Repository navigation
land: main-flake-fixes - #500
Merged
Merged
Conversation
…n Windows lane_paused_initial_build_search_shows_approved_sentence relied on a background trigram build finishing inside the search's interactive wait budget. Under Windows module load it did not, both lanes read not-ready (trigram building, semantic unavailable because the pause was found) and the search refused with search_lanes_unavailable. The pause lookup itself matched; the test now installs a ready trigram index up front. small_chunk_reuse_fill_completes_with_cold_cap_held decided pass/fail by a 2 s recv timeout. It now waits for either the fill's result or a queued waiter on the held cold-build limiter, which is the defect it guards. Co-authored-by: alfonso-aft <289616620+alfonso-aft@users.noreply.github.com>
…nts, not wall clock long_line_with_many_occurrences_builds_line_details_once_per_arm ran its scan against the real clock with a 2 s budget, so a loaded Windows runner could end the scan (scan_deadline_reached) before the count it checks was decided. The scan clock is now frozen and advances half the budget per line-details build: one build never reaches the deadline, a per-occurrence regression reaches it on the second build and fails the count quickly. subc_storm_same_or_lower_epoch_rebind_is_rejected waited for its setup bind with the debug-scaled 8 s storm assertion bound and skipped any Error reply for that bind, so a slow cold bind or a refusal both surfaced as 'timed out waiting for RouteBindAck'. The setup bind (not the subject of the test) now waits for its reply and fails with the refusal body if the module refuses it; the scenario watchdog bounds a missing reply. Co-authored-by: alfonso-aft <289616620+alfonso-aft@users.noreply.github.com>
Co-authored-by: alfonso-aft <289616620+alfonso-aft@users.noreply.github.com>
Let the existing private clock seam construct a deadline with independent scan and read instants. Production still computes and clamps one instant at the same point and assigns it to both fields, preserving its read and scan budgets. Keep the long-line test's frozen two-second scan budget and unchanged assertions, but give bounded file reads a real sixty-second hang guard. Retain a three-second pause before the second arm's read: it failed with zero line-detail calls before the fix and now passes. Reusing the frozen read deadline and restoring per-occurrence line-detail work each redden only this test. Verified the search_index filter (134 passed, 3 ignored), strict Windows-target test compilation, restored focused regression (1 passed), and cargo fmt --check. Co-authored-by: alfonso-aft <289616620+alfonso-aft@users.noreply.github.com>
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.
Accepted agent delivery; branch-keyed search descriptor included.
Summary by cubic
Fixes flaky tests in the aft crate that timed out on loaded Windows runners. The tests now wait on events instead of wall-clock budgets.
Bug Fixes
Written for commit 5d22113. Summary will update on new commits.


Confidence Score: 5/5
The PR appears safe to merge; no blocking issues remain.Summary
Replaces short timing assumptions with ready test data and event-based waits.
Reviews (2) · Last reviewed commit: "mason: separate real read deadlines from..." · Reviewed by Greptile