Repository navigation
docs: attribute default-test waits to executed operations - #1926
Conversation
|
Historical wait-evidence PR #1926 merged as Automatic ready CI 38619609256, attempt1, passed both selected documentation jobs and current contract proof passed. Planner logs identify the three docs files and deliberately exclude typecheck, unit/runtime, browser/layout, captures and packaging; aggregate validates those skips. Both executed checkout logs identify The data covers72runs across71files at frozen source04a, with source hashes, actual calls/spans, case names, missing timing nulls and explicit namespace profiles. Original1009cases/5337expectations and a separately repeated28/70 diagnostic remain distinguishable. Parent review independently checked all source hashes and observation expressions, including imported sources. Native/process/browser boundaries and overlapping spans are not collapsed into a total saving. New independent audit work continues on a separate tree; this merged branch receives no further pushes. |
|
Automatic main-push CI 38619808365, attempt1, passed both selected documentation jobs for #1926 squash The merged clean documentation worktree was confirmed unused and removed; branch and evidence remain. Independent audit work continues on its own frozen-source checkout. PR and advanced issue retain milestone0.44.0(58). |
|
Additional automatic website publication-proof occurrence after docs PR #1926: run 38619808068, attempt 1, artifact job 116045238202, fails the existing Actual website checkout is Static verification passed 157 HTML pages and 501 artifact digests; share/publication/release-selection guards and all four initial demo smoke scenes passed before the heading failure. Diagnostic artifact 11921217856 is retained. This follow-up inspected the source and complete failed job log, not the artifact's runtime DOM or network trace; it does not independently re-establish the false-404 causal diagnosis already owned here. The artifact job failed, and deploy was skipped. Execution cost was 368 job seconds; event-to-completion was 510 seconds including queue. No website publication, unchanged retry, timeout increase or source correction was attempted. Preserve this newer 0.43.0 occurrence under #1772 / its blocked PR #1776. The #1926 documentation CI proof passed separately and does not imply the website workflow passed. |
Refs #1855
What changed
Adds source-attributed observations for 72 default-test runs across 71 files, including actual awaited counts and elapsed spans, full-file/per-case timings, resource usage, source hashes, namespace profiles and explicit measurement limits. Corrects the track-draft timestamp-separation role without rewriting the historical source inventory.
Why this shape
The observations distinguish native waits, logical clock advances, fixture latency, overlapping helpers and losing watchdogs. They retain all original waits and assertions and do not claim a repository-wide runtime census, a speedup, or a universal minimum observation window. The guide preserves the minimum justified window policy and identifies native boundaries that a replacement must preserve.
Proof
04a659593379a0c8da9c049fa649124d6fd54f70; documentation base:f6f5c713b2c52b400a5d34fc1d04120d3ea54107.bun run check:pr --base origin/mainpassed on7f75a79fc5ab913c3c6b6919d26ff2602450eb75: lint/leakage, environment documentation and changeset checks. This documentation-only change adds no behavioral test or mutation requirement. Hosted selected proof remains pending.Checklist
docs/reflect the change, or nothing documented moved.Anything left open
#1855 remains open: imported callbacks, remaining default/native/experiment families, complete operational minimum-window justifications and final tier ownership are not covered by this sample. Existing native negative windows remain unchanged. #1918 owns imported scrape-retry clock coverage; #1922 owns the recovery deadline. Higher measurement cost alone does not exempt a family from required proof.