Skip to content

Worktrunk is a CLI for Git worktree management, designed for parallel AI agent workflows

fix: use quiescence drain for PTY prompt to prevent echo interleaving (#1133)

worktrunk-botd51527df654c
diff --git a/tests/common/pty.rs b/tests/common/pty.rs
index 47ee05824..399a79623 100644
--- a/tests/common/pty.rs
+++ b/tests/common/pty.rs
@@ -333,6 +333,28 @@ fn prompted_pty_interaction(
std::thread::sleep(poll);
}
+ // Quiescence drain: after detecting the marker, wait until the PTY
+ // goes quiet before sending input. Without this, trailing prompt bytes
+ // (ANSI resets, spaces) that arrive in a separate read chunk interleave
+ // with the echo of our input, producing non-deterministic output on macOS.
+ let quiescence = Duration::from_millis(20);
+ let drain_ceiling = Duration::from_millis(500);
+ let drain_start = Instant::now();
+ let mut last_data = Instant::now();
+ loop {
+ while let Ok(chunk) = rx.try_recv() {
+ accumulated.extend_from_slice(&chunk);
+ last_data = Instant::now();
+ }
+ if last_data.elapsed() >= quiescence {
+ break;
+ }
+ if drain_start.elapsed() >= drain_ceiling {
+ break;
+ }
+ std::thread::sleep(poll);
+ }
+
writer.write_all(input.as_bytes()).unwrap();
writer.flush().unwrap();
}