openai/codex
1
0
Fork 0
mirror of https://github.com/openai/codex.git synced 2026-10-30 19:18:49 +00:00
Code Issues Packages Projects Releases Wiki Activity

tui: resume queued drafts after inline app events

Browse Source 
Remove the extra app-side popup completion latch so queued replay resumes
whenever app events fully drain and ChatWidget is waiting to continue.
Add an app-level regression for queued /personality followed by a
queued user draft.

Co-authored-by: Codex <noreply@openai.com>
This commit is contained in:
Charles Cunningham
2026-03-15 23:36:33 -07:00
co-authored by Codex
parent 9ec1452655
commit d9325105a1
1 changed files with 101 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files
codex-rs/tui/src/app.rs
+101 -13
View File
@@ -730,7 +730,6 @@ pub(crate) struct App {
primary_thread_id: Option<ThreadId>,
primary_session_configured: Option<SessionConfiguredEvent>,
pending_primary_events: VecDeque<Event>,
resume_queued_inputs_after_app_events: bool,
}
#[derive(Default)]
@@ -758,16 +757,11 @@ fn normalize_harness_overrides_for_cwd(
}
impl App {
fn on_bottom_pane_view_completed(&mut self) {
self.resume_queued_inputs_after_app_events = true;
}
fn maybe_resume_queued_inputs_after_app_events(&mut self, app_events_drained: bool) {
if !self.resume_queued_inputs_after_app_events || !app_events_drained {
if !app_events_drained {
return;
}
self.resume_queued_inputs_after_app_events = false;
self.chat_widget.maybe_resume_queued_inputs_when_idle();
}
@@ -2307,7 +2301,6 @@ impl App {
primary_thread_id: None,
primary_session_configured: None,
pending_primary_events: VecDeque::new(),
resume_queued_inputs_after_app_events: false,
};
// On startup, if Agent mode (workspace-write) or ReadOnly is active, warn about world-writable dirs on Windows.
@@ -2785,9 +2778,7 @@ impl App {
self.chat_widget.handle_serialized_slash_command(draft);
self.refresh_status_line();
}
AppEvent::BottomPaneViewCompleted => {
self.on_bottom_pane_view_completed();
}
AppEvent::BottomPaneViewCompleted => {}
AppEvent::UpdateCollaborationMode(mask) => {
self.chat_widget.set_collaboration_mask(mask);
self.refresh_status_line();
@@ -6564,7 +6555,6 @@ guardian_approval = true
primary_thread_id: None,
primary_session_configured: None,
pending_primary_events: VecDeque::new(),
resume_queued_inputs_after_app_events: false,
}
}
@@ -6625,7 +6615,6 @@ guardian_approval = true
primary_thread_id: None,
primary_session_configured: None,
pending_primary_events: VecDeque::new(),
resume_queued_inputs_after_app_events: false,
},
rx,
op_rx,
@@ -7640,6 +7629,105 @@ guardian_approval = true
}
}
#[tokio::test]
async fn queued_personality_selection_resumes_followup_after_app_events_drain() {
let (mut app, mut app_event_rx, mut op_rx) = make_test_app_with_channels().await;
app.chat_widget.set_model("gpt-5.2-codex");
app.chat_widget
.set_feature_enabled(Feature::Personality, true);
app.chat_widget.handle_codex_event(Event {
id: "configured".into(),
msg: EventMsg::SessionConfigured(SessionConfiguredEvent {
session_id: ThreadId::new(),
forked_from_id: None,
thread_name: None,
model: "gpt-5.2-codex".to_string(),
model_provider_id: "test-provider".to_string(),
service_tier: None,
approval_policy: AskForApproval::Never,
approvals_reviewer: ApprovalsReviewer::User,
sandbox_policy: SandboxPolicy::new_read_only_policy(),
cwd: PathBuf::from("/home/user/project"),
reasoning_effort: Some(ReasoningEffortConfig::default()),
history_log_id: 0,
history_entry_count: 0,
initial_messages: None,
network_proxy: None,
rollout_path: None,
}),
});
app.chat_widget.handle_codex_event(Event {
id: "turn-started".into(),
msg: EventMsg::TurnStarted(TurnStartedEvent {
turn_id: "turn-1".to_string(),
model_context_window: None,
collaboration_mode_kind: Default::default(),
}),
});
while app_event_rx.try_recv().is_ok() {}
while op_rx.try_recv().is_ok() {}
app.chat_widget
.handle_serialized_slash_command(UserMessage::from("/personality pragmatic"));
app.chat_widget
.set_composer_text("followup".to_string(), Vec::new(), Vec::new());
app.chat_widget
.handle_key_event(KeyEvent::new(KeyCode::Tab, KeyModifiers::NONE));
app.chat_widget.handle_codex_event(Event {
id: "turn-complete".into(),
msg: EventMsg::TurnComplete(TurnCompleteEvent {
turn_id: "turn-1".to_string(),
last_agent_message: None,
}),
});
assert_eq!(
app.chat_widget.queued_user_message_texts(),
vec!["followup".to_string()]
);
assert!(
op_rx.try_recv().is_err(),
"queued follow-up should not submit before app events drain"
);
loop {
match app_event_rx.try_recv() {
Ok(AppEvent::CodexOp(Op::OverrideTurnContext {
personality: Some(Personality::Pragmatic),
..
})) => continue,
Ok(AppEvent::UpdatePersonality(Personality::Pragmatic)) => {
app.on_update_personality(Personality::Pragmatic);
break;
}
Ok(AppEvent::PersistPersonalitySelection {
personality: Personality::Pragmatic,
}) => continue,
Ok(_) => continue,
Err(TryRecvError::Empty) => panic!("expected personality update events"),
Err(TryRecvError::Disconnected) => panic!("expected personality update events"),
}
}
app.maybe_resume_queued_inputs_after_app_events(true);
match next_user_turn_op(&mut op_rx) {
Op::UserTurn {
items,
personality: Some(Personality::Pragmatic),
..
} => assert_eq!(
items,
vec![UserInput::Text {
text: "followup".to_string(),
text_elements: Vec::new(),
}]
),
other => panic!("expected Op::UserTurn with pragmatic personality, got {other:?}"),
}
}
#[tokio::test]
async fn shutdown_first_exit_returns_immediate_exit_when_shutdown_submit_fails() {
let mut app = make_test_app().await;