Files
kandev/apps/backend/internal/plugins/service_sync_test.go
T
João Salavisa e145d48f22 feat: add plugin system with grpc runtime and native ui plugins (#1742)
* feat: add plugin system with grpc runtime and native ui plugins

- go plugin backends ship as release tarballs (per-platform binaries, checksums) that
  kandev installs from url/upload or filesystem sideload+sync, then spawns and
  supervises via hashicorp/go-plugin (unix socket / loopback tcp+automtls)
- strict typed grpc contract (kandev.plugin.v1): DeliverEvent/InvokeTool/HandleWebhook
  plus a capability-gated host service (state, secrets, emit); author sdk in
  pkg/pluginsdk; packaging cli in cmd/plugin-pack
- native frontend bundles: window.registerKandevPlugin + registry/host loader with
  dynamic routes, nav items, slots and ws handlers; plugins settings page with
  install dialog, sync button and enable/disable/uninstall
- gated behind the plugins feature flag (prod off, dev/e2e on); covered by unit and
  integration tests plus playwright e2e against a real packaged fixture plugin

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(plugins): hydrate base app context on plugin routes

Plugin-owned SPA routes classify as RouteUnknown, so the boot payload shipped
only feature flags — no workspaces, workflows, or kanban context. Native plugin
UI (e.g. host.ui.TaskCreateDialog) then had an empty store. Hydrate the same
base context as the Home route for unknown SPA routes.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* feat(plugins): expose TaskCreateDialog on host.ui

Lets a native plugin open kandev's real create-task modal (prefilled via
initialValues) instead of POSTing directly.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(plugins): harden filesystem/URL sinks against CodeQL findings

Add sink-level guards on the plugin trust boundary: securejoin containment on
archive extraction (zip-slip), plugin-id validation before every fs_store path,
http/https scheme validation on install-from-URL, and non-interpolated ws-handler
error logging.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(plugins): address review findings (races, resource caps, lifecycle)

- serialize per-plugin startup to close a TOCTOU double-spawn race, and let a
  plugin restart after exhausting attempts
- bound plugin startup with an explicit 30s timeout
- reject backslash paths and cap decompressed bytes by actual size in pkgtar;
  only mark a package Signed when a verifier actually ran and passed
- delete plugin_state on uninstall; validate webhook keys against the manifest
  and cap webhook body size; restart an active plugin when a new version installs

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(plugins): sanitize plugin-id path sinks for CodeQL

Pin the plugin id to a single path segment with filepath.Base at recordPath/
configPath and pkgtar.Remove — a barrier the taint analysis recognizes.
Behaviour-identical for valid ids (already validated by safePluginID); clears
the go/path-injection alerts.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(plugins): inline plugin-id validation at path sinks for CodeQL

filepath.Base does not strip '..', so it is not a path-injection barrier. Call
the recognized safePluginID / id-pattern guard inline in Delete/GetConfig/
SetConfig and pkgtar.Remove — the same barrier that already cleared the Get sink.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(plugins): address code-review findings (backend)

- add capability-gating tests for the Host RPC service (was untested)
- close a stop-during-start orphan race so Disable/Uninstall can't leave a
  supervised subprocess behind
- on failed upgrade, remove only the new version dir (not all versions + data);
  restart the prior version
- validate manifest version and reject ids ending in .config (record-file alias)
- make Deliverer.Refresh non-blocking and cancel in-flight delivery on worker stop
- clamp out-of-range webhook status, skip-and-warn corrupt records in List,
  write records atomically, semver-aware sideload version pick

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(plugins): address code-review findings (frontend)

- wrap plugin routes + settings routes in a shared PluginErrorBoundary so a
  throwing plugin page shows a fallback instead of white-screening the SPA
- fix disable/re-enable in one session: keep the cached registration and
  re-run initialize instead of relying on a cached ES-module re-import
- prefix root-relative bundle/style URLs with apiBaseUrl (split-origin/Tauri)
- surface partial-install warnings via toast.warning instead of a bare success
- dedupe/remove injected style links on unload; gate PluginNavItems on the flag

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(plugins): serialize per-plugin lifecycle + guard config writes

- add a per-plugin keyed lock in Service so concurrent Enable/Disable/Install/
  Uninstall/UpdateConfig for one id are serialized (no status-machine race)
- take s.mu in SetConfig/GetConfig and write config atomically
- bound each plugin initialize() with a timeout so one hung bundle can't stall
  the rest of the boot load

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* test(plugins): split service_test.go to satisfy file-length lint

