Skip to content

docs(target): record Minitest 6.1.0 changelog review in minitest target - #32

Merged
developerz-ai[bot] merged 2 commits into
mainfrom
dz/review-minitest-6-1-0-against-the-target-and--93ece43f
Nov 12, 2026
Merged

developerz-ai[bot] merged 2 commits into
mainfrom
dz/review-minitest-6-1-0-against-the-target-and--93ece43f

Conversation

@developerz-ai

@developerz-ai developerz-ai Bot commented Nov 12, 2026 •

Copy link
Copy Markdown
Contributor

What

Review upstream Minitest 6.1.0 (2026-11-06) against the documented
surface: two additive APIs (Test.abstract_test_case!,
Minitest.installed_at_exit? + idempotent autorun) and tightened -W
category validation; no breaking changes to the drop-in contract.
Clears the upstream-changelog-review:minitest checkup row.

Acceptance

  • docs/target/minitest-rails.md names Minitest 6.1.0 (released 2026-11-06) as reviewed against upstream History.rdoc, with a review date
  • docs/target/minitest-rails.md records the 6.1.0 surface deltas: new Minitest::Test.abstract_test_case!, new Minitest.installed_at_exit? + idempotent autorun, -W unknown-category validation now exit!s, and notes no breaking change to the documented drop-in contract
  • The compat metric 'unreviewed upstream minor releases affecting this target' reads 0 for minitest because the target doc records the 6.1.0 review outcome

Changes

  • docs/target/minitest-rails.md (edit) — Record the Minitest 6.1.0 changelog review. Minitest 6.1.0 (released 2026-11-06, upstream History.rdoc) shipped vs the doc's pinned 6.0.6. Add a clearly-marked review record (new short section near the top, and adjust the line 3 'Sources read' note to state 6.1.0 was reviewed on 2026-11-11 via History.rdoc while mt6/ citations still point at the actually-read 6.0.6 tree). Surface deltas to record, verified against the v6.1.0 source: (1) NEW public API Minitest::Test.abstract_test_case!(klass = self) — class method that extends the class with a module whose run no-ops when `self = [trimmed]

Verification

./bin/check passed (exit 0) in 39s.

What the editors ran to see the change working, and what they saw:

  • curl -sf https://raw.githubusercontent.com/minitest/minitest/v6.1.0/lib/minitest.rb + lib/minitest/test.rb, then grep -n for the new/changed symbols (docs): v6.1.0 confirmed: def self.abstract_test_case! klass = self at test.rb:46 (extends with a module whose run calls super(*args) unless self == klass); @@installed_at_exit ||= false at minitest.rb:15, autorun guard } unless @@installed_at_exit at :88, def self.installed_at_exit? at :96; -W[<error>] option at :232 with warn "Unknown warning category: %p" at :243 and bare exit! at :245
  • python3 in-place edit of docs/target/minitest-rails.md with asserted anchors, then code_read of lines 1-45 (docs): Line 3 now states 'Minitest 6.1.0 (released 2026-11-06) reviewed 2026-11-11 via upstream History.rdoc ... mt6/ citations below still point at the actually-read 6.0.6 tree'; new '## 0. Changelog review — Minitest 6.1.0 (2026-11-06)' section at line 21 with a 4-row delta table (abstract_test_case!, installed_at_exit?/idempotent autorun, -W exit!, internal-only entries) and the note 'nothing in §5 (drop-in contract) changes'
  • curl raw v6.0.6 + v6.1.0 lib/minitest.rb from github.com/minitest/minitest, grep -n for installed_at_exit / autorun / -W / exit! (docs): 6.0.6: @@installed_at_exit ||= false at :15, autorun at :69, '} unless @@installed_at_exit' at :85, set at :86 — the idempotent guard predates 6.1.0, confirming finding (1). 6.1.0: same guard unchanged at :72/:88/:90 plus NEW reader 'def self.installed_at_exit?' at :96; -W[] at :232, bare exit! at :245 (row 31 citations correct). Drift: process_args 152→163 (+11), Minitest.run 299→319 (+20).
  • curl raw v6.0.6 + v6.1.0 lib/minitest/test.rb, grep -n abstract_test_case! / i_suck_and_my_tests (docs): 6.1.0: abstract_test_case! at test.rb:46, i_suck at :58 (row 29 citations correct); 6.0.6 has no abstract_test_case! (i_suck at :31) — row 29 is accurate, left unchanged.
  • python3 exact-string replacement of the two §0 lines in docs/target/minitest-rails.md, then code_read of lines 21-33 (docs): Row 30 now reads 'NEW public reader Minitest.installed_at_exit?' with the guard credited to pinned 6.0.6 (mt6/minitest.rb:69/:85/:86, §1.1) and v6.1.0 positions :72/:88/:90 noted as unchanged moves; line 23 drift now attributed to the new installed_at_exit? method + autorun doco (+11 by process_args) and the -W validation branch (+20 at Minitest.run).

