mirror of
https://github.com/pchuan98/codex.git
synced 2026-07-01 00:31:56 +08:00
c884536d84
## Why [#25345](https://github.com/openai/codex/pull/25345) was approved, green, and squash-merged into its stacked base branch, `fcoury/tokenmaxxing-api`. Four minutes later, that base branch was force-pushed back to an API-only rebased head while preparing [#25344](https://github.com/openai/codex/pull/25344) for `main`. As a result, the squash commit from #25345 was orphaned and the TUI command never reached `main` or a release. This PR relands the orphaned TUI change from [`411410b8`](https://github.com/openai/codex/commit/411410b85c2d8eb050d441f17396c5c4048d866f) on current `main`. ## What changed - Add `/usage`, `/usage daily`, `/usage weekly`, and `/usage cumulative` for account token activity. - Fetch account usage asynchronously through the existing `account/usage/read` app-server RPC. - Render daily, weekly, and cumulative activity with theme-aware terminal palettes and bounded transient cards. - Preserve transcript ordering while assistant streams, history consolidations, active cells, and hooks complete. - Hide `/usage` from completion when backend auth is unavailable while keeping typed-command guidance. - Carry current-main behavior forward for cwd-aware Markdown parsing, Windows Terminal color detection, and personal access token auth. - Clear pending usage cards on thread rollback and delay completed cards until live hook output is committed. - Add focused regression and snapshot coverage for loading, auth errors, invalid views, rollback, hook ordering, layout, and charts. ## Prior review The original implementation was approved by Eric Traut in #25345 after testing multiple themes and light/dark terminals. This PR preserves that reviewed implementation while adapting it to current `main` and adding regression coverage for newer rollback and hook lifecycle behavior. ## Validation - `just test -p codex-tui token_activity palette renderable usage_command` — 37 passed. - Focused rollback, hook-ordering, and error snapshot tests — 4 passed. - `just fix -p codex-tui` — passed. - `UV_CACHE_DIR=/private/tmp/codex-uv-cache just fmt` — passed. - `cargo insta pending-snapshots` — no pending snapshots. - `just test -p codex-tui` — 2,870 passed; two unrelated guardian feature-flag tests failed because their expected `OverrideTurnContext` event was absent: - `update_feature_flags_disabling_guardian_clears_manual_review_policy_without_history` - `update_feature_flags_disabling_guardian_clears_review_policy_and_restores_default` - `just argument-comment-lint` could not complete because the local Bazel LLVM `compiler-rt` repository is missing `include/sanitizer/*.h`. The touched Rust diff was manually inspected and no missing opaque-literal argument comments were found.
484 lines
20 KiB
Rust
484 lines
20 KiB
Rust
//! Agent turn lifecycle and runtime bookkeeping for `ChatWidget`.
|
|
//!
|
|
//! This module owns task start/completion state, runtime metrics, plan updates,
|
|
//! and final-message separator handling.
|
|
|
|
use super::*;
|
|
|
|
impl ChatWidget {
|
|
/// Synchronize the bottom-pane "task running" indicator with the current lifecycles.
|
|
///
|
|
/// The bottom pane only has one running flag, but this module treats it as a derived state of
|
|
/// both the agent turn lifecycle and MCP startup lifecycle.