Move the install/URL test group into service_install_test.go; no logic change.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* ci: build and ship the e2e fixture plugin package to E2E shards

The plugin E2E specs' global-setup requires apps/backend/.build/kandev-plugin-e2e-1.0.0.tar.gz.
Build it in the E2E build job, upload it as an artifact, and download it in both
the regular and container shard jobs.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Kandev Agent <agent@kandev.dev>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-07-17 12:00:26 +01:00

320 lines
11 KiB
Go

package plugins
import (
"context"
"os"
"path/filepath"
goruntime "runtime"
"testing"
"time"
)
// writeManifestOnly writes just a manifest.yaml (a "dir sideload": a
// version directory dropped straight onto disk, with no {id}.yml record and
// none of the other package files pkgtar.Install would extract).
func writeManifestOnly(t *testing.T, pluginsDir, id, version string) string {
t.Helper()
versionDir := filepath.Join(pluginsDir, id, version)
if err := os.MkdirAll(versionDir, 0o755); err != nil {
t.Fatalf("MkdirAll: %v", err)
}
manifestYAML := "id: " + id + "\napi_version: 1\nversion: \"" + version + "\"\ndisplay_name: Sideloaded\n" +
"runtime:\n type: binary\n executables:\n " + goruntime.GOOS + "-" + goruntime.GOARCH + ": server/plugin\n"
if err := os.WriteFile(filepath.Join(versionDir, "manifest.yaml"), []byte(manifestYAML), 0o644); err != nil {
t.Fatalf("WriteFile manifest.yaml: %v", err)
}
return versionDir
}
// dropTarball writes a valid plugin tar.gz package directly under
// pluginsDir (simulating an operator copying a package file onto disk
// instead of using the install UI/API).
func dropTarball(t *testing.T, pluginsDir, id, filename string) string {
t.Helper()
dest := filepath.Join(pluginsDir, filename)
data := testPackage(t, id, "1.0.0", false).Bytes()
if err := os.WriteFile(dest, data, 0o644); err != nil {
t.Fatalf("WriteFile %s: %v", filename, err)
}
return dest
}
func TestServiceSync_RegistersDirSideloadAsDisabled(t *testing.T) {
svc, dir, _, rt := newTestServiceWithDir(t)
versionDir := writeManifestOnly(t, dir, "kandev-plugin-side", "1.0.0")
result, err := svc.Sync(context.Background())
if err != nil {
t.Fatalf("Sync() unexpected error: %v", err)
}
if len(result.Added) != 1 || result.Added[0] != "kandev-plugin-side" {
t.Fatalf("Sync().Added = %v, want [kandev-plugin-side]", result.Added)
}
rec, err := svc.Get("kandev-plugin-side")
if err != nil {
t.Fatalf("Get(): %v", err)
}
if rec.Status != StatusDisabled {
t.Fatalf("sideloaded record Status = %q, want %q", rec.Status, StatusDisabled)
}
if rec.InstallPath != versionDir {
t.Fatalf("sideloaded record InstallPath = %q, want %q", rec.InstallPath, versionDir)
}
if rec.Signed {
t.Fatal("sideloaded record Signed = true, want false (unverified sideload)")
}
if rt.Running("kandev-plugin-side") {
t.Fatal("Sync() must never spawn a disabled sideload")
}
}
func TestServiceSync_ManifestIDMismatchIsAnError(t *testing.T) {
svc, dir, _, _ := newTestServiceWithDir(t)
writeManifestOnly(t, dir, "wrong-id-inside", "1.0.0")
// Rename the directory so the on-disk id ("dir-id") does not match the
// manifest's own id ("wrong-id-inside").
if err := os.Rename(filepath.Join(dir, "wrong-id-inside"), filepath.Join(dir, "dir-id")); err != nil {
t.Fatalf("os.Rename: %v", err)
}
result, err := svc.Sync(context.Background())
if err != nil {
t.Fatalf("Sync() unexpected error: %v", err)
}
if len(result.Added) != 0 {
t.Fatalf("Sync().Added = %v, want none (mismatched manifest id)", result.Added)
}
if len(result.Errors) != 1 {
t.Fatalf("Sync().Errors = %v, want exactly one entry", result.Errors)
}
if _, err := svc.Get("dir-id"); err == nil {
t.Fatal("Get(\"dir-id\") succeeded, want not found (mismatched manifest must not be registered)")
}
}
func TestServiceSync_MultipleVersionDirsPicksGreatestAndSkipsOthers(t *testing.T) {