🤖 Opened by developerz.ai, an automated agent, not a human, on developerz-ai/fast_rspec.

Raised automatically by a developerz.ai maintainer lane.

🤖 developerz.ai, an automated agent, not a human — this PR is one slice of a stacked plan.
Dispatched automatically by a developerz.ai maintainer lane
Raised automatically by the developerz.ai raw task triage maintainer lane, from manual.
Plan: https://app.developerz.ai/fleet/plans/rtk_577f6de802d64a4eb7f0310a7bb3afe5
Part 1/1
BYOK spend: $0.00 (flat-rate subscription: no per-token charge).

Review upstream Minitest 6.1.0 (2026-11-06) against the documented
surface: two additive APIs (Test.abstract_test_case!,
Minitest.installed_at_exit? + idempotent autorun) and tightened -W
category validation; no breaking changes to the drop-in contract.
Clears the upstream-changelog-review:minitest checkup row.

Dz-Task-Id: tsk_93ece43fc48fe1d38b82b4c35fa8b720
@developerz-ai developerz-ai Bot added the dz/fleet-pr Opened by developerz.ai, an automated agent, from a fleet run. label Nov 12, 2026
@developerz-ai

developerz-ai Bot commented Nov 12, 2026 •

Copy link
Copy Markdown
Contributor Author

Fix failing CI on #32

Work landed on #32 — watching it toward merge.

2m · 13 tool calls · 3 messages


🤖 developerz.ai, an automated agent, not a human — developerz-ai[bot]

@developerz-ai

developerz-ai Bot commented Nov 12, 2026

Copy link
Copy Markdown
Contributor Author

No code change warranted. The fix_pr follow-up for this pull request investigated its feedback and concluded no code change is warranted:

NO CHANGE — failing checks on PR #32 are unrelated to this PR's diff and unfixable from this round's file set (docs/target/minitest-rails.md only).

Failing checks named in the brief:

  1. "Demo apps / upstream vs fast_rspec" (CI job 131278731756, run 43245206976)
  2. "CI OK" — fails only because it aggregates the demo job ("Results: success success success success success failure"); same root item.

What actually failed (staged log tmp/dz/ci/[REDACTED:high_entropy].log):

  • demo/ruby: fully green (rspec 131 ex 0 fail; minitest 105 runs 0 fail). The fast_rspec stacks are red but that is by design and non-gating ("fast_rspec stack is expected red until roadmap M3-M4"; STRICT is not set in .github/workflows/ci.yml).
  • demo/rails UPSTREAM (the gating stack): rspec "98 ex, 1 fail" at "Failure/Error: visit new_post_path"; minitest "98 runs, 1 fail, 0 skip" reported as an Error.
  • The single failure in each suite is the Chrome/cuprite JS system test: demo/rails/spec/system/body_counter_spec.rb ("updates live and the post saves", first line visit new_post_path, line 6) and demo/rails/test/system/body_counter_test.rb (BodyCounterTest, same first line). Proof it is the JS test and not the rack_test posts tests: both suites have TWO rack_test tests that visit new_post_path (posts_spec.rb lines 7/18; posts_test.rb lines 7/19), so a server-side 500 would fail two tests per suite, not one; and the run reports 0 pending / 0 skips, so the Chrome binary was found (DemoJS.available? true) and the browser session itself errored at the first visit.

Why this PR cannot be the cause:

  • PR docs(target): record Minitest 6.1.0 changelog review in minitest target #32 changes exactly one file, docs/target/minitest-rails.md (see the PR's own "Changes" section and commit 1241d79). Nothing in the test suites, the demo apps, or CI loads a docs/target markdown file. The failure does not exercise code this PR changed.
  • Main was green the same day on identical demo code: CI run #68 for commit 8996b13 (Nov 12, 2026) "completed successfully", including this demo job. The demo gemfiles are unpinned, so I checked rubygems.org for releases between the green main run and the red PR run: none (cuprite 0.20.1 released 2026-11-08, minitest 6.1.0 and rspec-rails 8.1.0 on 2026-11-06 — all predate the green main run). Same code, same gems, different result.
  • The failure hit the initial run and the re-run (run 43245206976, re-run triggered Nov 12 15:16), and that run carries GitHub's annotation "The ubuntu-latest label will migrate to Ubuntu 26 beginning October 19, 2026". Runner-image drift breaking the headless-Chrome session (cuprite/ferrum) fits every observation: only the two Chrome-driven tests fail, everything else — including all non-JS Rails tests on the same stack — passes.

Why I changed nothing instead of "fixing" it:

  • The fix would live in demo/rails/spec/support/capybara.rb and demo/rails/test/support/js_driver.rb (browser options / driver config) — outside this round's exclusive file set, and these are tests/support OUTSIDE this PR's diff, which the brief forbids me to weaken, skip or rewrite. Widening was not requested: the widening exception covers tests that assert on what this PR changed or lockfiles the change needs; neither applies — this failure is unrelated to the docs change.
  • The full browser error is not in the staged logs (the GitHub logs endpoint needs auth; only the one-line hints are staged) and this box has no Ruby, so any driver-config edit would be an unverifiable guess.
  • If the Ubuntu 26 runner image has durably broken cuprite/Chrome, main CI will fail the same way and it needs its own PR (demo driver config, or pinning the runner in .github/workflows/ci.yml — a locked merge-gate file I may not touch). That is a separate task, not part of a docs-only changelog-review PR.

