Repository navigation
docs(magnetar): refresh issue 462 broker evidence - #885
Merged
Merged
Conversation
Signed-off-by: Florentin Dubois <florentin.dubois@clever.cloud>
…stable-broker-images
This was referenced Nov 2, 2026
FlorentinDUBOIS
added a commit
that referenced
this pull request
Nov 2, 2026
PR #885 carries the per-image measurement and the reason-string rewrite that satisfy ADR-0126's conditions 3 and 4, and it had a failing check when this reconciliation merged, so the ADR landed first. Say so instead of claiming a sequencing that did not happen: conditions 3 and 4 are owed by the tree until #885 merges, and the carve-out does not excuse a reason string that still names an unpublished image. Signed-off-by: Florentin Dubois <florentin.dubois@clever.cloud>
ADR-0126 landed while this branch was open, so the three places that called the permit test's `#[ignore]` an unreconciled deviation are now wrong: the attribute is narrowly authorized, on five conditions, and what it owes is its removal. Point the test's doc comment, the docs/testing.md row and ADR-0116's amendment at ADR-0126 instead of describing a conflict that no longer exists. This branch is also what makes that carve-out conformant. ADR-0126's conditions 3 and 4 require the reason string to carry the measured counter-evidence and the deviation to be written down per image; the reason string here names `totalAvailablePermits=-2` with its reproduce command, and the evidence row carries the index digests plus the failing run's raw-log SHA-256. ADR-0126 said those two surfaces were owed by the tree until this PR merged; it now records them as satisfied. The measured content is unchanged: `4.0.14` passed 3/3 bounded runs, `4.2.5` passed run 1 and reached a minimum of -2 on run 2, `5.0.0` was not re-run, and the earlier 7/10 results predate PR #399 and are kept separate. The test stays ignored, and the document still refuses to infer a fix from merged upstream patches or a short run of passes. Signed-off-by: Florentin Dubois <florentin.dubois@clever.cloud>
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.
What changed
Refreshes issue #462's broker evidence with a measured, published-image sample, and — since ADR-0126 landed while this branch was open — reconciles the three places that still described the permit test's
#[ignore]as an unreconciled deviation.The measurement (unchanged from the original campaign)
Run from source commit
46792879ecde265c50ffaf10616396f14a894f7b, multi-platform image-index digests recorded:4.0.144.2.50107369d8a0b47b7b93e9254f13f994e4f67d5768989e2a51c8000720aa825a75.0.0The earlier 2026-10-12 matrix (PR #881) recorded 5/5 on
4.0.14and 7/10 on each of4.2.5and5.0.0with a residual −2; those predate PR #399 and are kept separate rather than pooled.The ignore stays. The document explicitly refuses to infer a fix from merged upstream patches or a short run of passes — which is now also ADR-0126's condition 2.
Reconciliation with ADR-0126
ADR-0126 authorizes this one
#[ignore]narrowly, on five conditions, so "a separately tracked deviation from ADR-0046" is no longer accurate. Updated in the test's doc comment, thedocs/testing.mdrow, and ADR-0116's 2026-10-16 amendment.This branch is also what makes that carve-out conformant. ADR-0126's conditions 3 and 4 require the reason string to carry the measured counter-evidence and the deviation to be written down per image. The reason string here names
totalAvailablePermits=-2with its reproduce command; the evidence row carries the digests and the failing log hash. ADR-0126 recorded those two surfaces as owed by the tree until this PR merged — it now records them as satisfied.Why the previous CI failure does not apply
The run on the old head failed one check:
e2e (Pulsar 5.0.0) / e2e 2/4, one2e_hardened_scalable_stream_consumer_contract— a PIP-460 scalable-topics test, unrelated to this docs diff. That run is from 2026-10-16, and the branch was 19 commits behindmain. It has been merged up to currentmain(41f97a8), so CI is re-running against today's tree.Validation
cargo +nightly fmt --all --check,cargo run -p xtask -- check-e2e-fixture(tag constants still agree with the CI pre-pulls,FIXTURE_TAGSand the PIP-33 compose pin), andcargo test -p magnetar-driver --test e2e_shared_subscription_permit_accounting --no-runall exit 0.prettier --checkon the edited markdown deviates only wheremainalready deviates (docs/testing.mdrows 8-13, pre-existing and deliberately untouched).The suite itself cannot run on the authoring host — no container runtime — so the measured numbers above are the original campaign's, not re-measured here.