lizhimins commented on PR #5649:
URL:
https://github.com/apache/rocketmq-dashboard/pull/5649#issuecomment-6446514621
Merged into `rocketmq-studio` as `aaf1d367c`.
This PR carried the batch: we consolidated 3 of your related fixes (#5649,
#5297, #6106) into a single commit rather than landing 3 near-identical ones on
the trunk. Every change was cherry-picked onto the current baseline with you
kept as the author, and your tests came along with it. The commit body
describes each part and says which lines we added while landing the batch -
notably #5297's test fixture, which needed `nameserverId` after #7353 made it
mandatory on the APACHE create path.
All four `-1` sentinel leak surfaces are closed: the AI contract records
document the sentinel in structured output (the same split #5291 applied),
`message.yaml` declares `minimum: -1` with a description on all three size
properties, `format.ts` gained `formatOptionalBytes` next to the existing
`formatOptionalTimeMs` and is wired into the message table, the detail drawer
and QueueBrowser, and docs/api-spec.md states that -1 means the vendor did not
report a size.
Verified on the combined branch after rebasing onto the current trunk: 4161
backend tests, 2019 frontend tests across 166 files, the production build and
type check, license gate 8/8, eslint with no new warnings, and make
catalog-verify / go build / go test / go vet / make fmt all clean. Both fixes
are mutation-sensitive: dropping the `isUniqueInstanceNameViolation`
translation fails
`createInstanceTranslatesUniqueKeyViolationToDuplicateNameTest`, and making
`formatOptionalBytes` ignore the sentinel fails two rendering tests.
One note on CI: the Frontend Build job failed twice on this PR with two
ConsumerPage tests ("keeps the latest client stack when an older request
resolves last" and "highlights inconsistent subscriptions and refreshes the
check result"). Neither file is in this change set, the full suite passes
locally, and the same test fails intermittently on unrelated PRs - run
40471035645 for a DLQ CSV change hit the identical failure. We are tracking it
as a pre-existing flaky test, not as something this batch introduced.
The other 2 PRs are closed with a note pointing here. Thanks for the
carefully evidenced patches - the vendor-SDK取证 in particular made the
unknown-size case easy to defend.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
