perf: drop the progress overlay and dead diagnostics, tighten collection timing - #685
Merged
Conversation
…ion timing Measured on-device after the seeding fix: 21:01:21.147 page load -> 21:01:23.738 injected, so 2.59s, down from ~10.5s. A second playlist in the same session came back in 0.38s from cache. Trimming what is left. Removed: - The "Playlist: loading batch N…" overlay (_showProgress/_hideProgress and their call sites). It was useful while collection took ten seconds; now it appears and disappears inside a couple of seconds, immediately before a reload, which is just flicker. - The continuation stack-trace capture. It built and logged a 45-frame string on EVERY continuation request, for an investigation that is closed — the stack showed the request is driven through an observable chain with no callable trigger, which is recorded in the commit history. Pure per-request cost now. Timing: - NATIVE_SETTLE_MS 1200ms -> 700ms. Native continuations arrive ~200ms apart (measured: 21:01:21.508, .690, and the burst ends by .758), so 700ms of quiet is already a clear end-of-burst signal. - Pace only BETWEEN our own fetches. The first needs no delay: nothing of ours precedes it and the scheduler has just waited out the native burst. With seeding usually leaving a single batch to fetch, that delay was pure latency on the critical path. Together roughly 900ms off the common case. Note on unused code: nothing is now unreferenced — the checked suspects (_triggerReveal, _verifyAutoLoadEffect, dumpHelperTileStructure, attemptPlaylistAutoLoad) all still have live callers. The two removed here were live but obsolete rather than unreachable. _verifyAutoLoadEffect and dumpHelperTileStructure are the next candidates if the helper behaviour stays fixed, but both are capped and cheap, so they are left for now.
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
1 participant
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.
Confirmed working on-device: 21:01:21.147 page load → 21:01:23.738 injected = 2.59s, down from ~10.5s. A second playlist in the same session returned from cache in 0.38s.
Removed
Playlist: loading batch N…overlay. It earned its place when collection took ten seconds; now it flashes on and off within a couple of seconds immediately before a reload.Timing
NATIVE_SETTLE_MS1200ms → 700ms. Native continuations arrive ~200ms apart (21:01:21.508, .690, burst ends .758), so 700ms of quiet is already a clear end-of-burst signal.Roughly 900ms off the common case.
On unused functions: nothing is actually unreferenced. The suspects I checked (
_triggerReveal,_verifyAutoLoadEffect,dumpHelperTileStructure,attemptPlaylistAutoLoad) all still have live callers. The two removed here were live but obsolete._verifyAutoLoadEffectanddumpHelperTileStructureare the next candidates if helper behaviour stays fixed, but both are capped and cheap, so I left them.