26307512418
95%

Ran 22 May 2026 07:25PM UTC

Jobs 1

Files 54

Run time 1min

Badge

Embed ▾

Committed 22 May 2026 07:22PM UTC coverage: 95.893%. Remained the same

Build # 26307512418

Build Type

push

github

Committed by

web-flow

Commit Message

fix(tests): tolerate spawn/waiter reorder in backpressure chunk test (#147)

`first_chunk_delivers_before_slow_second_chunk_e2e` labeled the two
`tokio::spawn`'d batches as `first`/`second` and assumed `first` would
be assigned to chunk 1. But the client driver assigns waiters to chunks
in mpsc-arrival order, which under multi-threaded scheduling is
independent of spawn order — and under `cargo llvm-cov` instrumentation
the scheduling gap between spawn and first-poll widens enough that the
reverse assignment becomes common. When `first` was assigned chunk 2
(stalled), its 2s per-RPC `Endpoint::timeout` fired and the gRPC call
returned `Cancelled / "Timeout expired"`, panicking on
`.expect("first batch must succeed")` at ~2.03s.

Identify the chunk-1 (allowed) vs chunk-2 (stalled) batch by completion
order via `futures::future::select` instead of by spawn name. The
asserted property — one batch returns promptly, the other blocks until
release — is unchanged; only the way the test resolves which JoinHandle
holds which chunk has been made order-tolerant. No production code
changes.

Verified locally: 30/30 passes under `cargo llvm-cov` on the targeted
test, matching the failing CI environment.

Coverage Stats

7542 of 7865 relevant lines covered (95.89%)

698889.28 hits per line

Jobs

ID	Job ID	Ran	Files	Coverage
1	26307512418.1	22 May 2026 07:25PM UTC	54	95.89	GitHub Action Run

prisma-risk / tsoracle / 26307512418
95%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 26307512418

prisma-risk / tsoracle / 26307512418 95%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 26307512418

prisma-risk / tsoracle / 26307512418
95%

README BADGES
x