fix(tui): clear launch missing-key state when a provider switch lands on a keyed route

[Hmbown]

Root cause of the reported first-launch symptom

This is not a product bug. The tmux server used for the repro has CODEWHALE_PROVIDER=deepseek and CODEWHALE_MODEL=deepseek-v4-flash in its global environment, and every new session inherits them. Config::load then resolves DeepSeek under the documented env > file precedence. Instrumentation showed provider=deepseek already at load_config_from_cli_with_effective_profile, with the env var set. codewhale model resolve from a shell without those vars reported openai, which hid the difference. With the vars unset, the unmodified release build aaaadfe starts on the composer with OpenAI-compatible · gpui-fixture.

Real bug fixed here

After Esc on the recovery picker, /provider openai switched to a keyed route, but switch_provider never recomputed app.onboarding_needs_api_key. Only complete_provider_picker_onboarding clears it, and that runs only while onboarding == Provider. The stale true kept the footer on "model not connected" and left should_adopt_live_local_ollama armed against the provider the user had just chosen.

Fix (crates/tui/src/tui/ui/provider_routes.rs, switch_provider): on success, set onboarding_needs_api_key = !has_api_key(config) for the new route, and clear onboarding_missing_key_recovery when the key is present. The auth-failure rollback already restores the previous value.

Tests (focused, local)

• New first_run_switch_to_keyed_route_clears_launch_missing_key_state runs the real Config::load + App::new startup with the inherited env pair, then Esc, then switch_provider. It passes with the fix and fails without it.
• provider_switch|switch_provider|local_ollama|missing_key|onboarding: 109 passed, 0 failed.
• first_run_|provider_picker: 161 passed, 1 failed. The failure, first_run_ollama_choice_survives_restart_from_canonical_config ("index not found"), also fails on a clean aaaadfe tree and is not touched here.

E2E (tmux 130x40, debug build, fixture at 127.0.0.1:4880)

• Configured home, env clean: composer, footer OpenAI-compatible · gpui-fixture · thinking: max.
• Configured home with the inherited deepseek env: picker (correct per precedence), then Esc and /provider openai: footer OpenAI-compatible · gpt-5.6 …. The release build showed model not connected here.
• Unconfigured home with Ollama unreachable: Choose your model provider picker. With a live local Ollama it auto-adopts ollama/qwen3:4b, the same as the release build.

Out of scope, noted: under an env model override, /provider openai lands on the catalog default (gpt-5.6), not the file's default_text_model, because the env value replaced it in memory.

🤖 Generated with Claude Code

https://claude.ai/code/session_014ZwqatxgVFxHvovngywnks

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, add credits to your account and enable them for code reviews in your settings.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.