| 1 | package console |
| 2 | |
| 3 | import ( |
| 4 | "context" |
| 5 | "encoding/json" |
| 6 | "errors" |
| 7 | "fmt" |
| 8 | "io" |
| 9 | "net/http" |
| 10 | "sync/atomic" |
| 11 | "testing" |
| 12 | "time" |
| 13 | |
| 14 | "github.com/carlosframework/platform/internal/objstore" |
| 15 | "github.com/carlosframework/platform/internal/release" |
| 16 | ) |
| 17 | |
| 18 | // statusArmed is enqueueArmed plus a handle on the engine, because every test |
| 19 | // here has to stage what the object's tag says. |
| 20 | func statusArmed(t *testing.T, code string) (*consoleFixture, *objstore.MemScanner, string, string) { |
| 21 | t.Helper() |
| 22 | fx, _, token, _ := storeArmed(t, code) |
| 23 | grantStore(t, fx, token) |
| 24 | requestScanning(t, fx, token, false) |
| 25 | sc := objstore.NewMemScanner() |
| 26 | fx.srv.cfg.StoreScanner = sc |
| 27 | enableDeploymentScanning(t, fx) |
| 28 | fx.srv.cfg.StoreOperators = []string{"alice@keymail.dev"} |
| 29 | if s, o := postJSON(t, scanURL(fx, "titogo")+"/grant", map[string]any{}, token); s != http.StatusOK { |
| 30 | t.Fatalf("scan grant = %d %v", s, o) |
| 31 | } |
| 32 | rec, _, err := objstore.Get(context.Background(), fx.rel, release.OpsAccount, "titogo") |
| 33 | if err != nil || rec == nil || rec.Provisioned == nil { |
| 34 | t.Fatalf("no provisioned record to read a bucket from: %v", err) |
| 35 | } |
| 36 | return fx, sc, appToken(t, fx, "titobot"), rec.Provisioned.Bucket |
| 37 | } |
| 38 | |
| 39 | // getStatus asks the door about one key and decodes the answer. |
| 40 | func getStatus(t *testing.T, fx *consoleFixture, tok, query string) (int, map[string]any) { |
| 41 | t.Helper() |
| 42 | status, _, body := rawReq(t, http.MethodGet, enqueueURL(fx, "titogo")+"?"+query, tok, "") |
| 43 | out := map[string]any{} |
| 44 | if body != "" { |
| 45 | if err := json.Unmarshal([]byte(body), &out); err != nil { |
| 46 | t.Fatalf("decode %q: %v", body, err) |
| 47 | } |
| 48 | } |
| 49 | return status, out |
| 50 | } |
| 51 | |
| 52 | // getStatusSafe is getStatus for a NON-TEST goroutine. rawReq calls t.Fatal on |
| 53 | // a transport error, and FailNow from any goroutine but the test's own is |
| 54 | // undefined behaviour that go vet's testinggoroutine check exists to catch. |
| 55 | func getStatusSafe(fx *consoleFixture, tok, query string) (map[string]any, error) { |
| 56 | req, err := http.NewRequest(http.MethodGet, enqueueURL(fx, "titogo")+"?"+query, nil) |
| 57 | if err != nil { |
| 58 | return nil, err |
| 59 | } |
| 60 | req.Header.Set("Authorization", "Bearer "+tok) |
| 61 | resp, err := http.DefaultClient.Do(req) |
| 62 | if err != nil { |
| 63 | return nil, err |
| 64 | } |
| 65 | defer resp.Body.Close() |
| 66 | raw, err := io.ReadAll(resp.Body) |
| 67 | if err != nil { |
| 68 | return nil, err |
| 69 | } |
| 70 | out := map[string]any{} |
| 71 | if len(raw) > 0 { |
| 72 | if err := json.Unmarshal(raw, &out); err != nil { |
| 73 | return nil, fmt.Errorf("decode %q: %w", raw, err) |
| 74 | } |
| 75 | } |
| 76 | return out, nil |
| 77 | } |
| 78 | |
| 79 | // seedObject puts an object in the engine's world and returns its version, so |
| 80 | // a test can stage a tag or a result against the same bytes. |
| 81 | func seedObject(sc *objstore.MemScanner, bucket, key, etag string) string { |