svc, dir, _, _ := newTestServiceWithDir(t)
writeManifestOnly(t, dir, "kandev-plugin-multi", "1.0.0")
writeManifestOnly(t, dir, "kandev-plugin-multi", "2.0.0")
result, err := svc.Sync(context.Background())
if err != nil {
t.Fatalf("Sync() unexpected error: %v", err)
}
if len(result.Added) != 1 || result.Added[0] != "kandev-plugin-multi" {
t.Fatalf("Sync().Added = %v, want [kandev-plugin-multi]", result.Added)
}
if len(result.Errors) != 1 {
t.Fatalf("Sync().Errors = %v, want exactly one skip entry for the 1.0.0 dir", result.Errors)
}
rec, err := svc.Get("kandev-plugin-multi")
if err != nil {
t.Fatalf("Get(): %v", err)
}
if rec.Version != "2.0.0" {
t.Fatalf("registered Version = %q, want %q (semver-greatest)", rec.Version, "2.0.0")
}
}
// TestServiceSync_MultipleVersionDirsPicksSemverGreatestNotLexical pins the
// fix for a plain lexical sort picking the wrong "latest" version: "9.0.0"
// sorts after "10.0.0" as a string, but 10.0.0 is the actually-newer
// semver.
func TestServiceSync_MultipleVersionDirsPicksSemverGreatestNotLexical(t *testing.T) {
svc, dir, _, _ := newTestServiceWithDir(t)
writeManifestOnly(t, dir, "kandev-plugin-semver", "9.0.0")
writeManifestOnly(t, dir, "kandev-plugin-semver", "10.0.0")
result, err := svc.Sync(context.Background())
if err != nil {
t.Fatalf("Sync() unexpected error: %v", err)
}
if len(result.Added) != 1 || result.Added[0] != "kandev-plugin-semver" {
t.Fatalf("Sync().Added = %v, want [kandev-plugin-semver]", result.Added)
}
rec, err := svc.Get("kandev-plugin-semver")
if err != nil {
t.Fatalf("Get(): %v", err)
}
if rec.Version != "10.0.0" {
t.Fatalf("registered Version = %q, want %q (semver-greatest, not lexically greatest)", rec.Version, "10.0.0")
}
}
func TestServiceSync_TarballInstallSucceedsAndDeletesFile(t *testing.T) {
svc, dir, _, rt := newTestServiceWithDir(t)
tarPath := dropTarball(t, dir, "kandev-plugin-drop", "kandev-plugin-drop-1.0.0.tar.gz")
result, err := svc.Sync(context.Background())
if err != nil {
t.Fatalf("Sync() unexpected error: %v", err)
}
if len(result.Installed) != 1 || result.Installed[0] != "kandev-plugin-drop" {
t.Fatalf("Sync().Installed = %v, want [kandev-plugin-drop]", result.Installed)
}
if !rt.Running("kandev-plugin-drop") {
t.Fatal("Sync() did not spawn the successfully installed tarball plugin")
}
rec, err := svc.Get("kandev-plugin-drop")
if err != nil {
t.Fatalf("Get(): %v", err)
}
if rec.Status != StatusActive {
t.Fatalf("installed-from-tarball record Status = %q, want %q", rec.Status, StatusActive)
}
if _, statErr := os.Stat(tarPath); !os.IsNotExist(statErr) {
t.Fatalf("dropped tarball still exists after successful install, stat err = %v", statErr)
}
}
func TestServiceSync_TarballValidationFailureLeavesFileAndAddsError(t *testing.T) {
svc, dir, _, _ := newTestServiceWithDir(t)
junkPath := filepath.Join(dir, "junk.tar.gz")
if err := os.WriteFile(junkPath, []byte("not a real gzip archive"), 0o644); err != nil {
t.Fatalf("WriteFile: %v", err)
}
result, err := svc.Sync(context.Background())
if err != nil {
t.Fatalf("Sync() unexpected error: %v", err)
}
if len(result.Installed) != 0 {
t.Fatalf("Sync().Installed = %v, want none", result.Installed)
}
if len(result.Errors) != 1 || result.Errors[0].Path != junkPath {
t.Fatalf("Sync().Errors = %v, want one entry for %q", result.Errors, junkPath)
}
if _, statErr := os.Stat(junkPath); statErr != nil {
t.Fatalf("corrupt tarball was removed, want it left in place: %v", statErr)
}
}
func TestServiceSync_MissingInstallPathSetsErrorAndStopsRuntime(t *testing.T) {
svc, _, rt := newTestService(t)
rec := installTestPlugin(t, svc, "kandev-plugin-vanished")
if err := os.RemoveAll(rec.InstallPath); err != nil {
t.Fatalf("RemoveAll(InstallPath): %v", err)
}
result, err := svc.Sync(context.Background())
if err != nil {
t.Fatalf("Sync() unexpected error: %v", err)
}
if len(result.Missing) != 1 || result.Missing[0] != "kandev-plugin-vanished" {
