Repository navigation
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Signed-off-by: zlareb1 <lareb.zafar@clickhouse.com>
zlareb1
force-pushed
the
fix/graph-stats-pass-clock
branch
from
November 21, 2026 00:19
5e2e751 to
4a183d8
Compare
4 of 5 tasks
This branch has not been deployed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Related: #1009
Changelog category (leave one):
Changelog entry
The graph-stats cancellation fixture uses an idle clock for its pass deadline.
Why
The pass timeout starts before the cheap legs reach the blocked unit-risk query. Unrelated scheduling time can exhaust that allowance before the fixture observes its in-flight signal. The hosted failure occurs at the two-second signal watchdog; the actual scheduling cause of that run is unknown. A controlled monotonic jump independently reproduces this wall-clock dependency.
Scope and risk
One existing test module changes. The cancellation case and its controls use a private function loop with the existing manual clock. Pure calculations run inline through the fixture's thread-offload seam, with the original functions, arguments and results. The fixture therefore controls both loop time and executor scheduling while the real graph-stats pass, repository fake and cancellation paths run.
The original test body retains its 0.2-second pass deadline, two-second signal watchdog, five 0.01-second post-cancellation waits, two namespaces, statement-count freeze and completed cheap-leg assertions. Teardown cancels and drains unexpected tasks and rejects leaks. A negative control observes cancellation of a genuinely blocked edge query at the unchanged 0.2-second deadline and checks the complete 0.25-second logical timeline through the original freeze assertions. Independent review must assess the fixture's scheduling coverage.
Compatibility, rollout and rollback
Tests
Prior source qualification on Python 3.12 with the declared development and code extras showed that the controlled scheduler regression fails its 0.25-second jump case and passes its zero-jump control before the fixture fix. The identical command passed both cases afterward:
On the composed head, the complete directly related module passes all 24 cases, including the blocked-query deadline control. Scoped Ruff, Python formatting and whitespace checks pass. Full exact-head hosted CI and fresh independent whole review remain required before merge. No database, provider, runtime or benchmark was started.
AI assistance in the code (leave one; writing this description with AI is fine):
AI-generated: Codex diagnosed the timing dependency, wrote the scoped fixture and checked preservation of the original test bodies.
I can explain every line of this diff and how it interacts with the rest of Loom.
Checklist