Verdict: flake/environmental failure in unrelated demo-app JS system tests. Retire for this PR: check "Demo apps / upstream vs fast_rspec" (and dependent "CI OK"), tests demo/rails/spec/system/body_counter_spec.rb and demo/rails/test/system/body_counter_test.rb ("updates live and the post saves").

Posted automatically by the developerz bot; the reasoning above is the follow-up run's own (AI-authored).

@developerz-ai

developerz-ai Bot commented Nov 12, 2026 •

Copy link
Copy Markdown
Contributor Author

Updated this branch from its base; waiting for CI on the new head. Nothing is needed from you right now.

  • Red on this head: Demo apps / upstream vs fast_rspec, CI OK
  • What was done: Merged the base branch into this branch (no force-push).
  • What happens next: CI re-runs on the new head and the merge goes ahead once it is green.

Status for head 1241d79d0b6d. This comment is updated in place.

🤖 Posted by developerz.ai — an automated agent, not a human.

@developerz-ai

developerz-ai Bot commented Nov 12, 2026

Copy link
Copy Markdown
Contributor Author

No code change warranted. The fix_pr follow-up for this pull request investigated its feedback and concluded no code change is warranted:

developerz.ai[bot]: no code change this round. The failing checks are an environmental failure in the demo Chrome system test, which sits outside this PR's diff and doesn't run any code this PR changes.

Checks: "Demo apps / upstream vs fast_rspec" and "CI OK". CI OK fails only because it aggregates the demo job (RESULTS: success success success failure).

What the staged log shows ([REDACTED:high_entropy].log):

  • demo/ruby: upstream rspec and minitest both pass. The fast_rspec rows are expected red ("fast_rspec stack is expected red until roadmap M3–M4").
  • demo/rails: the UPSTREAM stack fails one example in each suite:
    rspec × upstream: 98 ex, 1 fail -> "Failure/Error: visit new_post_path" (38.0s wall)
    minitest × upstream: 98 runs, 1 fail (13.2s wall)
    visit new_post_path with a matching failure in both suites points to the Chrome/cuprite JS system test: demo/rails/spec/system/body_counter_spec.rb and demo/rails/test/system/body_counter_test.rb. The cuprite driver is set to process_timeout: 20 in demo/rails/spec/support/capybara.rb and demo/rails/test/support/js_driver.rb. On this box the same rspec run takes 12.0s, so the extra ~26s on CI fits the browser hanging until it times out. Inference: the job ran on a new runner image (ubuntu-26.04 20261111.196.1).

Why this PR didn't cause it:

  • The PR's file set is docs/target/minitest-rails.md only: the Minitest 6.1.0 changelog review in §0. Nothing in demo/, apps/fast-rspec/lib or the Gemfiles changes. A markdown target doc can't change how the upstream gems (rspec-rails, capybara, cuprite) drive Chrome.
  • I reproduced it on this box at the PR head, with real Chrome (Google Chrome 156.0.8078.79):
    demo/rails/bin/stacks upstream rspec -> | rspec | upstream | green | 98 ex, 0 fail, 0 pend | 12.0s |, EXIT=0
    demo/rails/bin/stacks upstream minitest -> 98 runs, 184 assertions, 0 failures, 0 errors, 0 skips, EXIT=0
    0 skips means the JS test really ran against Chrome and passed. The same upstream suites fail only on the CI runner.

Why there's no edit:

  • The only possible fix would be in demo/rails support or test files, or in CI config. Those are outside this PR's file set, and the platform refuses CI-config changes. The test also doesn't assert on anything this PR changed, so this round isn't entitled to widen to it. Raising the cuprite timeout to get green would be the loosened-timeout workaround the brief forbids.
  • Action for a human: re-run the "Demo apps / upstream vs fast_rspec" job. If it stays red on main too, open a separate task to make the demo Chrome system test robust on the new ubuntu-26.04 runner image. Start with the Ferrum/cuprite error text in tmp/stacks/rspec-upstream.log from the CI artifacts, which the staged log doesn't include.

Posted automatically by the developerz bot; the reasoning above is the follow-up run's own (AI-authored).

@developerz-ai developerz-ai Bot mentioned this pull request Nov 12, 2026
@developerz-ai
developerz-ai Bot merged commit 5de0764 into main Nov 12, 2026
9 checks passed
@developerz-ai
developerz-ai Bot deleted the dz/review-minitest-6-1-0-against-the-target-and--93ece43f branch November 12, 2026 19:35
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

dz/fleet-pr Opened by developerz.ai, an automated agent, from a fleet run.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

0 participants