|
|
pub(super) fn update_task_running_state(&mut self) {
|
|
self.bottom_pane.set_task_running(
|
|
self.turn_lifecycle.agent_turn_running || self.mcp_startup_status.is_some(),
|
|
);
|
|
self.refresh_plan_mode_nudge();
|
|
self.refresh_status_surfaces();
|
|
}
|
|
|
|
pub(super) fn collect_runtime_metrics_delta(&mut self) {
|
|
if let Some(delta) = self.session_telemetry.runtime_metrics_summary() {
|
|
self.apply_runtime_metrics_delta(delta);
|
|
}
|
|
}
|
|
|
|
pub(super) fn apply_runtime_metrics_delta(&mut self, delta: RuntimeMetricsSummary) {
|
|
let should_log_timing = has_websocket_timing_metrics(delta);
|
|
self.turn_runtime_metrics.merge(delta);
|
|
if should_log_timing {
|
|
self.log_websocket_timing_totals(delta);
|
|
}
|
|
}
|
|
|
|
pub(super) fn log_websocket_timing_totals(&mut self, delta: RuntimeMetricsSummary) {
|
|
if let Some(label) = history_cell::runtime_metrics_label(delta.responses_api_summary()) {
|
|
self.add_plain_history_lines(vec![
|
|
vec!["• ".dim(), format!("WebSocket timing: {label}").dark_gray()].into(),
|
|
]);
|
|
}
|
|
}
|
|
|
|
pub(super) fn refresh_runtime_metrics(&mut self) {
|
|
self.collect_runtime_metrics_delta();
|
|
}
|
|
|
|
// Raw reasoning uses the same flow as summarized reasoning
|
|
|
|
pub(super) fn on_task_started(&mut self) {
|
|
self.input_queue.user_turn_pending_start = false;
|
|
self.turn_lifecycle.start(Instant::now());
|
|
self.transcript.reset_turn_flags();
|
|
self.adaptive_chunking.reset();
|
|
if self.plan_stream_controller.take().is_some() {
|
|
self.request_completed_token_activity_output_insertion();
|
|
}
|
|
self.turn_runtime_metrics = RuntimeMetricsSummary::default();
|
|
self.session_telemetry.reset_runtime_metrics();
|
|
self.bottom_pane.clear_quit_shortcut_hint();
|
|
self.quit_shortcut_expires_at = None;
|
|
self.quit_shortcut_key = None;
|
|
self.update_task_running_state();
|
|
self.status_state.retry_status_header = None;
|
|
self.clear_active_hook_cell();
|
|
self.status_state.pending_status_indicator_restore = false;
|
|
self.bottom_pane
|
|
.set_interrupt_hint_visible(/*visible*/ true);
|
|
self.status_state.terminal_title_status_kind = TerminalTitleStatusKind::Working;
|
|
if self.mcp_startup_status.is_none() || !self.status_header_is_mcp_startup_owned() {
|
|
self.set_status_header(String::from("Working"));
|
|
}
|
|
self.full_reasoning_buffer.clear();
|
|
self.reasoning_buffer.clear();
|
|
self.set_ambient_pet_notification(
|
|
crate::pets::PetNotificationKind::Running,
|
|
/*body*/ None,
|
|
);
|
|
self.request_redraw();
|
|
}
|
|
|
|
pub(super) fn on_task_complete(
|
|
&mut self,
|
|
last_agent_message: Option<String>,
|
|
duration_ms: Option<i64>,
|
|
from_replay: bool,
|
|
) {
|
|
self.input_queue.submit_pending_steers_after_interrupt = false;
|
|
// Use `last_agent_message` from the turn-complete notification as the copy
|
|
// source only when no earlier item-level event (AgentMessageItem, plan
|
|
// commit, review output) already recorded markdown for this turn. This
|
|
// prevents the final summary from overwriting a more specific source.
|
|
let sanitized_last_agent_message = last_agent_message.as_deref().map(|message| {
|
|
parse_assistant_markdown(message, self.config.cwd.as_path()).visible_markdown
|
|
});
|
|
if let Some(message) = sanitized_last_agent_message
|
|
.as_ref()
|
|
.filter(|message| !message.is_empty())
|
|
&& !self.transcript.saw_copy_source_this_turn
|
|
{
|
|
self.record_agent_markdown(message);
|
|
}
|
|
// For desktop notifications: prefer the notification payload, fall back to
|
|
// the item-level copy source if present, otherwise send an empty string.
|
|
let notification_response = sanitized_last_agent_message
|
|
.as_ref()
|
|
.filter(|message| !message.is_empty())
|
|
.cloned()
|
|
.or_else(|| {
|
|
if self.transcript.saw_copy_source_this_turn {
|
|
self.transcript.last_agent_markdown.clone()
|
|
} else {
|
|
None
|
|
}
|
|
})
|
|
.unwrap_or_default();
|
|
self.transcript.saw_copy_source_this_turn = false;
|
|
// If a stream is currently active, finalize it.
