docs(macos): clarify repeat-task cleanup releases the key

Address @feschber review feedback on PR #441. The repeat-task cleanup
already releases the key with the correct CGKeyCode via the existing
key_event call at the end of the closure — this commit just expands
the surrounding comment to make that explicit and to document why
update_modifiers is intentionally NOT called from this path (Mac
CGKeyCode vs Linux evdev scancode collision).

No behavioral change.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: 2 people

input-emulation/src/macos.rs

@@ -106,14 +106,19 @@ impl MacOSEmulation {
                     }
                 }
             }
-            // release key when cancelled.
+            // Always release the key with the correct CGKeyCode, regardless of
+            // whether the repeat loop ran. This matches @feschber's review
+            // request: "still release the key repeat task but with the correct
+            // code."
+            //
             // Do NOT call update_modifiers here: `key` is a Mac CGKeyCode but
             // update_modifiers expects a Linux evdev scancode, and the two
             // codespaces collide (e.g. Mac LeftShift=56 == Linux KeyLeftAlt=56,
             // Mac Down=125 == Linux KeyLeftMeta=125), corrupting modifier
             // state for chords like Shift+Option+X or Cmd+Down. Modifier state
-            // is owned by the main consume() loop, which calls update_modifiers
-            // with the correct Linux scancode on the real release event.
+            // is owned by the main consume() loop, which already calls
+            // update_modifiers with the correct Linux scancode on the real key
+            // release event from the client.
             key_event(event_source.clone(), key, 0, modifiers.get());
         });
         self.repeat_task = Some(repeat_task);