Skip to content
This repository was archived by the owner on Oct 11, 2026. It is now read-only.

chore(analytics): note regen uptime parity — 0 diffs corpus-wide [#117] - #118

Closed
Roddygithub wants to merge 2 commits into
mainfrom
fix/regen-coverage-parity
Closed

Roddygithub wants to merge 2 commits into
mainfrom
fix/regen-coverage-parity

Conversation

@Roddygithub

Copy link
Copy Markdown
Owner

Summary

The regen coverage tracking via coverage_ms field (added in #117) matches EI's buffUptimes[718].uptime across the full corpus (0 differences across 20 logs).

Remaining known case

  • 20260125-001308: single player (Ovalkvadratcylinder.9365) with ~34ms active-window delta (our 52.791 vs EI 52.751, 0.04%).
    • Root cause: 7 long-duration WvW regen events — the chained coverage drains entries fully serially, but EI's window is computed from state transitions. Residual at the last entry boundary.
    • This is a distinct corner case, not a regression. Tracked for a future PR.

Test

  • uv run python scripts/ei-parity/ei_diff.py → 0 diffs corpus-wide

The coverage_ms/coverage-based reporting in b1fb47b matches EI's buffUptimes[718].uptime
across the full corpus (20 logs). The only outlier (20260125-001308) has a ~34ms active-window
delta with EI on a single player (long-duration WvW regen), traced to the last event
alignment. This is a known 0.04% residual independent of the main port.

Signed-off-by: Roland Salardon <r.salardon@gmail.com>
Signed-off-by: Roland Salardon <r.salardon@gmail.com>
@Roddygithub

Copy link
Copy Markdown
Owner Author

Closing: this PR contains two empty commits (+0/-0) and no file changes. The [skip ci] marker also means the required checks can never run, so it can never satisfy the ruleset.

Its useful content is the note about the remaining regen corner case on 20260125-001308 (Ovalkvadratcylinder.9365, ~34 ms active-window delta). That has been carried into BACKLOG.md so it is not lost.

Separately: the "0 diffs corpus-wide" figure in #117/#118 was measured over 20 logs. Re-running scripts/ei-parity/ei_diff.py over the committed 35-log corpus gives 4 503 differences. Being tracked and worked on now.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant