Skip to content

feat(tui): show live background shell output #2826

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Hmbown merged 1 commit into from
Jun 6, 2026
Merged

feat(tui): show live background shell output #2826

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,6 +66,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
visibility trail, and @BigBenLabs, @lzx1545642258, @yangdaowan,
@mangdehuang, @VerrPower, @hejia-v, @nasus9527, and @ygzhang-cn for the
GUI/VS Code demand and validation trail.
- Added inline live-output refresh for background shell Exec cards keyed by the
exact shell task id, so long-running commands can show bounded stdout/stderr
tails without consuming deltas or matching by command text. Thanks
@donglovejava for the live shell-output direction in #2048.
- Added a static prompt composer override for embedders that need to replace
the byte-stable base/personality prompt segment while leaving mode metadata,
approval policy, tool taxonomy, Context Management, and the Compaction Relay
Expand Down
4 changes: 4 additions & 0 deletions crates/tui/CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,6 +66,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
visibility trail, and @BigBenLabs, @lzx1545642258, @yangdaowan,
@mangdehuang, @VerrPower, @hejia-v, @nasus9527, and @ygzhang-cn for the
GUI/VS Code demand and validation trail.
- Added inline live-output refresh for background shell Exec cards keyed by the
exact shell task id, so long-running commands can show bounded stdout/stderr
tails without consuming deltas or matching by command text. Thanks
@donglovejava for the live shell-output direction in #2048.
- Added a static prompt composer override for embedders that need to replace
the byte-stable base/personality prompt segment while leaving mode metadata,
approval policy, tool taxonomy, Context Management, and the Compaction Relay
Expand Down
2 changes: 2 additions & 0 deletions crates/tui/src/tui/active_cell.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -331,6 +331,8 @@ mod tests {
command: command.to_string(),
status: ToolStatus::Running,
output: None,
live_output: None,
shell_task_id: None,
started_at: Some(Instant::now()),
duration_ms: None,
source: ExecSource::Assistant,
Expand Down
59 changes: 58 additions & 1 deletion crates/tui/src/tui/history.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -890,6 +890,8 @@ pub struct ExecCell {
pub command: String,
pub status: ToolStatus,
pub output: Option<String>,
pub live_output: Option<String>,
pub shell_task_id: Option<String>,
pub started_at: Option<Instant>,
pub duration_ms: Option<u64>,
pub source: ExecSource,
Expand Down Expand Up @@ -946,7 +948,7 @@ impl ExecCell {
}

if self.interaction.is_none() {
if let Some(output) = self.output.as_ref() {
if let Some(output) = self.output.as_ref().or(self.live_output.as_ref()) {
lines.extend(render_exec_output_mode(
output,
width,
Expand Down Expand Up @@ -4306,6 +4308,8 @@ mod tests {
command: "echo hi".to_string(),
status: ToolStatus::Running,
output: None,
live_output: None,
shell_task_id: None,
started_at,
duration_ms: None,
source: ExecSource::Assistant,
Expand Down Expand Up @@ -4645,6 +4649,8 @@ mod tests {
command: "ls".to_string(),
status: ToolStatus::Success,
output: Some("a\nb\n".to_string()),
live_output: None,
shell_task_id: None,
started_at: None,
duration_ms: Some(10),
source: ExecSource::Assistant,
Expand Down Expand Up @@ -4675,6 +4681,8 @@ mod tests {
command: "cargo test --workspace --all-features".to_string(),
status: ToolStatus::Running,
output: None,
live_output: None,
shell_task_id: None,
started_at: None,
duration_ms: None,
source: ExecSource::Assistant,
Expand Down Expand Up @@ -4991,6 +4999,8 @@ mod tests {
command: "false".to_string(),
status: ToolStatus::Failed,
output: Some("boom".to_string()),
live_output: None,
shell_task_id: None,
started_at: None,
duration_ms: Some(42),
source: ExecSource::Assistant,
Expand Down Expand Up @@ -5046,6 +5056,49 @@ mod tests {
lines.iter().map(line_text).collect::<Vec<_>>().join("\n")
}

#[test]
fn exec_cell_renders_live_shell_output_before_final_output() {
let cell = ExecCell {
command: "cargo test".to_string(),
status: ToolStatus::Running,
output: None,
live_output: Some("running line 1\nrunning line 2".to_string()),
shell_task_id: Some("shell_live".to_string()),
started_at: None,
duration_ms: None,
source: ExecSource::Assistant,
interaction: None,
output_summary: None,
};

let text = lines_text(&cell.lines_with_motion(80, true));

assert!(text.contains("running line 1"));
assert!(text.contains("running line 2"));
assert!(!text.contains("Ctrl+B opens shell controls"));
}

