Repository navigation
test(global-db): avoid unused raw fixture connections - #4897
Merged
ScriptedAlchemy merged 3 commits intoNov 15, 2026
Merged
Conversation
|
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
ScriptedAlchemy
enabled auto-merge
November 15, 2026 18:27
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why
The LCM schema fixtures still construct a raw
TestConnectionafter canonical registered initialization, then immediately discard it at 17 successful call sites. That starts an unused persistent writer and reader pool. The remaining callers only inspect initialization failures. This unnecessary work was found while investigating the global database unit-suite timeout on #4857.Changes
initialize_global_dbhelper and delete the redundantopen_global_dbwrapper.expect_errassertions at the two temporal-cursor refusal sites. These address the observed Clippy findings while preserving the same error values and messages.Against current master, the fixture cleanup changes six test files. Production code and execution limits are unchanged.
SDK failure-context follow-up
Master SDK conformance freshly passed 48/49 production cases. Its missing-workflow-definition check received a validated canonical HTTP 503 problem where it requires 404, but the failed assertion omitted that problem's details.
A separate test-only commit adds
{problem:?}to that existing assertion message. This exposes the already decoded code and envelope on failure, with the expected 404 and every following concealment, contract and retry assertion unchanged. It adds no requests or waits and does not identify or fix the 503 cause. This adds one SDK test file to the six fixture files above.Validation and current head
The current head
310bdd46integrates actual master97caca9f, the fixture correction and the SDK diagnostic. Its tree isd0dd3022494d2c0e7d8a9b3307fdaad7c741c4d7. Independent review reconstructed all 6,717 entries as the exact union of master, the six corrected fixture postimages and the one SDK assertion change. Pinned rustfmt, actual commitlint and diff checks passed.On previous head
92b2a787, ordinary CI passed all 180 Linux targets with 20 fresh executions, including global database unit tests freshly in 123.6 seconds. Its ordinary Windows job passed 467 build/typecheck targets and 26,371 actions. The only failed gate was the two Clippy findings corrected here. All prior-head jobs were terminal, and the complete Linux, Windows and Clippy logs were preserved and independently verified before the non-force update.Current-head ordinary CI passes Clippy. Linux job 132601737654 completed 179/180 targets, with 21 fresh executions; global database passed freshly in 96.8 seconds. The sole failed target was
hooks_lsp_suite(28/29): its receipt-backlog test still saw all 65 receipts after the existing 10-second bound. The restarted daemon log records readiness at 521 ms, but does not establish whether startup discovery, project admission or a drain pass completed. The same fixture and startup/replay sources produced the same 65-receipt symptom on the earlier diagnostic run 43631589275; the source comparison does not establish identical execution conditions or a cause.Windows job 132607685215 passed all 467 build/typecheck targets and 26,371 actions. The complete Linux and Windows logs are retained and independently hash-verified, along with the failing test artifact. Tested synthetic merge
dcdabf03has the exact current PR tree and parents actual master97caca9fplus310bdd46. The Linux failure still prevents qualification.The dedicated SDK conformance workflow is required to execute its production-router case with the configured
--include-ignoredselection; this workflow has master-push and manual-dispatch triggers only. It has not been dispatched for this head. Ordinary PR CI excludes the installed production case and does not supply that owning execution. The 503 cause remains unresolved.The fixture cleanup removes proven unused actor startup and teardown. It does not establish the cause or resolution of the observed 302-second timeout, and no measured speedup is claimed.