|
|
self.flush_answer_stream_with_separator();
|
|
if let Some(mut controller) = self.plan_stream_controller.take() {
|
|
let had_live_tail = controller.has_live_tail();
|
|
self.clear_active_stream_tail();
|
|
let (cell, source) = controller.finalize();
|
|
if !had_live_tail && let Some(cell) = cell {
|
|
self.add_boxed_history(cell);
|
|
}
|
|
if let Some(source) = source {
|
|
self.note_stream_consolidation_queued();
|
|
self.app_event_tx
|
|
.send(AppEvent::ConsolidateProposedPlan(source));
|
|
}
|
|
self.request_completed_token_activity_output_insertion();
|
|
}
|
|
self.flush_unified_exec_wait_streak();
|
|
if !from_replay {
|
|
self.collect_runtime_metrics_delta();
|
|
let runtime_metrics =
|
|
(!self.turn_runtime_metrics.is_empty()).then_some(self.turn_runtime_metrics);
|
|
let show_work_separator = self.transcript.had_work_activity
|
|
&& (self.transcript.needs_final_message_separator || runtime_metrics.is_some());
|
|
if show_work_separator || runtime_metrics.is_some() {
|
|
let elapsed_seconds = if show_work_separator {
|
|
duration_ms
|
|
.and_then(|duration_ms| u64::try_from(duration_ms).ok())
|
|
.map(|duration_ms| duration_ms / 1_000)
|
|
.or_else(|| {
|
|
self.bottom_pane
|
|
.status_widget()
|
|
.map(crate::status_indicator_widget::StatusIndicatorWidget::elapsed_seconds)
|
|
})
|
|
} else {
|
|
None
|
|
};
|
|
self.add_to_history(history_cell::FinalMessageSeparator::new(
|
|
elapsed_seconds,
|
|
runtime_metrics,
|
|
));
|
|
}
|
|
self.turn_runtime_metrics = RuntimeMetricsSummary::default();
|
|
self.transcript.needs_final_message_separator = false;
|
|
self.transcript.had_work_activity = false;
|
|
self.request_status_line_branch_refresh();
|
|
self.request_status_line_git_summary_refresh();
|
|
}
|
|
// Mark task stopped and request redraw now that all content is in history.
|
|
self.status_state.pending_status_indicator_restore = false;
|
|
self.input_queue.user_turn_pending_start = false;
|
|
self.clear_active_hook_cell();
|
|
self.turn_lifecycle.finish();
|
|
self.update_task_running_state();
|
|
self.running_commands.clear();
|
|
self.suppressed_exec_calls.clear();
|
|
self.last_unified_wait = None;
|
|
self.unified_exec_wait_streak = None;
|
|
if !from_replay {
|
|
let body = Notification::agent_turn_preview(¬ification_response);
|
|
self.set_ambient_pet_notification(crate::pets::PetNotificationKind::Review, body);
|
|
}
|
|
self.request_redraw();
|
|
|
|
let had_pending_steers = !self.input_queue.pending_steers.is_empty();
|
|
self.refresh_pending_input_preview();
|
|
|
|
if !from_replay && !self.has_queued_follow_up_messages() && !had_pending_steers {
|
|
self.maybe_prompt_plan_implementation();
|
|
}
|
|
// Keep this flag for replayed completion events so a subsequent live TurnComplete can
|
|
// still show the prompt once after thread switch replay.
|
|
if !from_replay {
|
|
self.transcript.saw_plan_item_this_turn = false;
|
|
}
|
|
// If there is a queued user message, send exactly one now to begin the next turn.
|
|
let follow_up_started = self.maybe_send_next_queued_input();
|
|
let active_goal_continuing = self
|
|
.current_goal_status
|
|
.as_ref()
|
|
.is_some_and(GoalStatusState::is_active);
|
|
// Emit a notification when the agent is truly waiting for the user.
|
|
// Queued follow-up input and active goal continuation both start the
|
|
// next turn immediately, so notifying at that boundary would feel like
|
|
// a false "needs attention".
