flyemoji
9896fe16c3
An out-of-band curl of the live Fill v3 endpoint contradicted two claims the previous commit shipped and surfaced a mis-tiering: - User-Agent is NOT enforced: fill.papermc.io/v3/projects/velocity returned HTTP 200 to a bare curl UA. The comments claimed a generic UA "is refused" and the API "REQUIRES" a contact UA. Reword to what is true — PaperMC's usage policy asks for a descriptive UA and may block generic ones, but sending it is etiquette/defensive here, not a gate Felis depends on. - The test fixture's shape was invented, not captured: the real "versions" object groups the entire 3.x line under a single key "3.0.0", not the per-minor keys the fixture used. Replace it with the real body (keys and version strings as returned). The key-agnostic parser already produced the right answer, and an independent max-stable check confirms 3.4.0. - Re-tier doc.go: the GitHub Releases source is verifiable-here (the same httptest-testable shape as PaperMC), not integration remainder. It is why 3 of 4 components report "latest unknown" today and is the next verifiable slice — the release-source work is only ~half done until it exists. No production logic changed. WSL oracle: build + vet clean, internal/updater 10/10, full tree go test RC=0 (19 ok, 0 fail).
40 lines
2.5 KiB
Go
40 lines
2.5 KiB
Go
// Package updater is the integration/caller side of the component self-update
|
|
// subsystem — it gives the pure decision core in internal/updates its first real
|
|
// caller. internal/updates declares the seams (ReleaseSource / Notifier / Applier)
|
|
// and orchestrates them (updates.Run) but performs no I/O; this package supplies the
|
|
// concrete wiring: the platform topology (which components Felis tracks and under
|
|
// what policy), the upstream release sources, and the Runner that gathers current
|
|
// versions, runs the plan against a maintenance window, and renders the "版本号状态"
|
|
// report.
|
|
//
|
|
// Verification boundary — stated honestly so a green test suite is not mistaken for
|
|
// "the updater works against real infra":
|
|
//
|
|
// - BUILT + UNIT-VERIFIED (Go tests, WSL oracle): the topology, the PaperMC Fill
|
|
// v3 parser, the routing source, and the Runner's report-only composition. The
|
|
// PaperMC fixture is captured from the live endpoint, and the live call was
|
|
// exercised out-of-band on 2026-07-04 (curl fill.papermc.io/v3/projects/velocity):
|
|
// the response shape matches the fixture and 3.4.0 is confirmed the newest stable
|
|
// (3.5.0-SNAPSHOT correctly filtered). These prove the parse/plan/compose LOGIC
|
|
// and that the core is now wired to a caller.
|
|
//
|
|
// - NOT YET BUILT, but VERIFIABLE HERE (same technique as PaperMC — HTTP GET, JSON
|
|
// decode, tolerant Parse, prerelease filter, all httptest-testable): the GitHub
|
|
// Releases source. It is why 3 of the 4 components (felis-api, k3s, cloudflared)
|
|
// currently report "latest unknown" — RoutingSource returns errGitHubNotWired for
|
|
// them. This is the next VERIFIABLE slice, not integration remainder; until it
|
|
// exists the verifiable release-source work is only ~half done.
|
|
//
|
|
// - CAVEATS on what the tests do NOT prove: they run against httptest, not the live
|
|
// host, so future upstream shape drift is not caught; and while Felis sends a
|
|
// descriptive User-Agent (PaperMC etiquette), upstream UA enforcement was not
|
|
// active on the project endpoint on 2026-07-04 (a bare UA got HTTP 200), so the UA
|
|
// is defensive, not load-bearing.
|
|
//
|
|
// - REMAINING INTEGRATION (pure I/O, no verifiable-here logic): the concrete
|
|
// VersionGatherer (`k3s --version`, image-tag / jar inspection), the concrete
|
|
// Notifier (SMTP + in-game) and Applier (control-plane image bump, cloudflared
|
|
// swap), the `felis update` CLI + CronJob entry point, and the runtime append of
|
|
// the live Pinned Minecraft fleet.
|
|
package updater
|