You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
test(runtime): compare ?version= response bodies against ONE installed row (#17416)
CI's Test Core (2/6) failed §3 on a 1 ms timestamp difference. Cause measured:
`get()` built a fresh host per call, so the package was installed twice, and
`SchemaRegistry.installPackage` stamps `installedAt`/`updatedAt` from one
`new Date()` per install. Two hosts therefore hold two rows whose stamps differ
whenever the installs straddle a millisecond boundary — a flake, not a product
defect: nothing on the read path reads a clock (`toPackageResponse` is an
allowlist copy, `withWritableVerdict` a spread, and `success()` adds no
timestamp), so one install cannot produce two answers.
Every case that compares two response bodies now issues both requests through
one host, which is the shape §2's byte-identical pin already had and the reason
it stayed green in the same shard. §1's pair moves onto one host too: the
criterion is «the SAME request with and without `?version=`», and two hosts let
a difference come from the rows instead of from the parameter. §4's
array-vs-string pin carried the identical latent flake and is repaired with it.
Whole-body `toEqual` is preserved everywhere — the stamps stay in the
comparison, which is what makes "the same request" mean the same response.
Claude-Session: https://claude.ai/code/session_01TSf4DV7ziu4V5j73e46b7c
Co-authored-by: Claude <noreply@anthropic.com>
0 commit comments