Repository navigation
fix: prevent periodic stalls in rolling waveform projections - #4
Merged
Merged
Conversation
added 4 commits
November 12, 2026 11:22
cuimmm
approved these changes
Nov 12, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
A 192 kHz waveform retaining 60 seconds eventually projects distinct X samples to the same Float32 coordinate. The rolling display cache then falls back to all 11.52 million samples per channel, causing repeated half-second renderer stalls.
Add a projection-rebase work counter, regression coverage for repeated rolling-window drift, and consistent package/runtime/gallery version
1.0.1-opentest.30. This branch includes the latestmainchanges.Evidence
Before: the 192 kHz / 60-second regression failed on its first rolling update with
display cache must not fall back to all 11520000 raw samples.After: repeated precision crossings keep display reduction and existing summaries; page and certified-arena tests preserve peaks, NaN gaps and subsequent incremental updates. Irrecoverable spacing still falls back after one attempt.
Latest merged tree: typecheck and build passed; 1,221 source tests passed, 2 skipped; 31 OpenTest contracts passed.
Actual Windows/Electron Monitor, two channels, full 60-second retention, 2560×1392 viewport and DPR 1.5:
Each observed rebase scanned only 9,600 new samples and projected about 22,500 display samples. Both channels continued receiving data. A separate native-window capture contained one 116 ms display reselection without a raw-geometry fallback; this change does not promise every frame fits 16.7 ms.
Host integration: 59 targeted waveform tests and 8 artifact checks passed. The existing host browser-smoke fixture references a missing calibration component and could not initialize; actual Windows validation above completed independently.
Merge Danger
Door: two-way
No persisted-data migration. Raw samples and strict projected-order validation are retained; a consumer can revert its pinned artifact.
Blast Radius: rendering
Affects the display-cache recovery path used by paged/arena line rendering. Rebase frames replace display/GPU/ledger caches while preserving source summaries.