feat(swift-ios): choose the reasoning level from the composer

[saphid]

Model options required opening the model picker. This adds composer controls for the selected model's reasoning, service tier, and other supported options.

Changing one option preserves unlisted saved values and leaves options without declared defaults unset. The control uses the current model-selection policy and shows the saved value instead of substituting a menu default.

Validation: 68 focused native tests passed at 4e361bf1e4ce04f07584d4680690b9cff16edad3, rebased onto SwiftUI base 614a7b672a86ee638f284298fe26e8b0194e0eaa. Coverage includes option preservation, native dispatch, and keyboard dismissal. No new live interaction capture was made.

Historical images and videos

These captures compare the original base with head 2524de5c55914a9364c56f1dcff7c5953b74ea98. They were not recaptured or revalidated for the modernized branch.

Dark mode before:

Dark mode after:

Dark mode comparison

Light mode before:

Light mode after:

Light mode comparison

Interaction GIF · Video · Annotated video

Tracking: native work item.

Modernized and fixed with GPT-6 in Codex.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 0fda1ea5-6762-467b-a742-8559ba6ed64a

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

Warning

Your free Security trial is over. An organization admin can activate Security or dismiss this notice.

---

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — This adds a new Swift iOS composer traits picker and routes selected reasoning and service-tier values into model submissions. Its multi-file production-path integration and nontrivial selection, default-resolution, and persistence logic make the impact broader than a small isolated UI adjustment.

Not approved because:

• Monthly spending limit reached (workspace setting). Approvability relies on correctness review in order to determine eligibility

^{Review your spending limits in Billing settings. You can add or adjust custom eligibility rules. Learn more.}

[t3dotgg]

Need this updated and also screenshots/vids

[saphid]

@t3dotgg Updated: the branch is now on the current t3code/rebuild-mobile-app-swift base (961a3ae), with the final selector changes at 6cf8887; GitHub reports it 0 commits behind and mergeable. The requested media is still pending: the required canonical proof simulator lock (~/.t3/locks/swiftui-proof-sim.lock) is currently held, so I did not disturb its owner. We still need both light- and dark-mode captures plus a short selector video, and I’ll attach them in a follow-up when that simulator is available.

[saphid]

@t3dotgg Updated again: I rebased the branch onto the current t3code/rebuild-mobile-app-swift head (c2276320) and pushed exact head b849a8800. The PR body now contains exact-head light and dark screenshots plus a short interaction recording (Low → High, then reopen with High checked). Focused FeatureComposerPowerTests are 25/25 passing, and the exact-head SwiftUI simulator build succeeded. I also fixed the selector label to meet the 44-point minimum tap target in b849a8800. GitHub reports the PR mergeable; no inline human review threads currently exist.

[saphid]

@t3dotgg Updated against the live SwiftUI target and refreshed the requested proof.

• PR head: 696f9d428e5069da90e5d758a45017d1c9b0d409
• live target: b67837984e5650888b1eed0e1cd7fff83625f696 (ancestor of the PR head)
• exact-head dark-first and light before/after screenshots plus comparison GIFs are now in the PR body
• exact-head interaction GIF plus clean and annotated 23-second MP4s are linked in the PR body
• the installed simulator app reported the exact PR SHA and its executable matched the preserved build byte-for-byte
• the three-action proof packet passed validation; current CI native tests and the main checks are green, with Cursor's head rerun still pending

The prior b849a880 media has been replaced rather than presented as current proof.

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 46b2288. Configure here.}

[saphid]

@t3dotgg Updated to the superseding acceptance target at exact head 2524de5c55914a9364c56f1dcff7c5953b74ea98.

The SwiftUI composer now renders every supported model-option descriptor section in descriptor order. The Codex proof shows Reasoning followed by Service Tier, Standard as Default, Fast with the live descriptor text, and retained/effective selections through the existing turn-dispatch path. Unsupported sections stay hidden; provider/model changes, inherited values, prompt-injected hidden values, and default materialization are covered by focused tests.

Focused verification: 32 tests passed, 0 failed; swiftc -parse and git diff --check passed. Exact base b67837984e5650888b1eed0e1cd7fff83625f696 remains an ancestor. GitHub reports no conflicts.

Dark mode — exact base → exact head

Interaction proof

Clean playable video: https://github.com/user-attachments/assets/b17c2f48-294f-46da-ad87-9da57ba01e12

Annotated playable video: https://github.com/user-attachments/assets/22c313e1-94a5-400c-b124-4b6dedb07d49

Light mode — exact base → exact head

The raw exact-head interaction capture is 76.758 seconds (SHA-256 1f14953d7c351a31498a776a9daef347a058258af7b3eed27dfb234a4a1fdbf4). Timeline validation passed all 5 declared actions. The live backend currently supplies 1.5x speed, increased usage without terminal punctuation; the UI renders the descriptor verbatim, while the focused fixture also covers the supplied punctuated form.

