rob thijssen
452d7d9b3d
All checks were successful
CI / CUDA type-check (push) Successful in 1m32s
CI / Format (push) Successful in 57s
CI / Clippy (push) Successful in 2m51s
CI / Test (push) Successful in 8m0s
CI / Build cortex SRPM (push) Has been skipped
CI / Build neuron SRPM (push) Has been skipped
CI / Publish cortex to COPR (push) Has been skipped
CI / Publish neuron to COPR (push) Has been skipped
CI / Bump version in source (push) Has been skipped
Ships the mesh authority as an RPM on the same prebuilt-binary pipeline as
helexa-bench.
- rpm/helexa-upstream-prerelease.spec: wraps the CI-built binary; installs
the systemd unit, sysusers, firewalld service (tcp/8090), and
/etc/helexa-upstream/helexa-upstream.toml (config noreplace).
- data/helexa-upstream.{service,sysusers.conf,firewalld.xml}: Type=simple
unit (serve --config), dedicated system user with a StateDirectory home,
inbound 8090 (/authz/v1 + /web/v1). PostgreSQL is reached outbound; the
schema migrates on startup.
- build-prerelease.yml: build-upstream + package-upstream jobs with
change-detection over crates/helexa-upstream/ (UPSTREAM_RE), gated into
publish. SQLX_OFFLINE=true is set defensively — helexa-upstream uses the
sqlx runtime query API (no compile-time macros), so it builds with no
database and no .sqlx cache; DB integration tests stay gated behind
UPSTREAM_TEST_DATABASE_URL.
Validated: workflow YAML parses, rpmspec expands, and
`SQLX_OFFLINE=true cargo build --release -p helexa-upstream` succeeds.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F6o3ddqmYNh9kzdwq6eowh