|
|
if !follow_up_started && !active_goal_continuing {
|
|
self.notify(Notification::AgentTurnComplete {
|
|
response: notification_response,
|
|
});
|
|
}
|
|
|
|
self.maybe_show_pending_rate_limit_prompt();
|
|
}
|
|
|
|
pub(super) fn maybe_prompt_plan_implementation(&mut self) {
|
|
if !self.collaboration_modes_enabled() {
|
|
return;
|
|
}
|
|
if self.has_queued_follow_up_messages() {
|
|
return;
|
|
}
|
|
if self.active_mode_kind() != ModeKind::Plan {
|
|
return;
|
|
}
|
|
if !self.transcript.saw_plan_item_this_turn {
|
|
return;
|
|
}
|
|
if !self.bottom_pane.no_modal_or_popup_active() {
|
|
return;
|
|
}
|
|
|
|
if matches!(
|
|
self.rate_limit_switch_prompt,
|
|
RateLimitSwitchPromptState::Pending
|
|
) {
|
|
return;
|
|
}
|
|
|
|
self.open_plan_implementation_prompt();
|
|
}
|
|
|
|
pub(super) fn open_plan_implementation_prompt(&mut self) {
|
|
let default_mask = collaboration_modes::default_mode_mask(self.model_catalog.as_ref());
|
|
let context_usage_label = self.plan_implementation_context_usage_label();
|
|
|
|
self.bottom_pane
|
|
.show_selection_view(plan_implementation::selection_view_params(
|
|
default_mask,
|
|
self.transcript.latest_proposed_plan_markdown.as_deref(),
|
|
context_usage_label.as_deref(),
|
|
));
|
|
self.notify(Notification::PlanModePrompt {
|
|
title: PLAN_IMPLEMENTATION_TITLE.to_string(),
|
|
});
|
|
}
|
|
|
|
/// Returns a context-used label for the plan implementation prompt.
|
|
///
|
|
/// The footer reports context remaining because it is ambient status, but
|
|
/// this prompt is asking whether to discard prior conversation state before
|
|
/// implementing a plan. Reporting used context makes the cleanup tradeoff
|
|
/// explicit. A fully fresh or unknown context window returns no label so
|
|
/// the clear-context option does not imply urgency without evidence.
|
|
pub(super) fn plan_implementation_context_usage_label(&self) -> Option<String> {
|
|
let info = self.token_info.as_ref()?;
|
|
let percent = self.context_remaining_percent(info);
|
|
|
|
let used_tokens = self.context_used_tokens(info, percent.is_some());
|
|
if let Some(percent) = percent {
|
|
let used_percent = 100 - percent.clamp(0, 100);
|
|
if used_percent <= 0 {
|
|
return None;
|
|
}
|
|
return Some(format!("{used_percent}% used"));
|
|
}
|
|
|
|
if let Some(tokens) = used_tokens
|
|
&& tokens > 0
|
|
{
|
|
return Some(format!("{} used", format_tokens_compact(tokens)));
|
|
}
|
|
|
|
None
|
|
}
|
|
|
|
pub(super) fn has_queued_follow_up_messages(&self) -> bool {
|
|
self.input_queue.has_queued_follow_up_messages()
|
|
}
|
|
|
|
pub(super) fn handle_app_server_steer_rejected_error(
|
|
&mut self,
|
|
codex_error_info: &AppServerCodexErrorInfo,
|
|
) -> bool {
|
|
matches!(
|
|
codex_error_info,
|
|
AppServerCodexErrorInfo::ActiveTurnNotSteerable { .. }
|
|
) && self.enqueue_rejected_steer()
|
|
}
|
|
|
|
/// Finalize any active exec as failed and stop/clear agent-turn UI state.
|
|
///
|
|
/// This does not clear MCP startup tracking, because MCP startup can overlap with turn cleanup
|
|
/// and should continue to drive the bottom-pane running indicator while it is in progress.