t.Fatalf("Sync().Missing = %v, want [kandev-plugin-vanished]", result.Missing)
}
if !rt.stopped("kandev-plugin-vanished") {
t.Fatal("Sync() did not stop the runtime process for a plugin with a missing install path")
}
got, err := svc.Get("kandev-plugin-vanished")
if err != nil {
t.Fatalf("Get(): %v", err)
}
if got.Status != StatusError {
t.Fatalf("Status after missing-install detection = %q, want %q", got.Status, StatusError)
}
}
// TestServiceSync_MissingInstallAlreadyErrorIsIdempotent proves the "direct
// status write if transition invalid" fallback: once a plugin is already
// StatusError, canTransition(error, error) is false (SetStatus rejects
// same-status "transitions"), so Sync must not itself fail or panic —
// running it again still reports the plugin as missing.
func TestServiceSync_MissingInstallAlreadyErrorIsIdempotent(t *testing.T) {
svc, _, _ := newTestService(t)
rec := installTestPlugin(t, svc, "kandev-plugin-vanished")
if err := os.RemoveAll(rec.InstallPath); err != nil {
t.Fatalf("RemoveAll(InstallPath): %v", err)
}
if _, err := svc.Sync(context.Background()); err != nil {
t.Fatalf("first Sync() unexpected error: %v", err)
}
result, err := svc.Sync(context.Background())
if err != nil {
t.Fatalf("second Sync() unexpected error: %v", err)
}
if len(result.Missing) != 1 || result.Missing[0] != "kandev-plugin-vanished" {
t.Fatalf("second Sync().Missing = %v, want [kandev-plugin-vanished]", result.Missing)
}
got, err := svc.Get("kandev-plugin-vanished")
if err != nil {
t.Fatalf("Get(): %v", err)
}
if got.Status != StatusError {
t.Fatalf("Status after second Sync() = %q, want %q", got.Status, StatusError)
}
}
// TestServiceSync_ConcurrentCallsAreSerialized proves Sync is guarded by a
// mutex: a second concurrent Sync call must not proceed until the first one
// has fully returned. Pauses the first call mid-Install (inside the fake
// runtime's Start), starts the competing call, proves it is blocked out,
// then releases the first call and confirms the second only completes
// afterward.
func TestServiceSync_ConcurrentCallsAreSerialized(t *testing.T) {
svc, dir, _, rt := newTestServiceWithDir(t)
dropTarball(t, dir, "kandev-plugin-first", "kandev-plugin-first-1.0.0.tar.gz")
started, release := rt.blockNextStart()
firstDone := make(chan *SyncResult, 1)
go func() {
result, err := svc.Sync(context.Background())
if err != nil {
t.Errorf("first Sync() unexpected error: %v", err)
}
firstDone <- result
}()
select {
case <-started:
case <-time.After(5 * time.Second):
t.Fatal("first Sync() never reached the blocked runtime Start call")
}
secondDone := make(chan *SyncResult, 1)
go func() {
result, err := svc.Sync(context.Background())
if err != nil {
t.Errorf("second Sync() unexpected error: %v", err)
}
secondDone <- result
}()
// The second call must still be blocked on the sync mutex: give it a
// bounded window to (incorrectly) complete, and fail if it does.
select {
case <-secondDone:
t.Fatal("second Sync() completed while the first was still in flight — Sync is not mutex-guarded")
case <-time.After(200 * time.Millisecond):
}
release()
var firstResult, secondResult *SyncResult
select {
case firstResult = <-firstDone:
case <-time.After(5 * time.Second):
t.Fatal("first Sync() never returned after release")
}
select {
case secondResult = <-secondDone:
case <-time.After(5 * time.Second):
t.Fatal("second Sync() never returned after the first released the mutex")
}
if len(firstResult.Installed) != 1 || firstResult.Installed[0] != "kandev-plugin-first" {
t.Fatalf("first Sync().Installed = %v, want [kandev-plugin-first]", firstResult.Installed)
}
// The tarball is already gone by the time the second call scans, so it
// must not be double-installed.
if len(secondResult.Installed) != 0 {
t.Fatalf("second Sync().Installed = %v, want none (tarball already consumed)", secondResult.Installed)
}
}