| 82 | mod := time.Now().Truncate(time.Second) |
| 83 | sc.PutObject(bucket, key, etag, mod, 10) |
| 84 | return objstore.ScanObject{Key: key, ETag: etag, Modified: mod, Size: 10}.Version() |
| 85 | } |
| 86 | |
| 87 | // TestStatusDoorReadsTagThenResultObject pins the ORDER, which is the whole of |
| 88 | // §5.2's read rule: the object's own tag is the more recent and more |
| 89 | // authoritative answer, and a result object exists only because a tag could |
| 90 | // not be written. |
| 91 | func TestStatusDoorReadsTagThenResultObject(t *testing.T) { |
| 92 | ctx := context.Background() |
| 93 | fx, sc, tok, bucket := statusArmed(t, "code-status-order") |
| 94 | |
| 95 | // (a) a tag and nothing else. |
| 96 | vTagged := seedObject(sc, bucket, "uploads/tagged.pdf", "etag-t") |
| 97 | sc.Put(bucket, "uploads/tagged.pdf", vTagged, objstore.VerdictClean) |
| 98 | if status, out := getStatus(t, fx, tok, "key=uploads/tagged.pdf"); status != http.StatusOK || out["state"] != "clean" { |
| 99 | t.Fatalf("tagged = %d %v, want 200 clean", status, out) |
| 100 | } |
| 101 | |
| 102 | // (b) a result object and no tag. |
| 103 | vResult := seedObject(sc, bucket, "uploads/untaggable.pdf", "etag-u") |
| 104 | if err := objstore.WriteScanResult(ctx, fx.rel, release.OpsAccount, "titogo", &objstore.ScanResult{ |
| 105 | Key: "uploads/untaggable.pdf", Verdict: "infected", Version: vResult, |
| 106 | Engine: "clamav", Signature: "Eicar-Test", At: time.Now().UTC(), |
| 107 | }); err != nil { |
| 108 | t.Fatal(err) |
| 109 | } |
| 110 | status, out := getStatus(t, fx, tok, "key=uploads/untaggable.pdf") |
| 111 | if status != http.StatusOK || out["state"] != "infected" { |
| 112 | t.Fatalf("result-object read = %d %v, want 200 infected", status, out) |
| 113 | } |
| 114 | if out["engine"] != "clamav" || out["signature"] != "Eicar-Test" { |
| 115 | t.Errorf("provenance missing from a result-object answer: %v", out) |
| 116 | } |
| 117 | // The app must be told the verdict is NOT on the object: an infected |
| 118 | // object answered this way is not covered by the enforcement Deny, which |
| 119 | // matches on a tag that is not there. |
| 120 | if out["recorded"] != false { |
| 121 | t.Errorf("recorded = %#v, want JSON false — the app cannot tell this verdict is not enforceable", out["recorded"]) |
| 122 | } |
| 123 | |
| 124 | // (c) BOTH, disagreeing. The tag wins. |
| 125 | vBoth := seedObject(sc, bucket, "uploads/both.pdf", "etag-b") |
| 126 | sc.Put(bucket, "uploads/both.pdf", vBoth, objstore.VerdictClean) |
| 127 | if err := objstore.WriteScanResult(ctx, fx.rel, release.OpsAccount, "titogo", &objstore.ScanResult{ |
| 128 | Key: "uploads/both.pdf", Verdict: "infected", Version: vBoth, |
| 129 | Engine: "clamav", At: time.Now().UTC(), |
| 130 | }); err != nil { |
| 131 | t.Fatal(err) |
| 132 | } |
| 133 | if status, out := getStatus(t, fx, tok, "key=uploads/both.pdf"); out["state"] != "clean" { |
| 134 | t.Fatalf("with both present = %d %v, want the TAG's answer (clean)", status, out) |
| 135 | } |
| 136 | } |
| 137 | |
| 138 | // TestStatusDoorRefusesStaleResultObject is the laundering bug as a |
| 139 | // correctness bug: v1 was judged, v2 was uploaded over it, and the recorded |
| 140 | // answer is about bytes that are no longer there. |
| 141 | func TestStatusDoorRefusesStaleResultObject(t *testing.T) { |
| 142 | ctx := context.Background() |
| 143 | fx, sc, tok, bucket := statusArmed(t, "code-status-stale") |
| 144 | |
| 145 | if err := objstore.WriteScanResult(ctx, fx.rel, release.OpsAccount, "titogo", &objstore.ScanResult{ |
| 146 | Key: "uploads/x.pdf", Verdict: "clean", Version: "etag-v1@2020-01-01T00:00:00Z", |
| 147 | Engine: "clamav", At: time.Now().UTC(), |
| 148 | }); err != nil { |
| 149 | t.Fatal(err) |
| 150 | } |
| 151 | // The object now holds DIFFERENT bytes. |
| 152 | seedObject(sc, bucket, "uploads/x.pdf", "etag-v2") |
| 153 | |
| 154 | status, out := getStatus(t, fx, tok, "key=uploads/x.pdf") |
| 155 | if status != http.StatusOK { |
| 156 | t.Fatalf("status = %d %v", status, out) |
| 157 | } |
| 158 | if out["state"] != "pending" { |
| 159 | t.Fatalf("state = %v, want pending — a verdict about replaced bytes is not an answer about this object", out["state"]) |
| 160 | } |
| 161 | } |
| 162 | |
| 163 | // TestStatusDoorReportsFailedAndDenied: as themselves, never as pending. |
| 164 | // |
| 165 | // The landed TAG READER maps both back to unknown so the sweep retries them, |
| 166 | // which is right as retry policy and wrong as an answer to a question. If this |
| 167 | // door applied that rule, a failed scan would read as pending forever — the |
| 168 | // drain deletes the item as handled, and nothing changes the answer until a |
| 169 | // sweep happens to re-fail it. |
| 170 | func TestStatusDoorReportsFailedAndDenied(t *testing.T) { |
| 171 | fx, sc, tok, bucket := statusArmed(t, "code-status-fd") |
| 172 | |
| 173 | v := seedObject(sc, bucket, "uploads/failed.pdf", "etag-f") |
| 174 | sc.Put(bucket, "uploads/failed.pdf", v, objstore.VerdictFailed) |
| 175 | if _, out := getStatus(t, fx, tok, "key=uploads/failed.pdf"); out["state"] != "failed" { |
| 176 | t.Errorf("state = %v, want failed", out["state"]) |
| 177 | } |
| 178 | |
| 179 | seedObject(sc, bucket, "uploads/denied.pdf", "etag-d") |
| 180 | sc.DenyHead(bucket, "uploads/denied.pdf") |
| 181 | if _, out := getStatus(t, fx, tok, "key=uploads/denied.pdf"); out["state"] != "denied" { |
| 182 | t.Errorf("state = %v, want denied", out["state"]) |
| 183 | } |
| 184 | |
| 185 | // And skipped is not clean. An app treating it as success reproduces the |
| 186 | // dishonesty the two-number rule exists to prevent. |
| 187 | vs := seedObject(sc, bucket, "uploads/huge.bin", "etag-h") |
| 188 | sc.Put(bucket, "uploads/huge.bin", vs, objstore.VerdictSkipped) |
| 189 | if _, out := getStatus(t, fx, tok, "key=uploads/huge.bin"); out["state"] != "skipped" { |
| 190 | t.Errorf("state = %v, want skipped", out["state"]) |
| 191 | } |
| 192 | } |
| 193 | |
| 194 | // TestStatusDoorGoneForMissingObject: an object that is not there is `gone`, |
| 195 | // which is distinct from both `pending` and `clean`. |
| 196 | func TestStatusDoorGoneForMissingObject(t *testing.T) { |
| 197 | fx, _, tok, _ := statusArmed(t, "code-status-gone") |
| 198 | |
| 199 | status, out := getStatus(t, fx, tok, "key=uploads/never-existed.pdf") |
| 200 | if status != http.StatusOK || out["state"] != "gone" { |
| 201 | t.Fatalf("missing object = %d %v, want 200 gone", status, out) |
