Repository navigation
fix(assistant): align sidebar, refine dictation and respond to stop - #262
Merged
Merged
Conversation
chaxus
enabled auto-merge (rebase)
October 13, 2026 02:10
Deploying document with
|
| Latest commit: |
7e912b0
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://42877869.document-7hm.pages.dev |
| Branch Preview URL: | https://fix-assistant-interactions.document-7hm.pages.dev |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Merge Verdict
[APPROVE] — Local verification and all hosted CI checks pass on
7e912b06b39b92a358ed282b6b557053fd032678.Summary
The sidebar used different insets for its heading, empty state and composer, while model buttons stacked outer and inner padding. Align those surfaces, use compact settings controls and group model preparation/release actions in one wrapping row. Dictation now has distinct microphone, listening-stop and finishing states, with a compact cancel icon in the toolbar. Stop now returns cancellation to the UI immediately while WebLLM drains the underlying operation before starting another inference.
Commits
fix(assistant): align sidebar content and compact model settingsfix(assistant): clarify compact speech input statesfix(assistant): acknowledge inference cancellation immediatelytest(assistant): enter chat before measuring sidebar layoutRisk Analysis
Design Decisions
Retain ranui and the existing SpeechRecognition integration. Reference AI Elements' compact microphone toggle, listening feedback and processing states without adding React, shadcn or a new transcription service. Live announcements and reduced-motion behavior remain available.
Code Concerns
No outstanding blocking findings from source review. A stopped WebLLM worker may continue draining internally; subsequent inference waits for that drain to finish instead of sharing the engine concurrently.
Verification
Full unit suite: 5,112 tests passed across 172 files.
Focused panel, speech, WebLLM and chat-layout tests: 202 passed.
Regression for immediate cancellation failed before the fix and passed after it; checks queue serialization as well.
Format, oxlint, TypeScript and 1,511 immutable evaluation artifacts validated.
Production Vite build completed.
Local Chromium browser acceptance: 6/6 passed. Layout checks passed at 320px and 1280px (shared left edges, 30–34px model action height, no settings overflow). Word, Excel, PowerPoint and PDF dictation checks passed for transcript finalization on explicit Send and cancellation preserving the draft.
Additional local Chromium coverage: 11 tests passed for composer behavior, locale handling and model-loading geometry at 360px and 640px.
CI follow-up: the initial run failed only in the new layout test, which measured a hidden composer while the default view was Settings. Enter Chat through the actual editor and onboarding controls first. The updated test passed at 320px and 1280px against the deployed Cloudflare preview and all three hosted environments.
Conclusion basis
Verified: unit suites, static checks, build and remote trees matching the local commit trees. Reviewed: cancellation callbacks, speech session transitions, moved runtime-row consumers and responsive CSS overrides. Verified in Chromium: responsive geometry and controlled speech-event interaction across all four editors. Hosted verification: all checks passed on
7e912b06b39b92a358ed282b6b557053fd032678, including standard E2E, Docker-image E2E, Cloudflare Pages semantics, preview smoke and lint/validation. Logs confirm both responsive layout cases and all four editor dictation cases passed in each of the three environments. Not covered: physical microphone recognition accuracy and hardware-specific real-model cancellation latency.