Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

Extend the walltime drift ratio from shell-seeded tests to compiled tests

Open
  • Is something wrong?
#1,653 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
4/5
Estimated time
3-5 days
Newbie friendliness
48/100
Issue type
Feature
Clarity
Mostly clear
Activity status
Active
Tech stack
c, python
Domain
ci-cd, testing

Research direction

Start with scripts/ci/check-shell-gate-walltime.py and inspect how it joins meson-info/intro-tests.json with meson-logs/testlog.json; compare that population filter with compiled test registrations. Run the existing walltime check and examine the reported test count and ratios. Done means compiled tests are included, the distribution and outliers are addressed, undersized populations fail, and a near-budget test is shown to trip the check.

Written by the indexing model from the issue text.

Description

enhancement p2-medium

Triage — 2026-09-22

Target: Infra-0.A. Priority: p2-medium.

Introduce compiled-test drift measurement report-only first, validate population completeness, then justify outlier budgets/enforcement. Share measurement logic with #1797.

This scheduling/ownership update supersedes older milestone assignments below; original acceptance criteria remain binding. A target milestone is a completion target, not a requirement to delay ready work.

Discovery context

PR #1615. col_rel_deep_copy sat at 24.54s against Meson's 30s default on a green main TSan run — 82% of budget, with no contention. Nothing reported that. It was discovered only when a new test's contention pushed it over and the leg went red.

The repo already owns the instrument that would have caught it months earlier; it is just scoped narrowly.

What exists

scripts/ci/check-shell-gate-walltime.py joins meson-info/intro-tests.json (registered timeout) with meson-logs/testlog.json (measured duration) and fails when duration > 0.5 * timeout. Its own docstring states the motivation precisely: a timeout alone is "deliberately useless for detecting drift: a gate that quietly goes from 3 s to 110 s passes everything".

It applies only to shell-seeded tests — shell_seed_from_cmd requires argv[0] to be a .sh, or bash/sh with a .sh in argv[1]. Compiled test executables are outside the population by construction, so col_rel_deep_copy was never measured against its own budget.

Proposal

Generalise the same join to compiled tests. The data is already in the two JSON files and the comparison is unchanged; only the population filter differs.

This will surface a triage queue rather than a clean pass — roughly 268 tests currently sit at the 30s default — so it needs a plan rather than a flag flip:

  • Run it in report-only mode first and publish the ratio distribution.
  • Give the tests above the bar explicit, justified timeouts (the rule the surrounding registrations already follow: explicit and justified, not uniform).
  • Only then make it enforcing, possibly with a separate fraction for compiled tests if the shell-gate 0.5 proves wrong for them.

Note the sharp edge in the existing gate, which any generalisation inherits: it escalates only on a fully empty population, so a truncated testlog.json — for example from a targeted meson test overwriting the full-run log — yields a confident green over a silently shrunken set, and WIRELOG_WALLTIME_REQUIRED=1 does not catch it. The printed test count is the only thing distinguishing a real pass from a hollow one. Worth fixing while touching this, by asserting an expected floor on the population size.

Acceptance criteria

  • Ratio reporting covers compiled tests as well as shell seeds.
  • The distribution is published and the outliers are either given justified timeouts or recorded as acceptable.
  • The gate fails when the measured population is implausibly small, not only when it is empty.
  • A deliberately near-budget test is shown to trip the new check.

Related

  • #1487 — the shell-gate walltime ratio this generalises
  • #1488, #1489 — extending the explicit-timeout rule to other seed kinds
  • #1652 — the specific tests currently at the default
Dominant language
C
Stars
18
Forks
2
Avg merge
9h 28m
Merged PRs (30d)
169

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from semantic-reasoning/wirelog

All issues in semantic-reasoning/wirelog

Similar issues

More C issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.