| 202 | } |
| 203 | } |
| 204 | |
| 205 | // TestStatusDoorPendingForAnUnjudgedObject is the positive control for |
| 206 | // everything above: without it, a door that answered `pending` to absolutely |
| 207 | // everything would pass the stale test and the long-poll tests. |
| 208 | func TestStatusDoorPendingForAnUnjudgedObject(t *testing.T) { |
| 209 | fx, sc, tok, bucket := statusArmed(t, "code-status-pending") |
| 210 | seedObject(sc, bucket, "uploads/new.pdf", "etag-n") |
| 211 | |
| 212 | if _, out := getStatus(t, fx, tok, "key=uploads/new.pdf"); out["state"] != "pending" { |
| 213 | t.Fatalf("state = %v, want pending", out["state"]) |
| 214 | } |
| 215 | } |
| 216 | |
| 217 | // TestStatusDoorSharesTheConsentChain: the same table the enqueue door reads, |
| 218 | // so the two doors cannot answer one refusal differently — a disagreement is |
| 219 | // itself something an app could conclude from. |
| 220 | func TestStatusDoorSharesTheConsentChain(t *testing.T) { |
| 221 | fx, _, member, _ := storeArmed(t, "code-status-consent") |
| 222 | grantStore(t, fx, member) |
| 223 | fx.srv.cfg.StoreScanner = objstore.NewMemScanner() |
| 224 | enableDeploymentScanning(t, fx) |
| 225 | tok := appToken(t, fx, "bot") |
| 226 | |
| 227 | // ARMED TO FAIL ON ANY READ. MemScanner counts Scans, not Status calls, so |
| 228 | // "no bytes were read" cannot be asserted by counting — but an engine that |
| 229 | // errors on Status turns any read into a 502, which a 409 proves did not |
| 230 | // happen. That is what makes this a test of ORDER and not just of wording. |
| 231 | sc := fx.srv.cfg.StoreScanner.(*objstore.MemScanner) |
| 232 | sc.Fail("Status", errors.New("the engine must not have been consulted")) |
| 233 | |
| 234 | status, out := getStatus(t, fx, tok, "key=uploads/x.pdf") |
| 235 | if status != http.StatusConflict || out["reason"] != "not_requested" { |
| 236 | t.Fatalf("status door = %d %v, want the enqueue door's 409 not_requested", status, out) |
| 237 | } |
| 238 | if out["state"] != "unavailable" { |
| 239 | t.Errorf("state = %v, want unavailable", out["state"]) |
| 240 | } |
| 241 | |
| 242 | // And the reversible refusal, which is a different status code — one row |
| 243 | // proved nothing about the mapping being shared, only about one entry. |
| 244 | requestScanning(t, fx, member, false) |
| 245 | fx.srv.cfg.StoreOperators = []string{"alice@keymail.dev"} |
| 246 | if s, o := postJSON(t, scanURL(fx, "titogo")+"/grant", map[string]any{}, member); s != http.StatusOK { |
| 247 | t.Fatalf("scan grant = %d %v", s, o) |
| 248 | } |
| 249 | setDeploymentScanning(t, fx, false) |
| 250 | status, out = getStatus(t, fx, tok, "key=uploads/x.pdf") |
| 251 | if status != http.StatusServiceUnavailable || out["reason"] != "disabled" { |
| 252 | t.Fatalf("disabled = %d %v, want 503 disabled", status, out) |
| 253 | } |
| 254 | } |
| 255 | |
| 256 | // TestStatusDoorNeedsAKey and rejects a malformed wait. |
| 257 | func TestStatusDoorArgumentErrors(t *testing.T) { |
| 258 | fx, _, tok, _ := statusArmed(t, "code-status-args") |
| 259 | |
| 260 | if status, _ := getStatus(t, fx, tok, ""); status != http.StatusBadRequest { |
| 261 | t.Errorf("no key = %d, want 400", status) |
| 262 | } |