#[test]
fn exec_cell_prefers_final_output_over_live_shell_tail() {
let cell = ExecCell {
command: "cargo test".to_string(),
status: ToolStatus::Success,
output: Some("final output".to_string()),
live_output: Some("stale live tail".to_string()),
shell_task_id: Some("shell_live".to_string()),
started_at: None,
duration_ms: None,
source: ExecSource::Assistant,
interaction: None,
output_summary: None,
};

let text = lines_text(&cell.lines_with_motion(80, true));

assert!(text.contains("final output"));
assert!(!text.contains("stale live tail"));
}

#[test]
fn long_thinking_display_is_shorter_than_transcript() {
// Build a multi-paragraph thinking body so the live view has
Expand Down Expand Up @@ -5162,6 +5215,8 @@ mod tests {
command: "noisy_script.sh".to_string(),
status: ToolStatus::Success,
output: Some(output),
live_output: None,
shell_task_id: None,
started_at: None,
duration_ms: Some(120),
source: ExecSource::Assistant,
Expand Down Expand Up @@ -5613,6 +5668,8 @@ mod tests {
command: command.to_string(),
status: ToolStatus::Success,
output: Some("ok".to_string()),
live_output: None,
shell_task_id: None,
started_at: None,
duration_ms: None,
source: ExecSource::Assistant,
Expand Down
10 changes: 10 additions & 0 deletions crates/tui/src/tui/sidebar.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2874,6 +2874,8 @@ mod tests {
command: command.to_string(),
status: ToolStatus::Running,
output: None,
live_output: None,
shell_task_id: None,
started_at: None,
duration_ms: Some(ACTIVE_TOOL_STALE_RUNNING_ROW_TTL.as_millis() as u64 + 1),
source: ExecSource::Assistant,
Expand Down Expand Up @@ -2906,6 +2908,8 @@ mod tests {
command: "cargo test --workspace".to_string(),
status: ToolStatus::Running,
output: None,
live_output: None,
shell_task_id: None,
started_at: Some(std::time::Instant::now()),
duration_ms: None,
source: ExecSource::Assistant,
Expand Down Expand Up @@ -3040,6 +3044,8 @@ mod tests {
.to_string(),
status: ToolStatus::Failed,
output: Some("Lint pending\nTest pending".to_string()),
live_output: None,
shell_task_id: None,
started_at: None,
duration_ms: Some(15_000),
source: ExecSource::Assistant,
Expand Down Expand Up @@ -3080,6 +3086,8 @@ mod tests {
command: "cargo test -p codewhale-tui".to_string(),
status: ToolStatus::Failed,
output: Some("test failed".to_string()),
live_output: None,
shell_task_id: None,
started_at: None,
duration_ms: Some(1_250),
source: ExecSource::Assistant,
Expand Down Expand Up @@ -3109,6 +3117,8 @@ mod tests {
command: "cargo check".to_string(),
status: ToolStatus::Success,
output: Some("Finished".to_string()),
live_output: None,
shell_task_id: None,
started_at: None,
duration_ms: Some(1_250),
source: ExecSource::Assistant,
Expand Down
20 changes: 20 additions & 0 deletions crates/tui/src/tui/tool_routing.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -101,6 +101,8 @@ pub(super) fn handle_tool_call_started(
command,
status: ToolStatus::Running,
output: None,
live_output: None,
shell_task_id: None,
started_at: Some(Instant::now()),
duration_ms: None,
source,
Expand Down Expand Up @@ -133,6 +135,8 @@ pub(super) fn handle_tool_call_started(
command,
status: ToolStatus::Running,
output: None,
live_output: None,
shell_task_id: None,
started_at: Some(Instant::now()),
duration_ms: None,
source,
Expand Down Expand Up @@ -506,6 +510,16 @@ pub(super) fn handle_tool_call_complete(
HistoryCell::Tool(ToolCell::Exec(exec)) => {
exec.status = status;
if let Ok(tool_result) = result.as_ref() {
let shell_task_id = tool_result
.metadata
.as_ref()
.and_then(|m| m.get("task_id"))
.and_then(serde_json::Value::as_str)
.filter(|task_id| !task_id.trim().is_empty())
.map(str::to_string);
if shell_task_id.is_some() {
exec.shell_task_id = shell_task_id;
}
if let Some(meta_command) = tool_result
.metadata
.as_ref()
Expand Down Expand Up @@ -538,6 +552,12 @@ pub(super) fn handle_tool_call_complete(
exec.output = Some(tool_result.content.clone());
exec.output_summary =
Some(super::history::summarize_tool_output(&tool_result.content));
exec.live_output = None;
} else if status == ToolStatus::Running
&& exec.interaction.is_none()
&& !tool_result.content.is_empty()
{
exec.live_output = Some(tool_result.content.clone());
}
} else if let Err(err) = result.as_ref()
&& exec.interaction.is_none()
Expand Down
2 changes: 2 additions & 0 deletions crates/tui/src/tui/transcript.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -616,6 +616,8 @@ mod tests {
command: command.to_string(),
status: ToolStatus::Running,
output: None,
live_output: None,
shell_task_id: None,
started_at: None,
duration_ms: None,
source: ExecSource::Assistant,
Expand Down
87 changes: 87 additions & 0 deletions crates/tui/src/tui/ui.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -65,6 +65,7 @@ use crate::settings::Settings;
use crate::task_manager::{
NewTaskRequest, SharedTaskManager, TaskManager, TaskManagerConfig, TaskStatus, TaskSummary,
};
use crate::tools::shell::{ShellJobSnapshot, ShellStatus};
use crate::tools::spec::{RuntimeToolServices, ToolResult};
use crate::tools::subagent::SubAgentStatus;
use crate::tui::app::HuntVerdict;
Expand Down Expand Up @@ -1014,6 +1015,89 @@ async fn refresh_active_task_panel(app: &mut App, task_manager: &SharedTaskManag
app.task_panel = entries;
}