|
|
pub(super) fn finalize_turn(&mut self) {
|
|
// Drop preview-only stream tail content on any termination path before
|
|
// failed-cell finalization, so transient tail cells are never persisted.
|
|
self.clear_active_stream_tail();
|
|
// Ensure any spinner is replaced by a red ✗ and flushed into history.
|
|
self.finalize_active_cell_as_failed();
|
|
// Turn-scoped hook rows are transient live state; once the turn is over,
|
|
// do not leave an orphaned running row behind if no matching completion
|
|
// event arrived before cancellation.
|
|
self.clear_active_hook_cell();
|
|
// Reset running state and clear streaming buffers.
|
|
self.input_queue.user_turn_pending_start = false;
|
|
self.turn_lifecycle.finish();
|
|
self.update_task_running_state();
|
|
self.running_commands.clear();
|
|
self.suppressed_exec_calls.clear();
|
|
self.last_unified_wait = None;
|
|
self.unified_exec_wait_streak = None;
|
|
self.adaptive_chunking.reset();
|
|
self.stream_controller = None;
|
|
self.plan_stream_controller = None;
|
|
self.request_completed_token_activity_output_insertion();
|
|
self.status_state.pending_status_indicator_restore = false;
|
|
self.clear_cancel_edit();
|
|
self.request_status_line_branch_refresh();
|
|
self.request_status_line_git_summary_refresh();
|
|
self.maybe_show_pending_rate_limit_prompt();
|
|
}
|
|
|
|
pub(super) fn on_server_overloaded_error(&mut self, message: String) {
|
|
self.input_queue.submit_pending_steers_after_interrupt = false;
|
|
self.finalize_turn();
|
|
|
|
let message = if message.trim().is_empty() {
|
|
"Codex is currently experiencing high load.".to_string()
|
|
} else {
|
|
message
|
|
};
|
|
|
|
self.add_to_history(history_cell::new_warning_event(message));
|
|
self.request_redraw();
|
|
self.maybe_send_next_queued_input();
|
|
}
|
|
|
|
pub(super) fn on_error(&mut self, message: String) {
|
|
self.input_queue.submit_pending_steers_after_interrupt = false;
|
|
self.flush_answer_stream_with_separator();
|
|
self.finalize_turn();
|
|
self.add_to_history(history_cell::new_error_event(message));
|
|
self.set_ambient_pet_notification(
|
|
crate::pets::PetNotificationKind::Failed,
|
|
/*body*/ None,
|
|
);
|
|
self.request_redraw();
|
|
|
|
// After an error ends the turn, try sending the next queued input.
|
|
self.maybe_send_next_queued_input();
|
|
}
|
|
|
|
pub(super) fn on_cyber_policy_error(&mut self) {
|
|
self.input_queue.submit_pending_steers_after_interrupt = false;
|
|
self.finalize_turn();
|
|
self.add_to_history(history_cell::new_cyber_policy_error_event());
|
|
self.request_redraw();
|
|
|
|
// After an error ends the turn, try sending the next queued input.
|
|
self.maybe_send_next_queued_input();
|
|
}
|
|
|
|
pub(super) fn on_rate_limit_error(&mut self, error_kind: RateLimitErrorKind, message: String) {
|
|
let rate_limit_reached_type = self.codex_rate_limit_reached_type.map(|kind| {
|
|
if matches!(error_kind, RateLimitErrorKind::UsageLimit) {
|
|
match kind {
|
|
RateLimitReachedType::WorkspaceOwnerCreditsDepleted => {
|
|
RateLimitReachedType::WorkspaceOwnerUsageLimitReached
|
|
}
|
|
RateLimitReachedType::WorkspaceMemberCreditsDepleted => {
|
|
RateLimitReachedType::WorkspaceMemberUsageLimitReached
|
|
}
|
|
other => other,
|
|
}
|
|
} else {
|
|
kind
|
|
}
|
|
});
|
|
self.codex_rate_limit_reached_type = rate_limit_reached_type;
|
|
|
|
match rate_limit_reached_type {
|
|
Some(RateLimitReachedType::WorkspaceOwnerCreditsDepleted) => {
|
|
self.on_error(
|
|
"You're out of credits. Your workspace is out of credits. Add credits to continue using Codex."