Direct Claude Opus high review was attempted read-only on the frozen change, but the process exited 1 with HTTP 429 and zero model usage; no Claude model participated and cross-provider review was skipped because capacity was exhausted.

[saphid]

GIF crop correction

The three earlier GIFs were cropped too tightly and hid most of the descriptor menu. I replaced their live PR-body embeds without changing the PR head.

• Exact PR head remains 2524de5c55914a9364c56f1dcff7c5953b74ea98.
• The corrected GIFs are full-width 720×1158 crops. They keep the complete Reasoning section, the complete Service Tier section including Fast's description, the popover anchor, and the composer trigger in one frame.
• Dark, light, and interaction GIFs use one explicit source-bound crop decision. All three anonymous downloads returned HTTP 200 as image/gif and matched the inspected local hashes byte-for-byte.
• Media stays outside the product diff at saphid/t3code media commit 6dcfcf115e9cea4ac2aee0464bc89c23db520f6f.
• The durable renderer fix is local commit db1596389 on t3code/review-project-skills: GIF rendering now requires a reasoned crop decision that binds exact source hashes, dimensions, required visible details, and a receipt. Its focused suite passes 37/37 tests.

Dark comparison

Interaction

Light mode comparison

Crop decision · dark receipt · light receipt · interaction receipt

[saphid]

Current status: do not merge yet. Theo's SwiftUI target advanced to c49bcc5d8a56324eb0ad9cf1b19d5dcc6ca0caa0 with overlapping model-option behavior after this PR's proof was captured. The head is textually mergeable and its current CI/AI reviews are green, but it is classified blocked until the combined picker/composer behavior is retested on the live target, refreshed Simulator proof is published, and the integrated result receives fresh AI review.

[saphid]

Upstream successor handoff — #110 / SwiftUI Test Build 86

This records the contribution state only. The coordinator-owned phone acceptance remains on saphid/t3code-personal#110: Alex’s recorded verdict was “Reasoning menu looks good.” I did not create or alter that acceptance.

• Accepted candidate / current PR head: 2524de5c55914a9364c56f1dcff7c5953b74ea98.
• Test Build 86 generation receipt: SHA-256 0b46a98cbc196efa58a1ddc17fb20309706082c7fc8e7208be14f0cbf294a4fb; its feat: copy button on user messages, hover on message for buttons #110 candidate entry names the same head and binds proof 6ffac5eb28335a3806525dec09021f38dea8b91747f0ec227f93a77bb1d1e4a5 plus inspection 56dc80709a602bd9b68b42d6dce2dcfbaa62f4b2f5e26b2a853bcf0b41cf7f28.
• Installed-and-launched device receipt: SHA-256 2cea2febb651bf6675e09ca75889c41d77b014077c4d0dbd4b16d4a15023e654.
• Live target: b67837984e5650888b1eed0e1cd7fff83625f696; fetched ancestry check passes (0 behind, 9 ahead).
• Review: all four published review threads are resolved. Theo’s request to update the branch and provide screenshots/video is addressed by the live exact-head PR body and the full-menu GIF correction.
• Evidence: exact-base/exact-head dark and light images, full-width dark/light/interaction GIFs, and clean/annotated MP4s are live. Anonymous range checks returned 206 with the expected PNG/GIF/MP4 content types for every current embed/link.
• Verification: focused native tests recorded 32 passed, 0 failed, 0 skipped; Swift parsing and git diff --check passed. Current CI’s substantive jobs are green, including SwiftUI native tests, CI checks/server partitions, Rust, release smoke, vouch, size, Cursor Bugbot, and Macroscope correctness; non-applicable preview/static-analysis jobs are skipped and Approvability is neutral.
• The earlier direct Claude Opus high review attempt on this frozen head exited with HTTP 429, zero tokens, and no model usage, so cross-provider review remains truthfully recorded as skipped.

GitHub currently reports MERGEABLE / CLEAN; the PR is open, trusted, and labeled size:XL. No upstream human approval is recorded (reviewDecision is unset), so the accurate handoff is: phone-accepted and proof-complete, ready for upstream maintainer review/decision, but not independently classified as merge-ready under the contribution gate yet.

No product commit, push, rebase, proof mutation, media upload, or PR-body change was made during this handoff; the accepted candidate was preserved exactly.