mirror of
https://github.com/zeronsh/comet.git
synced 2026-11-16 18:57:19 +00:00
The track's bars moved from the time alone. Now the host keeps the old app's meter (crates/ui/src/dictation/meter.rs, ported to apps/zgx/src/dictation/meter.rs) fed from the session's take_level() while listening, and publishes the newest 128 bars, newest first, in a shared value of their own (`dictation_levels`) as each 60 ms slot fills. The poll loop wakes on the slot while listening. Nothing is published while not listening. ds/voice_wave draws them in the old app's proportions: 3 pt bars with 3 pt gaps, 16 high, the quietest a dot at 0.08, newest at the trailing edge, fading at the leading edge, and a still dot baseline past the bars. Starting is a breathing baseline (a clocked fill shader). Finalizing holds the bars while a highlight sweeps across them (a clocked layer shader). ds.Composer takes it from the dictation prop's `levels` and `phase`. Without them (the gallery, fixtures), it keeps the time-driven shader. Cost: about 256 script calls and 0.1 ms per slot (16 Hz), none in the frames between slots beyond the track's existing clock tick. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: zeronsh[bot] <346184242+zeronsh[bot]@users.noreply.github.com> |
History
|
|
|---|---|---|
| .. | ||
| archived_list.rs | ||
| chat_drop.rs | ||
| chat_paste.rs | ||
| chat_window.rs | ||
| cli.rs | ||
| composer_tray.rs | ||
| design_lint.rs | ||
| dictation.rs | ||
| full_screen_titlebar.rs | ||
| hover_cards.rs | ||
| leftovers.rs | ||
| onboarding_continue_ink.rs | ||
| pointer_cursor.rs | ||
| reload.rs | ||
| sidebar_row_ink.rs | ||
| streaming_cost.rs | ||
| thought_streaming.rs | ||