feat(tui): dynamic status indicator for Execute vs Streaming states

[echobt]

Summary

Add distinct status indicators for different streaming phases:

• 'Execute' when request is pending (waiting for first token)
• 'Streaming..' when actively receiving tokens from the LLM

Motivation

Previously, both states showed 'Working' which didn't distinguish between waiting for a response and actively streaming one. This provides clearer feedback to users about the current state of their request.

Changes

• Add is_actively_streaming field to StreamingState
• Update status_header() to show different text based on streaming phase
• Mark is_actively_streaming = true when first delta arrives

Testing

• cargo check passes
• cargo test passes
• Manual verification of status changes

[greptile-apps]

Greptile Overview

Greptile Summary

This PR adds a new StreamingState.is_actively_streaming flag to distinguish the “waiting for first token” phase (Execute) from the “receiving tokens” phase (Streaming..). The flag is reset when streaming starts/stops, and is set when the event loop receives the first StreamEvent::Delta, which then drives MinimalSessionView::status_header() to render the updated status text.

Main thing to validate is that StreamEvent::Delta corresponds to real user-visible token output; if Delta can be empty/keepalive, the UI will incorrectly switch to Streaming.. before any content is shown.

Confidence Score: 4/5

• This PR is safe to merge after addressing the Delta->active-streaming transition condition.
• Changes are small and localized, but the new status transition currently triggers on any StreamEvent::Delta, which can misrepresent the UI state if deltas can be empty/keepalive or non-content.
• src/cortex-tui/src/runner/event_loop/streaming.rs

Important Files Changed

Filename	Overview
src/cortex-tui/src/app/streaming.rs	Adds is_actively_streaming flag to StreamingState, resets it on start/stop, and exposes start_active_streaming() helper.
src/cortex-tui/src/runner/event_loop/streaming.rs	Marks start_active_streaming() on every StreamEvent::Delta; needs guarding to avoid incorrectly switching to Streaming on empty/non-content deltas.
src/cortex-tui/src/views/minimal_session/view.rs	Updates status header to show Execute vs Streaming.. based on new is_actively_streaming flag.

Sequence Diagram

sequenceDiagram
  participant U as User
  participant L as EventLoop
  participant S as StreamingState
  participant V as MinimalSessionView
  participant P as LLM Provider

  U->>L: "Submit prompt"
  L->>S: "start_streaming()"
  V-->>U: "Status shows Execute"

  L->>P: "Begin provider stream"
  P-->>L: "Delta arrives"
  L->>S: "start_active_streaming()"
  V-->>U: "Status shows Streaming.."

  P-->>L: "Done"
  L->>S: "stop_streaming()"
  V-->>U: "Status shows Idle"

Loading

[greptile-apps]

_{3 files reviewed, 1 comment}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

Active-streaming flips too early

start_active_streaming() is called on every StreamEvent::Delta, but Delta can be an empty/whitespace keepalive (or other non-user-visible chunk). In that case the UI will switch from Execute to Streaming.. even though the user still hasn’t received the first token. Consider guarding this with something like if !delta.is_empty() / if delta.chars().any(|c| !c.is_whitespace()) (or using a dedicated “first content token received” event/flag from the streaming layer).

Prompt To Fix With AI

This is a comment left during a code review.
Path: src/cortex-tui/src/runner/event_loop/streaming.rs
Line: 305:308

Comment:
**Active-streaming flips too early**

`start_active_streaming()` is called on every `StreamEvent::Delta`, but `Delta` can be an empty/whitespace keepalive (or other non-user-visible chunk). In that case the UI will switch from `Execute` to `Streaming..` even though the user still hasn’t received the first token. Consider guarding this with something like `if !delta.is_empty()` / `if delta.chars().any(|c| !c.is_whitespace())` (or using a dedicated “first content token received” event/flag from the streaming layer).

How can I resolve this? If you propose a fix, please make it concise.