Repository navigation
Make Voice Input work on first launch, redesign it, add e2e tests - #67
Merged
Merged
Conversation
- Bundle the pinned Whisper Tiny starter model and admit it through the normal installer on first launch (decision 0061). - Fix History failing to load after a reinstall or update: audio paths are rebased onto the current app container. - Add the Signal Bridge app icon and a monochrome accent. - Replace the single scrolling page with Dictate, History, and Settings tabs; one record button, amber only while recording, result with Copy/Share. - UI tests dictate a recorded clip end to end through capture, ASR, formatting, and History search (Debug-only microphone seam). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On a device, directory listing spells the container /private/var while constructed paths use /var, so an installed package never equalled itself and every selection failed with "no longer installed". Installed packages now carry one canonical root path; a symlinked-root test reproduces it. Provisioning also selects a starter left unselected by an earlier launch. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The 4 Hz refresh loop republished unchanged state, redrawing SwiftUI and keeping the app from idling; the style UI test took 258 s and timed out on CI. Unchanged refreshes now publish nothing (17 s; the iOS check runs in about a minute). check_ios.sh prints each failing test's message, since -quiet hid it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
CodeQL built the iOS app from its own copy of the preparation steps and lacked the bundled starter model. prepare_ios_build_inputs.sh is now the single source for project generation and the traced build, and the starter's inputs route to the iOS CodeQL lane. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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.
Why
Voice Input could not dictate on a fresh install (no speech model without a Mac-side prepare + Files import), had no icon, a cluttered single-page UI, and History broke after any reinstall or update.
What
testFreshInstallDictatesIntoSearchableHistory(~16 s) drives fresh install → record → ASR → result → History search withjfk.wavreplacing only the microphone (Debug-only seam). Plus Settings, style menu, and large-text tests.Verification
scripts/check_ios.sh: build, lint, all unit tests, 4 UI tests pass (304 s).scripts/check_fast.shpasses.🤖 Generated with Claude Code