Repository navigation
Visible row colors, clearer row wording, skip granted permission page - #6
Merged
jappeace merged 3 commits intoJul 14, 2026
Merged
Conversation
Three field findings from Jappie: - The rate coloring never rendered: WidgetStyle properties apply to the single node they wrap (Hatter.Render.applyStyle), so styling the row's column set text color on a container, a silent no-op. Every text node of the row now carries the color itself; the color test asserts all six texts are colored (it fails against the old container styling). - "~1024 ms" read as meters/second: the rate now renders as "every ~1024 ms". The "found" status now reads "not yet configured". - When both permissions are already granted the app skips the permission page: the first render probes checkPermission (the platform bridges are registered right before the first render, see jni_bridge.c) exactly once and flips straight to the scanner. The emulator test asserts the skip since CI pre-grants via pm; the permission-page unit tests pre-mark the probe because the desktop stub always grants. Both new behaviours sabotage-verified to fail their tests alone. Prompt: we still don't have the colored row if the configured interval mismatches what's on the ble device, also what does the m/s mean? and what does found mean? use the latest master. also if all permissions are set we should skip the permission screen Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Author
|
Heads-up: the last commit pins hatter at the jappeace/hatter#243 branch (text color moved into the text-bearing configs, resolving hatter#242 so this class of invisible styling cannot recur) and migrates |
hatter #243 moved text color from WidgetStyle into the text-bearing configs, resolving #242 so a container color is unrepresentable. rowText uses the new coloredText constructor, the TextInput records carry the new field, and the test walker reads the config colors. Pin moves to the hatter PR branch; repoint to master after #243 merges (hash-only). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Identical tree to the reviewed branch head; the text-color API is upstream. Co-Authored-By: Claude Fable 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.
Field findings from testing the merged #5:
Row colors actually render now
The rate coloring never showed on-device:
WidgetStyleproperties apply to the single node they wrap (Hatter.Render.applyStylesets the property on that node), so styling the row's column set text color on a LinearLayout, a silent no-op. Every text node of the row now carries the color itself. The color test asserts all six row texts are colored and was verified to fail against the old container styling. Colors remain: green when the measured advertisement rate matches the expected interval input within 25%, red when it deviates, uncolored until a rate is measured (second advertisement).Wording
~1024 msread as meters/second: now renders asevery ~1024 ms(the measured advertising interval).foundnow readsnot yet configured.Permission page skips when everything is granted
The first render probes
checkPermissionfor both permissions (safe: jni_bridge.c registers the platform bridges immediately before invoking the first render) exactly once, and lands directly on the scanner when both are granted. A denial keeps the page as before. The emulator test now asserts the skip (CI pre-grants viapm); the permission-page unit tests pre-mark the probe since the desktop stub always grants.65 tests green, zero warnings, shellcheck clean; both new behaviours sabotage-verified to fail their tests alone.
🤖 Generated with Claude Code