fn refresh_shell_exec_live_output(app: &mut App) -> bool {
let Some(shell_mgr) = app.runtime_services.shell_manager.as_ref().cloned() else {
return false;
};
let jobs = {
let Ok(mut mgr) = shell_mgr.lock() else {
return false;
};
mgr.list_jobs()
.into_iter()
.map(|job| (job.id.clone(), job))
.collect::<std::collections::HashMap<_, _>>()
};
if jobs.is_empty() {
return false;
}

let mut changed = false;
for index in 0..app.virtual_cell_count() {
let Some((task_id, next_status, next_live, next_duration)) =
shell_exec_live_update(app, index, &jobs)
else {
continue;
};
let Some(HistoryCell::Tool(ToolCell::Exec(exec))) = app.cell_at_virtual_index_mut(index)
else {
continue;
};
if exec.output.is_some() || exec.shell_task_id.as_deref() != Some(task_id.as_str()) {
continue;
}
exec.status = next_status;
exec.live_output = next_live;
exec.duration_ms = Some(next_duration);
changed = true;
}
changed
}

fn shell_exec_live_update(
app: &App,
index: usize,
jobs: &std::collections::HashMap<String, ShellJobSnapshot>,
) -> Option<(String, ToolStatus, Option<String>, u64)> {
let HistoryCell::Tool(ToolCell::Exec(exec)) = app.cell_at_virtual_index(index)? else {
return None;
};
if exec.output.is_some() {
return None;
}
let task_id = exec.shell_task_id.as_deref()?;
let job = jobs.get(task_id)?;
let next_status = shell_job_tool_status(&job.status);
let next_live = shell_job_live_output(job).or_else(|| exec.live_output.clone());
if exec.status == next_status
&& exec.live_output == next_live
&& exec.duration_ms == Some(job.elapsed_ms)
{
return None;
}
Some((task_id.to_string(), next_status, next_live, job.elapsed_ms))
}

fn shell_job_tool_status(status: &ShellStatus) -> ToolStatus {
match status {
ShellStatus::Running => ToolStatus::Running,
ShellStatus::Completed => ToolStatus::Success,
ShellStatus::Failed | ShellStatus::Killed | ShellStatus::TimedOut => ToolStatus::Failed,
}
}

fn shell_job_live_output(job: &ShellJobSnapshot) -> Option<String> {
match (job.stdout_tail.is_empty(), job.stderr_tail.is_empty()) {
(true, true) => None,
(false, true) => Some(job.stdout_tail.clone()),
(true, false) => Some(format!("STDERR:\n{}", job.stderr_tail)),
(false, false) => Some(format!(
"{}\n\nSTDERR:\n{}",
job.stdout_tail, job.stderr_tail
)),
}
}

fn active_reasoning_task_entries(app: &App) -> Vec<TaskPanelEntry> {
let Some(active) = app.active_cell.as_ref() else {
return Vec::new();
Expand Down Expand Up @@ -1301,6 +1385,9 @@ async fn run_event_loop(

if last_task_refresh.elapsed() >= Duration::from_millis(2500) {
refresh_active_task_panel(app, &task_manager).await;
if refresh_shell_exec_live_output(app) {
app.needs_redraw = true;
}
last_task_refresh = Instant::now();
app.needs_redraw = true;
}
Expand Down
Loading
Loading