| 263 | if status, _ := getStatus(t, fx, tok, "key=uploads/x.pdf&wait=soon"); status != http.StatusBadRequest { |
| 264 | t.Errorf("malformed wait = %d, want 400", status) |
| 265 | } |
| 266 | if status, _ := getStatus(t, fx, tok, "key=uploads/x.pdf&wait=-5s"); status != http.StatusBadRequest { |
| 267 | t.Errorf("negative wait = %d, want 400", status) |
| 268 | } |
| 269 | } |
| 270 | |
| 271 | // TestScanWaitForClamps: an over-long wait is the client asking for the |
| 272 | // longest it can have, not an error to special-case. |
| 273 | func TestScanWaitForClamps(t *testing.T) { |
| 274 | for _, tc := range []struct { |
| 275 | in string |
| 276 | want time.Duration |
| 277 | bad bool |
| 278 | }{ |
| 279 | {in: "", want: 0}, |
| 280 | {in: "5s", want: 5 * time.Second}, |
| 281 | {in: "30s", want: scanWaitMax}, |
| 282 | {in: "5m", want: scanWaitMax}, |
| 283 | {in: "24h", want: scanWaitMax}, |
| 284 | {in: "soon", bad: true}, |
| 285 | {in: "-1s", bad: true}, |
| 286 | } { |
| 287 | got, err := scanWaitFor(tc.in) |
| 288 | if tc.bad { |
| 289 | if err == nil { |
| 290 | t.Errorf("scanWaitFor(%q) = %v, want an error", tc.in, got) |
| 291 | } |
| 292 | continue |
| 293 | } |
| 294 | if err != nil { |
| 295 | t.Errorf("scanWaitFor(%q): %v", tc.in, err) |
| 296 | continue |
| 297 | } |
| 298 | if got != tc.want { |
| 299 | t.Errorf("scanWaitFor(%q) = %v, want %v", tc.in, got, tc.want) |
| 300 | } |
| 301 | } |
| 302 | } |
| 303 | |
| 304 | // TestLongPollWakesOnDrain: a request parked on `wait` returns as soon as the |
| 305 | // drain records a verdict, rather than sitting out its window. |
| 306 | // |
| 307 | // The wait is long and the assertion is that the answer arrives EARLY, which |
| 308 | // is the only formulation that fails if the wake is never delivered — a test |
| 309 | // that merely checked the final answer would pass against a door that always |
| 310 | // slept the full duration. |
| 311 | func TestLongPollWakesOnDrain(t *testing.T) { |
| 312 | ctx := context.Background() |
| 313 | fx, sc, tok, bucket := statusArmed(t, "code-status-poll") |
| 314 | seedObject(sc, bucket, "uploads/x.pdf", "etag-x") |
| 315 | if _, err := objstore.Enqueue(ctx, fx.rel, fx.srv.scanQueue, release.OpsAccount, "titogo", |
| 316 | &objstore.QueueItem{Key: "uploads/x.pdf", By: "test"}, time.Now()); err != nil { |
| 317 | t.Fatal(err) |
| 318 | } |
| 319 | |
| 320 | type reply struct { |
| 321 | out map[string]any |
| 322 | took time.Duration |
| 323 | state string |
| 324 | } |
| 325 | done := make(chan reply, 1) |
| 326 | errs := make(chan error, 1) |
| 327 | go func() { |
| 328 | start := time.Now() |
| 329 | out, err := getStatusSafe(fx, tok, "key=uploads/x.pdf&wait=25s") |
| 330 | if err != nil { |
| 331 | errs <- err |
| 332 | return |
| 333 | } |
| 334 | st, _ := out["state"].(string) |
| 335 | done <- reply{out: out, took: time.Since(start), state: st} |
| 336 | }() |
| 337 | |
| 338 | // Give the request time to park before the drain runs, so the wake is |
| 339 | // what returns it rather than the first read. |
| 340 | time.Sleep(150 * time.Millisecond) |
| 341 | if err := fx.srv.scanDrainTick(ctx); err != nil { |
| 342 | t.Fatal(err) |
| 343 | } |
| 344 | |
| 345 | select { |
| 346 | case got := <-done: |
| 347 | if got.state != "clean" { |
