Commit b2e4bc2
committed
ci: race-detector on -short, separate full pass without race
The combined `go test -race ./...` run blew past 30m on the runner —
TestHandleVKV3SolidityShape alone burns 51s+ under race because gnark's
G2 Setup batch-mul is goroutine-heavy and race instrumentation taxes
every step. That single test dragged the whole job past the deadline.
Split the test job into two passes:
- `go test -race -short -count=1 -timeout 15m ./...` — race detector
with the heavy ZK end-to-end tests skipped (they're already gated on
testing.Short() across internal/server/*_test.go and prover/*_test.go).
Locally ~50s; expect 5-10 min on the runner.
- `go test -count=1 -timeout 15m ./...` — full coverage without race
to keep functional regression coverage on the heavy paths.
Locally ~73s; expect 5-10 min on the runner.
Race detector still covers everything that actually races — HTTP
handlers, storage, parsing, in-memory state machine. The ZK solver
runs sequentially in the prover path, so racing it adds little signal
and a lot of wall clock.1 parent 2e600e2 commit b2e4bc2
1 file changed
Lines changed: 15 additions & 6 deletions
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
51 | 51 | | |
52 | 52 | | |
53 | 53 | | |
54 | | - | |
55 | | - | |
56 | | - | |
57 | | - | |
58 | | - | |
59 | | - | |
| 54 | + | |
| 55 | + | |
| 56 | + | |
| 57 | + | |
| 58 | + | |
| 59 | + | |
| 60 | + | |
| 61 | + | |
| 62 | + | |
| 63 | + | |
| 64 | + | |
| 65 | + | |
| 66 | + | |
| 67 | + | |
| 68 | + | |
60 | 69 | | |
61 | 70 | | |
62 | 71 | | |
| |||
0 commit comments