[CODE-1832] Show the last response's duration and credits in the TUI indicator slot

[ianhodge]

Description

Stacked on #13492. Shows the last response's duration and credits in the TUI's indicator slot (CODE-1832, parent CODE-1828; Figma).

What

• While a response streams, the ⋮ Warping (Ns) row (TUI warping indicator #13442) is unchanged — the mocks keep usage hidden during streaming (the usage element in frame 323:17216 is opacity: 0).
• Once the response completes, the same slot rests on a static summary row: ∷ 5s • 0.5 credits — resting spinner glyph, wall-to-wall duration, and the credits the response spent (the mock's ∷ 1s • … completed state, with credits instead of tokens per the CODE-1831 direction).
• The credits segment is omitted until any are reported; brand-new conversations show nothing.

How

• render_response_summary joins the indicator family in warping_indicator.rs; muted styling per the mock, credits via the GUI's format_credits.
• Wired in TuiTerminalSessionView::render from existing block-level accessors — wall_to_wall_response_time_since_last_query() and credits_spent_for_last_block() (the GUI footer's "Credits spent (last response)"). No app-side changes.
• Repaints reuse the CODE-1831 ConversationUsageMetadataUpdated subscription plus the status-change repaint that already removes the Warping row.
• Historical per-block rows in the transcript would need per-block stamping app-side; deferred (noted in specs/CODE-1828/TECH.md).

Linked Issue

Internal Linear-driven work — CODE-1832; no GitHub issue.

• The linked issue is labeled ready-to-spec or ready-to-implement. (N/A — Linear-tracked internal task)
• Where appropriate, screenshots or a short video of the implementation are included below (especially for user-visible or UI changes).

Testing

• Unit tests in warping_indicator_tests.rs: the summary row renders ∷ 5s • 2.5 credits with no repaint scheduling; the credits segment is omitted for None/zero. cargo nextest run -p warp_tui (97/97), clippy -D warnings, ./script/format all clean.

• E2E verified on a live PTY run against staging (two full prompt/response cycles). Rendered-screen captures confirmed:

  • Streaming: ⋮ Warping (0s) with no credits segment, for both prompts.
  • Completion: the slot rests on ∷ 1s • 2.4 credits after the first response and ∷ 1s • 1 credit after the second (singular form correct).
  • Footer (CODE-1831 regression): auto (genius) ~/warp • 2.4 credits click-toggles to • $0.04 and back; footer stays hidden until usage is first reported.
  • Cycling: the second prompt swaps the summary row back to the Warping row, then a fresh summary; footer cumulative updates 2.4 → 3.4 credits (last-response 1 credit + prior 2.4).

• I have manually tested my changes locally with ./script/run-tui

Screenshots / Videos

Live PTY verification screen captures (streaming, completed summary rows for both responses, and the footer credits/$ toggle both directions) recorded from the run described in Testing above.

⋮ Warping (0s)                                     ← streaming (no credits)
∷ 1s • 2.4 credits                                 ← completed, response 1
∷ 1s • 1 credit                                    ← completed, response 2
auto (genius) ~/warp • 2.4 credits  ⇄  • $0.04     ← footer toggle

Agent Mode

• Warp Agent Mode - This PR was created via Warp's AI Agent Mode

Warp conversation

Co-Authored-By: Oz oz-agent@warp.dev

[oz-for-oss]

@ianhodge

I'm starting a first review of this pull request.

You can view the conversation on Warp.

I completed the review and no human review was requested for this pull request.

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

[oz-for-oss]

Overview

This PR adds a completed-response summary row to the TUI indicator slot, showing the resting spinner, wall-to-wall duration, and optional credits for the last response. I did not find actionable correctness, security, or spec-drift issues in the annotated diff.

Concerns

• This is a user-facing TUI change, but the attached PR description does not include actual screenshots or a screen recording. It only says E2E verification with recording is queued. For this user-facing change, please include screenshots or a screen recording demonstrating it working end to end.

Verdict

Found: 0 critical, 1 important, 0 suggestions

Request changes

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

[kevinyang372]

very much a nit -- but I don't think these two unit tests are that useful