| 348 | t.Fatalf("state = %v (%v), want clean", got.state, got.out) |
| 349 | } |
| 350 | if got.took > 20*time.Second { |
| 351 | t.Errorf("took %v against a 25s wait — it sat out the window rather than being woken", got.took) |
| 352 | } |
| 353 | case err := <-errs: |
| 354 | t.Fatalf("long-poll request: %v", err) |
| 355 | case <-time.After(20 * time.Second): |
| 356 | t.Fatal("the parked request was never woken by the drain") |
| 357 | } |
| 358 | } |
| 359 | |
| 360 | // TestLongPollReturnsPendingWhenNothingHappens: the window is a ceiling, not a |
| 361 | // promise, and it must end. |
| 362 | func TestLongPollReturnsPendingWhenNothingHappens(t *testing.T) { |
| 363 | fx, sc, tok, bucket := statusArmed(t, "code-status-timeout") |
| 364 | seedObject(sc, bucket, "uploads/quiet.pdf", "etag-q") |
| 365 | |
| 366 | start := time.Now() |
| 367 | status, out := getStatus(t, fx, tok, "key=uploads/quiet.pdf&wait=300ms") |
| 368 | if status != http.StatusOK || out["state"] != "pending" { |
| 369 | t.Fatalf("quiet wait = %d %v, want 200 pending", status, out) |
| 370 | } |
| 371 | if took := time.Since(start); took < 250*time.Millisecond { |
| 372 | t.Errorf("returned after %v — it did not wait the window it was given", took) |
| 373 | } |
| 374 | } |
| 375 | |
| 376 | // TestLongPollCapsWaiters covers BOTH caps, and the per-app one is not implied |
| 377 | // by the global one: without it a single tenant fills the global budget and |
| 378 | // every other tenant's wait degrades to polling. |
| 379 | func TestLongPollCapsWaiters(t *testing.T) { |
| 380 | fx, _, _, _ := statusArmed(t, "code-status-caps") |
| 381 | |
| 382 | // Per app: one app's own keys, up to its cap and then no further. |
| 383 | ref := objstore.AppRef{Account: release.OpsAccount, App: "titogo"} |
| 384 | var release1 []func() |
| 385 | for i := 0; i < scanWaitersPerApp; i++ { |
| 386 | _, rel, ok := fx.srv.scanWait(ref, string(rune('a'+i%26))+"/x") |
| 387 | if !ok { |
| 388 | t.Fatalf("refused waiter %d, below the per-app cap of %d", i, scanWaitersPerApp) |
| 389 | } |
| 390 | release1 = append(release1, rel) |
| 391 | } |
| 392 | if _, _, ok := fx.srv.scanWait(ref, "one/too/many"); ok { |
| 393 | t.Errorf("parked a waiter past the per-app cap of %d", scanWaitersPerApp) |
| 394 | } |
| 395 | // A DIFFERENT app is unaffected — which is the whole point of the per-app |
| 396 | // cap existing beside the global one. |
| 397 | other := objstore.AppRef{Account: release.OpsAccount, App: "otherapp"} |
| 398 | if _, rel, ok := fx.srv.scanWait(other, "x"); !ok { |
| 399 | t.Errorf("one app at its cap blocked another app entirely") |
| 400 | } else { |
| 401 | rel() |
| 402 | } |
| 403 | for _, rel := range release1 { |
| 404 | rel() |
| 405 | } |
| 406 | // Releasing gives the capacity back rather than burning it. |
| 407 | if _, rel, ok := fx.srv.scanWait(ref, "after/release"); !ok { |
| 408 | t.Error("capacity was not returned when the waiters were released") |
| 409 | } else { |
| 410 | rel() |
| 411 | } |
| 412 | |
| 413 | // Global: enough distinct apps that no per-app cap is ever reached. |
| 414 | var release2 []func() |
| 415 | for i := 0; ; i++ { |
| 416 | app := "app" + string(rune('A'+i%26)) + string(rune('A'+(i/26)%26)) |
