mirror of
https://github.com/block/buzz.git
synced 2026-10-30 05:51:46 +00:00
### Summary Typing in the mobile composer felt slower than a native iOS text field. A device trace on an iPhone 12 mini showed the cause: while anyone in the channel is typing, the whole channel page rebuilds once per second, including the visible message list. - `ChannelTypingNotifier._prune()` runs on a 1 s timer and assigned a new list to `state` on every tick, even when nothing had changed. - `_ResolvedChannelDetailPage` watched the whole list and only filtered out the current user afterwards, so every tick rebuilt the page and its `PositionedList` (12 to 35 ms per frame, 8 to 19 ms for the message list alone). - Your own typing indicator echoes back from the relay, so the timer kept running the whole time you typed. In a 30-key trace, 33 of the 35 frames over 12 ms were these rebuilds, 1.00 s apart. ### Fix - `_prune()` publishes only when the visible entries change. `TypingEntry` now compares by value. - The channel and thread pages watch `visibleTypingProvider`, a `VisibleTyping` value that already excludes the current user and applies the page's scope. Your own echo refreshes its expiry every 3 s, which used to rebuild the page. Now it doesn't. Other `channelTypingProvider` consumers are unchanged. ### Related issue None found. ### Testing - Provider tests: a typing entry held across five prune ticks notifies listeners once, and the pages' shared `visibleTypingProvider` is not notified by the user's own refreshed echo but is by another user. Breaking either guard fails its test. - `just mobile-check` and `just mobile-test` pass at8b8a16845. - Device trace on an iPhone 12 mini: 30 keys typed into a channel composer, Profile builds with build tracing on both sides. Control is6661d594a, which has the same page code as main. | | control | this PR (5ff754131) | |---|---|---| | Channel page rebuilds while typing | about 1 per second (44 in 40 s) | 0 | | Largest UI build per key, median / p90 | 4.9 / 29.3 ms | 1.0 / 1.4 ms | | Keys with a build over 16.7 ms | 9/30 | 0/30 | | Frames with UI work over 12 ms | 35 (33 page rebuilds) | 1 (a semantics update from the test driver) | | ComposeBar frame, median | 10.2 ms | 3.0 ms | The test driver turns on accessibility, so semantics runs on every frame in both traces. About 1 key in 4 still has a slow frame with about 1 ms of build and 2 ms of raster. A Time Profiler capture put that time in the driver's accessibility queries on the main thread (52% of main-thread samples while typing, none before the first automated tap), so it is a test-harness artifact rather than app work. - Dictation on an iPhone SE (2nd gen): a 15 s paragraph dictated into a channel composer, Profile builds with build tracing. | | main | this PR (5ff754131) | |---|---|---| | UI frame, median / p99 / max | 0.9 / 10.4 / 26.1 ms | 0.9 / 5.3 / 10.8 ms | | Frames over 16.7 ms | 10 | 0 | | Channel page rebuilds | 26 (once a second) | 6 (composer growth, none over budget) | Each dictation update costs about 1 ms. The caret stayed at the end and nothing flickered on either build. buzz-review-completed 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Signed-off-by: Tom Brow <tomb@block.xyz> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>