Files
felhom-controller/controller/internal/infra/images_test.go
T
admin b5d78d1e0f v0.147.0 — feedback slice 1: pressing a button says something
The systemic complaint, twice in one evening: you press a button and nothing
happens. No progress, no ETA, no named result. Three worst offenders, fixed on
the two patterns already here (deploy 3-step panel, storage-init status poll).
No new framework — that is a ROADMAP item; three targeted cards ship tonight.

4a — a verification restore names its result. The flash said the app had been
restored "to a verification folder on the drive"; which folder, on which drive,
was invisible, so the customer could not go and look at what they had just asked
for. Full path now. The restore page gained a listing of existing verification
copies (app, size, date, path) — nothing anywhere showed these, so they piled up
and the only way to find them was SSH — each with a double-confirmed delete.

That delete is the only one this release adds, so it names a STACK, never a
path: the Manager resolves the name inside a backups/offsite-restore root it
computed itself and refuses anything landing outside. Red-proofed — neutralise
the name guard and stack:"" resolves to the offsite-restore ROOT and takes every
copy with it. Refusals are asserted as non-effects.

4b — Megosztás enable shows what it is waiting for. Enabling ran ReconcileSamba
synchronously inside the POST handler; on a golden without felhom-samba baked
that is compose pulling ~100MB, i.e. minutes of an apparently-hung form post
followed by "Beállítás mentve." whether or not anything came up. Detached +
polled now, distinguishing "képfájl letöltése" from "indítás" — decided BEFORE
the work starts, since afterwards the image is always present. Success is
probed, not inferred (compose up -d exits 0 on a crash-loop). The password form
starts the same job: with UserSet false reconcile deploys nothing, so on a fresh
box that is where the pull actually happens.

4c — "Távoli mentés most" streams real progress. restic was already reporting
bytes and percent; the runner seam used CombinedOutput() and discarded them. The
manual run now passes --json and scans stdout line-by-line: total bytes, percent,
current app. Manual only — the nightly stays silent, pinned by a test that fails
if it ever passes --json. The poll now arms unconditionally, closing a race the
manual trigger always ran: the redirect rendered before the goroutine wrote
LastStatus=running, so the poll never armed and the page sat static during the
very run just started. Red-proofed twice.

Also closes the golden/controller infra-image drift at the source: infra.Images()
derives from the existing pins and --print-infra-images exposes it, so the golden
bake can stop carrying its own copy. That copy had already drifted — felhom-samba
was never added, so the golden baked 3 of 4, which is why enabling Megosztás
pulled at runtime in the first place.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Nn3VgQk9iwEGgyx6QJ2NvE
2026-07-19 09:30:30 +02:00

87 lines
2.7 KiB
Go

package infra
import (
"go/ast"
"go/parser"
"go/token"
"strconv"
"strings"
"testing"
)
// TestImagesCoversEveryPin is the anti-drift gate for the golden bake.
//
// Images() feeds `felhom-controller --print-infra-images`, which build-golden.sh uses to decide what
// to pre-pull into the appliance image. If someone adds a fifth infra stack — a new `FooImage` const
// — and forgets to add it to Images(), the golden silently bakes 4 of 5 and enabling that stack on a
// fresh box goes back to being a multi-minute silent registry pull. That is exactly how felhom-samba
// was missed, so this test reads the CONST BLOCK OUT OF THE SOURCE rather than restating the list:
// a hand-written expected list would need the same edit and would rot the same way.
func TestImagesCoversEveryPin(t *testing.T) {
fset := token.NewFileSet()
f, err := parser.ParseFile(fset, "infra.go", nil, 0)
if err != nil {
t.Fatalf("parsing infra.go: %v", err)
}
pins := map[string]string{} // const name -> image ref
for _, decl := range f.Decls {
gd, ok := decl.(*ast.GenDecl)
if !ok || gd.Tok != token.CONST {
continue
}
for _, spec := range gd.Specs {
vs, ok := spec.(*ast.ValueSpec)
if !ok || len(vs.Names) != 1 || len(vs.Values) != 1 {
continue
}
name := vs.Names[0].Name
if !strings.HasSuffix(name, "Image") {
continue
}
lit, ok := vs.Values[0].(*ast.BasicLit)
if !ok || lit.Kind != token.STRING {
continue
}
val, err := strconv.Unquote(lit.Value)
if err != nil {
t.Fatalf("unquoting %s: %v", name, err)
}
pins[name] = val
}
}
if len(pins) == 0 {
t.Fatal("found no *Image consts in infra.go — the parser walk is broken, not the pins")
}
got := map[string]bool{}
for _, img := range Images() {
got[img] = true
}
for name, img := range pins {
if !got[img] {
t.Errorf("const %s = %q is not in Images() — the golden bake would not pre-pull it, so "+
"enabling that stack on a fresh box would block on a silent registry pull", name, img)
}
}
if len(Images()) != len(pins) {
t.Errorf("Images() has %d entries but infra.go declares %d *Image consts (%v) — they must "+
"correspond one-to-one", len(Images()), len(pins), pins)
}
}
// TestImagesArePinned guards the other half of the contract: a floating tag would make the golden
// bake unreproducible (the bake and a later deploy could resolve the same name to different digests).
func TestImagesArePinned(t *testing.T) {
for _, img := range Images() {
if !strings.Contains(img, ":") {
t.Errorf("infra image %q has no tag — implicitly :latest", img)
continue
}
if strings.HasSuffix(img, ":latest") {
t.Errorf("infra image %q is :latest — a floating tag breaks reproducible golden bakes", img)
}
}
}