|
|
.to_string(),
|
|
);
|
|
}
|
|
Some(RateLimitReachedType::WorkspaceOwnerUsageLimitReached) => {
|
|
self.on_error(
|
|
"Usage limit reached. You've reached your usage limit. Increase your limits to continue using codex."
|
|
.to_string(),
|
|
);
|
|
}
|
|
Some(RateLimitReachedType::WorkspaceMemberCreditsDepleted) => {
|
|
self.on_error(message);
|
|
self.open_workspace_owner_nudge_prompt(AddCreditsNudgeCreditType::Credits);
|
|
}
|
|
Some(RateLimitReachedType::WorkspaceMemberUsageLimitReached) => {
|
|
self.on_error(message);
|
|
self.open_workspace_owner_nudge_prompt(AddCreditsNudgeCreditType::UsageLimit);
|
|
}
|
|
Some(RateLimitReachedType::RateLimitReached) | None => {
|
|
self.on_error(message);
|
|
}
|
|
}
|
|
}
|
|
|
|
pub(super) fn handle_non_retry_error(
|
|
&mut self,
|
|
message: String,
|
|
codex_error_info: Option<AppServerCodexErrorInfo>,
|
|
) {
|
|
if codex_error_info
|
|
.as_ref()
|
|
.is_some_and(|info| self.handle_app_server_steer_rejected_error(info))
|
|
{
|
|
} else if codex_error_info
|
|
.as_ref()
|
|
.is_some_and(is_app_server_cyber_policy_error)
|
|
{
|
|
self.on_cyber_policy_error();
|
|
} else if let Some(info) = codex_error_info
|
|
.as_ref()
|
|
.and_then(app_server_rate_limit_error_kind)
|
|
{
|
|
match info {
|
|
RateLimitErrorKind::ServerOverloaded => self.on_server_overloaded_error(message),
|
|
RateLimitErrorKind::UsageLimit | RateLimitErrorKind::Generic => {
|
|
self.on_rate_limit_error(info, message)
|
|
}
|
|
}
|
|
} else {
|
|
self.on_error(message);
|
|
}
|
|
}
|
|
|
|
pub(super) fn on_warning(&mut self, message: impl Into<String>) {
|
|
let message = message.into();
|
|
if !self.warning_display_state.should_display(&message) {
|
|
return;
|
|
}
|
|
self.add_to_history(history_cell::new_warning_event(message));
|
|
self.request_redraw();
|
|
}
|
|
|
|
pub(super) fn on_app_server_model_verification(
|
|
&mut self,
|
|
verifications: &[AppServerModelVerification],
|
|
) {
|
|
if verifications.contains(&AppServerModelVerification::TrustedAccessForCyber) {
|
|
self.on_warning(TRUSTED_ACCESS_FOR_CYBER_VERIFICATION_WARNING);
|
|
}
|
|
}
|
|
|
|
pub(super) fn on_plan_update(&mut self, update: UpdatePlanArgs) {
|
|
self.transcript.saw_plan_update_this_turn = true;
|
|
let total = update.plan.len();
|
|
let completed = update
|
|
.plan
|
|
.iter()
|
|
.filter(|item| match &item.status {
|
|
StepStatus::Completed => true,
|
|
StepStatus::Pending | StepStatus::InProgress => false,
|
|
})
|
|
.count();
|
|
self.transcript.last_plan_progress = (total > 0).then_some((completed, total));
|
|
self.refresh_status_surfaces();
|
|
self.add_to_history(history_cell::new_plan_update(update));
|
|
}
|
|
|
|
pub(super) fn interrupted_turn_message(&self, reason: TurnAbortReason) -> String {
|
|
if reason == TurnAbortReason::BudgetLimited {
|
|
return "Goal budget reached - the turn was stopped.".to_string();
|
|
}
|
|
|
|
"Conversation interrupted - tell the model what to do differently. Something went wrong? Hit `/feedback` to report the issue.".to_string()
|
|
}
|
|
}
|