mirror of
https://github.com/hanzoai/insights-go.git
synced 2026-09-30 22:46:12 +00:00
zeekayandHanzo Dev
d4a92a1907
github.com has no runner for hanzo-build-linux-amd64, so a caller under .github/workflows is a gate that can never be scheduled. This is the ~7-line caller on the plane that can (git.hanzo.ai git-runner fleet), pinned @v1. Co-authored-by: Hanzo Dev <[email protected]>
19 lines
868 B
YAML
19 lines
868 B
YAML
# hanzoai/insights-go — the Go client SDK for Insights. A library: ships from a
|
|
# tag as a Go module, builds no image, deploys nowhere.
|
|
#
|
|
# This is the repo's own `make test` target inlined (vet, then the suite with
|
|
# -race -count=1), not a call to make: hanzo.yml is read by the runner and by
|
|
# platform.hanzo.ai, and neither is promised a make binary.
|
|
#
|
|
# -race is not decoration here. The client batches captures on a background
|
|
# goroutine and evaluates feature flags from a polled cache, so every one of the
|
|
# 41 suites touches shared state across goroutines; a data race in flag
|
|
# evaluation returns the wrong variant to a caller and nothing ever reports it.
|
|
# -count=1 defeats the test cache so a green run means these tests actually ran.
|
|
test:
|
|
- name: vet-and-test
|
|
run: |
|
|
set -e
|
|
go vet ./...
|
|
go test -race -count=1 -timeout=10m ./...
|