, msg: UINT, wparam: WPARAM, lparam: LPARAM) -> Option<L
                 // rather than calling TranslateMessage to do it for us,
                 // so that we have tighter control over the key processing.
                 let mut out = [0u16; 16];
+
+                let win32_uni_char = {
                     let res = ToUnicode(
                         wparam as u32,
                         scan_code as u32,
@@ -2598,6 +2600,35 @@ unsafe fn key(hwnd: HWND, msg: UINT, wparam: WPARAM, lparam: LPARAM) -> Option<L
                     );
 
                     match res {
+                        1 => Some(std::char::from_u32_unchecked(out[0] as u32)),
+                        0 => Some('\x00'),
+                        _ => None,
+                    }
+                };
+
+                let mut keys = keys;
+                // If control is pressed, clear that out and remember it in our
+                // own set of modifiers.
+                // We used to also remove shift from this set, but it impacts
+                // handling of eg: ctrl+shift+' (which is equivalent to ctrl+" in a US English
+                // layout.
+                // The shift normalization is now handled by the normalize_shift() method.
+                if modifiers.contains(Modifiers::CTRL) {
+                    keys[VK_CONTROL as usize] = 0;
+                    keys[VK_LCONTROL as usize] = 0;
+                    keys[VK_RCONTROL as usize] = 0;
+                }
+
+                let res = ToUnicode(
+                    wparam as u32,
+                    scan_code as u32,
+                    keys.as_ptr(),
+                    out.as_mut_ptr(),
+                    out.len() as i32,
+                    0,
+                );
+
+                let key = match res {
                     1 => Some(KeyCode::Char(std::char::from_u32_unchecked(out[0] as u32))),
                     // No mapping, so use our raw info
                     0 => {
@@ -2628,7 +2659,9 @@ unsafe fn key(hwnd: HWND, msg: UINT, wparam: WPARAM, lparam: LPARAM) -> Option<L
                         KeyboardLayoutInfo::clear_key_state();
                         None
                     }
-                }
+                };
+
+                (key, win32_uni_char)
             }
         }
     };
@@ -2644,6 +2677,7 @@ unsafe fn key(hwnd: HWND, msg: UINT, wparam: WPARAM, lparam: LPARAM) -> Option<L
             leds,
             repeat_count: repeat,
             key_is_down: !releasing,
+            win32_uni_char,
             raw: Some(raw_key_event),
         }
         .normalize_shift();