Skip to content

Commit 1b01244

Browse files
luisleo526claude
andcommitted
CI D1 fix: the preflight failures of lab job rj-20261019t202712-e0ff98 (test_ci_preflight, doc-anchors, doc-reverts) and L6's rule for the deferred rows
Fixes the three failing preflight stages of the D1 policy change: - scripts/test_ci_preflight.py: the must-fail test for the full-population rule injected a `steps:` line into the job `corpus-parity-subset`, which is a reusable-workflow call (`uses:` + `with:`) and has no `steps:`; the mutation now goes into its `with:` block. The rule it exercises (corpus-parity-subset keeps its full population) is unchanged. - doc anchors: four citations of `PRIVATE_FIXTURE_PROFILE_FLOORS` moved from ci_verify.py:789 to ci_verify.py:802, because the EXCLUDED_REGISTERED_MIN comment above it grew (CONTRIBUTING.md:162, docs/pages/contributing-llm.md:136 and :284, docs/pages/native-engine.md:4616). - doc-reverts: docs/ci.md intentionally replaces the old statement of the PR registration floors ("The PR registration floors after lane FIX-E1E2 are 773 for Debug and sanitizers and 782 for native ...", introduced by b2a578c "docs: make every README and doc page true before the release (DOCS-ENGINE) (#300)"): those numbers were already stale against scripts/ci_verify.py (876/867/963) and are now the keys of EXCLUDED_REGISTERED_MIN. This commit names that deletion on purpose: b2a578c. - docs/ci.md also gains L6's rule for the deferred rows: a red on main in an excluded row or class blocks the next release cut and is fixed or reverted within the hour. The registration floors for Release and kernel-only (895 and 309) are still derived, not measured; a lab run of the release and kernel steps confirms or corrects them. Co-Authored-By: Claude Sonnet 5.5 <[email protected]>
1 parent 11cd290 commit 1b01244

5 files changed

Lines changed: 19 additions & 7 deletions

File tree

‎CONTRIBUTING.md‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -159,7 +159,7 @@ What each gate refuses:
159159

160160
`ci_verify.py` counts the CTest rows that actually **ran** and fails below a
161161
floor. The floors' numbers live in one place, the
162-
`PRIVATE_FIXTURE_PROFILE_FLOORS` ci_verify.py:789 table: per profile, `full`
162+
`PRIVATE_FIXTURE_PROFILE_FLOORS` ci_verify.py:802 table: per profile, `full`
163163
is its overlay-present registered population and `pinned` is how many rows of
164164
`tests/private_fixture_rows.txt` that registry registers, so with the private
165165
fixtures absent the floor is `full` minus `pinned`; every other profile

‎docs/ci.md‎

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -839,6 +839,13 @@ and 1838 s, and a push run on the current pool has not been measured. The
839839
pull-request wall time under D1 is to be measured on real PR runs, typical and
840840
tail; it is not claimed as met against the owner's 15-minute target.
841841

842+
**Rule for the deferred rows (L6 condition, 10-20).** Every row and class this
843+
section excludes from a pull request still runs on the merge commit. A red
844+
result on `main` in an excluded row or class (the Release and kernel-only
845+
`slow` rows, or a SLOW verifier class) BLOCKS the next release cut, and the
846+
author or the release owner fixes or reverts the change within the hour; the
847+
red is not carried into a release.
848+
842849
### Documentation-only changes
843850

844851
A pull request or a push to `main` whose change is documentation only runs

‎docs/pages/contributing-llm.md‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -133,7 +133,7 @@ finding to report, not a step to take.
133133

134134
10. **A test row never silently disappears.** Each profile counts the rows that
135135
*ran* against a floor: the `PRIVATE_FIXTURE_PROFILE_FLOORS`
136-
ci_verify.py:789 table. When rows land, re-measure the profile's own
136+
ci_verify.py:802 table. When rows land, re-measure the profile's own
137137
overlay-present run and edit that table in the same commit.
138138

139139
## The recipe for a lane
@@ -281,7 +281,7 @@ measurement that produced it, so a later change to it is visible as a change to
281281
the record, not as an edit to a literal.
282282

283283
**floor** — the minimum number of CTest rows a profile must actually run
284-
(the `PRIVATE_FIXTURE_PROFILE_FLOORS` ci_verify.py:789 table: with the private
284+
(the `PRIVATE_FIXTURE_PROFILE_FLOORS` ci_verify.py:802 table: with the private
285285
overlay present, the profile's full registered population; without it, that
286286
population minus the pinned private-fixture rows it registers as SKIP
287287
placeholders). It counts rows that ran, so a skipped row does not pad it.

‎docs/pages/native-engine.md‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -4613,7 +4613,7 @@ own features, not only the rows that happened to be source-free.
46134613
`python3 scripts/ci_verify.py kernel` is the profile that verifies the
46144614
kernel-only build (Release, live runner ON, tutorial OFF, source layer OFF);
46154615
CI runs it as the `kernel-only` job. The profile carries a **row floor**, its
4616-
number in the `PRIVATE_FIXTURE_PROFILE_FLOORS` ci_verify.py:789 table in
4616+
number in the `PRIVATE_FIXTURE_PROFILE_FLOORS` ci_verify.py:802 table in
46174617
`scripts/ci_verify.py`: the count the kernel-only CTest set is
46184618
expected to run with the private overlay present, minus the pinned
46194619
private-fixture rows of `tests/private_fixture_rows.txt` it registers as SKIP

‎scripts/test_ci_preflight.py‎

Lines changed: 8 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -174,11 +174,16 @@ def test_slow_exclusion_is_exactly_pr_matrix_kernel_and_sanitizers(self):
174174
self.assertIn(f'{job} must exclude exactly slow on PRs only, once',
175175
ci_workflow_findings(*changed))
176176
# A fourth excluding job, or a parity subset with any exclusion, fails.
177-
for job in ('preflight', 'corpus-parity-subset'):
177+
# `preflight` is a steps job; `corpus-parity-subset` is a reusable-workflow
178+
# call (`uses:` + `with:`, no `steps:`), so the flag goes into its `with:`.
179+
for job, before, after in (
180+
('preflight', ' steps:\n',
181+
f' steps:\n - run: python3 scripts/ci_verify.py debug {flag}\n'),
182+
('corpus-parity-subset', ' mode: subset\n',
183+
f' mode: subset\n note: {flag}\n')):
178184
with self.subTest(job=job):
179185
changed = original.copy()
180-
changed[0] = in_job(changed[0], job, ' steps:\n',
181-
f' steps:\n - run: python3 scripts/ci_verify.py debug {flag}\n')
186+
changed[0] = in_job(changed[0], job, before, after)
182187
findings = ci_workflow_findings(*changed)
183188
self.assertIn(f'{job} must keep its full population', findings)
184189
self.assertIn('CI must carry exactly three PR-only slow exclusion flags', findings)

0 commit comments

Comments
 (0)