| 417 | r := objstore.AppRef{Account: release.OpsAccount, App: app} |
| 418 | ch, rel, ok := fx.srv.scanWait(r, "k") |
| 419 | if !ok { |
| 420 | break |
| 421 | } |
| 422 | _ = ch |
| 423 | release2 = append(release2, rel) |
| 424 | if i > scanWaitersGlobal+8 { |
| 425 | t.Fatalf("parked %d waiters with a global cap of %d", i, scanWaitersGlobal) |
| 426 | } |
| 427 | } |
| 428 | if len(release2) != scanWaitersGlobal { |
| 429 | t.Errorf("global cap admitted %d waiters, want %d", len(release2), scanWaitersGlobal) |
| 430 | } |
| 431 | for _, rel := range release2 { |
| 432 | rel() |
| 433 | } |
| 434 | } |
| 435 | |
| 436 | // TestReleaseIsIdempotent: the door defers the release and a caller may call it |
| 437 | // early. A second call that decremented again would leak capacity downwards |
| 438 | // until the console stopped parking anything at all. |
| 439 | func TestReleaseIsIdempotent(t *testing.T) { |
| 440 | fx, _, _, _ := statusArmed(t, "code-status-idem") |
| 441 | ref := objstore.AppRef{Account: release.OpsAccount, App: "titogo"} |
| 442 | |
| 443 | _, rel, ok := fx.srv.scanWait(ref, "k") |
| 444 | if !ok { |
| 445 | t.Fatal("idle console refused a waiter") |
| 446 | } |
| 447 | rel() |
| 448 | rel() |
| 449 | rel() |
| 450 | |
| 451 | fx.srv.scanWaits.mu.Lock() |
| 452 | total := fx.srv.scanWaits.total |
| 453 | fx.srv.scanWaits.mu.Unlock() |
| 454 | if total != 0 { |
| 455 | t.Errorf("total = %d after releasing one waiter three times, want 0", total) |
| 456 | } |
| 457 | } |
| 458 | |
| 459 | // TestStatusDoorAnswersWhenTheWaitCapIsFull: over the cap nothing is refused. |
| 460 | // The read still happens and still answers truthfully; only the waiting is |
| 461 | // skipped, so a full console degrades to polling rather than to errors. |
| 462 | func TestStatusDoorAnswersWhenTheWaitCapIsFull(t *testing.T) { |
| 463 | fx, sc, tok, bucket := statusArmed(t, "code-status-full") |
| 464 | // PENDING, not clean. The first version staged a CLEAN object, so the door |
| 465 | // answered on its first read and returned before the wait mattered at all |
| 466 | // — it passed against a door that ignored the caps entirely AND against |
| 467 | // one that parked forever when refused. The object has to be one the door |
| 468 | // would WANT to wait for. |
| 469 | seedObject(sc, bucket, "uploads/x.pdf", "etag-x") |
| 470 | |
| 471 | ref := objstore.AppRef{Account: release.OpsAccount, App: "titogo"} |
| 472 | var rels []func() |
| 473 | for i := 0; i < scanWaitersPerApp; i++ { |
| 474 | if _, rel, ok := fx.srv.scanWait(ref, "filler/"+string(rune('a'+i%26))); ok { |
| 475 | rels = append(rels, rel) |
| 476 | } |
| 477 | } |
| 478 | defer func() { |
| 479 | for _, rel := range rels { |
| 480 | rel() |
| 481 | } |
| 482 | }() |
| 483 | |
| 484 | start := time.Now() |
| 485 | status, out := getStatus(t, fx, tok, "key=uploads/x.pdf&wait=10s") |
| 486 | if status != http.StatusOK || out["state"] != "pending" { |
| 487 | t.Fatalf("over the cap = %d %v, want the truthful 200 pending", status, out) |
| 488 | } |
| 489 | if took := time.Since(start); took > 3*time.Second { |
| 490 | t.Fatalf("took %v against a 10s wait — a request that cannot park must answer at once, not park anyway", took) |
| 491 | } |
| 492 | |
| 493 | // THE CONTROL, without which "answered immediately" proves nothing: under |
| 494 | // the cap, the same request on the same pending object waits. |
| 495 | for _, rel := range rels { |
| 496 | rel() |
| 497 | } |
| 498 | rels = nil |
| 499 | start = time.Now() |
| 500 | if _, out := getStatus(t, fx, tok, "key=uploads/x.pdf&wait=400ms"); out["state"] != "pending" { |
| 501 | t.Fatalf("under the cap = %v, want pending", out) |
| 502 | } |
| 503 | if took := time.Since(start); took < 350*time.Millisecond { |
| 504 | t.Errorf("under the cap it returned after %v — it did not park, so the test above says nothing", took) |
| 505 | } |
| 506 | } |
| 507 | |
| 508 | // TestLongPollSurvivesADrainDuringItsFirstRead runs a whole drain INSIDE the |
| 509 | // door's first Status call and requires a prompt, correct answer. |
| 510 | // |
| 511 | // WHAT IT DOES NOT PROVE, stated because the obvious reading of it is wrong: it |
| 512 | // does not pin the register-before-read ORDER. That was the intent, and a |
| 513 | // mutation that moves registration after the first read SURVIVES it. The |
| 514 | // reason is that the hook fires at the top of Status, so by the time the read |
| 515 | // returns the drain has already tagged the object and the door answers "clean" |
| 516 | // from the read itself — in both orderings, without the wake mattering at all. |
| 517 | // |
| 518 | // The ordering therefore remains unasserted, and the comment on the door says |
| 519 | // so. What this does pin is that a drain landing mid-read cannot make the door |
| 520 | // lose the answer or hang, which is worth having on its own. |
| 521 | func TestLongPollSurvivesADrainDuringItsFirstRead(t *testing.T) { |
| 522 | ctx := context.Background() |
| 523 | fx, sc, tok, bucket := statusArmed(t, "code-status-order2") |
| 524 | seedObject(sc, bucket, "uploads/x.pdf", "etag-x") |
| 525 | if _, err := objstore.Enqueue(ctx, fx.rel, fx.srv.scanQueue, release.OpsAccount, "titogo", |
| 526 | &objstore.QueueItem{Key: "uploads/x.pdf", By: "test"}, time.Now()); err != nil { |
| 527 | t.Fatal(err) |
| 528 | } |
| 529 | |
| 530 | // FIRE ONCE, and with a compare-and-swap rather than sync.Once: the drain |
| 531 | // this hook starts calls Status itself, so the hook re-enters — and |
| 532 | // Once.Do is not reentrant, it DEADLOCKS the second caller against the |
| 533 | // first. A CAS lets the inner call see "already fired" and return. |
| 534 | var fired atomic.Bool |
| 535 | sc.OnStatus(func() { |
| 536 | if !fired.CompareAndSwap(false, true) { |
| 537 | return |
| 538 | } |
| 539 | if err := fx.srv.scanDrainTick(ctx); err != nil { |
| 540 | t.Errorf("drain inside the door's read: %v", err) |
| 541 | } |
| 542 | }) |
| 543 | |
| 544 | start := time.Now() |
| 545 | status, out := getStatus(t, fx, tok, "key=uploads/x.pdf&wait=20s") |
| 546 | took := time.Since(start) |
| 547 | if status != http.StatusOK { |
| 548 | t.Fatalf("status = %d %v", status, out) |
| 549 | } |
| 550 | if out["state"] != "clean" { |
| 551 | t.Fatalf("state = %v, want clean — the drain ran during the first read", out["state"]) |
| 552 | } |
| 553 | if took > 15*time.Second { |
| 554 | t.Errorf("took %v against a 20s wait — a drain during the read must not cost the whole window", took) |
| 555 | } |
| 556 | } |
| 557 | |
amadan