Files
multica/server/internal/daemon/gc_test.go
Bohan Jiang 8f10741a4d feat(daemon/gc): tighten GC defaults + flex duration suffix (#1559)
* feat(daemon/gc): tighten GC defaults + flex duration suffix

Driven by user feedback in #1539 (40 GB VPS filling within 24h of heavy
AI-coding usage): the existing TTLs were sized for desktop/laptop
deployments and are too lenient for small-disk, long-running daemons.

- GCTTL: 5d → 24h. Done/canceled issues almost never need a multi-day
  grace period in AI-coding workflows.
- GCOrphanTTL: 30d → 72h. Covers crash-leftover and pre-GC directories
  without a month-long wait.
- Issue-deleted orphans (API returns 404) are now cleaned on the next GC
  cycle regardless of mtime. The issue row is gone; there is nothing
  left to protect.
- parseFlexDuration: accept a `d` (day) suffix in addition to the stdlib
  time.ParseDuration syntax. MULTICA_GC_TTL=5d now works; previously only
  120h was accepted.

* fix(daemon/gc): address review — 404 safety + decimal/overflow in duration parser

Two issues flagged in PR review:

1. 404-immediate-clean is unsafe. The /gc-check endpoint returns 404 for
   both "issue deleted" AND "daemon token has no access to the workspace"
   (anti-enumeration, see requireDaemonWorkspaceAccess). Clean-on-404
   would let a scoped-down daemon token wipe taskDirs whose issues are
   still live. Restore the mtime gate against GCOrphanTTL. With the new
   72h default we still shrink the original 30d window dramatically
   without the cross-workspace hazard. Lock the behavior in with a new
   test that asserts a recent 404 is skipped.

2. parseFlexDuration mishandled decimals and swallowed Atoi errors:
   "0.5d" → 7m12s (regex matched only the "5d"), "1.5d" → 1h7m12s,
   and 20+ digit day values Atoi-errored silently to 0. Match the full
   decimal number with `\d*\.\d+|\d+` and parse with ParseFloat so
   fractional days and oversized inputs both go through
   time.ParseDuration correctly — fractions as sub-hour durations,
   overflow as a returned error.
2026-04-23 17:40:09 +08:00

329 lines
9.8 KiB
Go

package daemon
import (
"context"
"encoding/json"
"fmt"
"log/slog"
"net/http"
"net/http/httptest"
"os"
"path/filepath"
"testing"
"time"
"github.com/multica-ai/multica/server/internal/daemon/execenv"
)
// newGCTestDaemon creates a minimal Daemon for GC testing with a mock HTTP server.
func newGCTestDaemon(t *testing.T, handler http.Handler) *Daemon {
t.Helper()
srv := httptest.NewServer(handler)
t.Cleanup(srv.Close)
root := t.TempDir()
cfg := Config{
WorkspacesRoot: root,
GCEnabled: true,
GCInterval: 1 * time.Hour,
GCTTL: 5 * 24 * time.Hour,
GCOrphanTTL: 30 * 24 * time.Hour,
}
d := New(cfg, slog.Default())
d.client = NewClient(srv.URL)
d.client.SetToken("test-token")
return d
}
// createTaskDir creates a task directory with optional GC metadata.
func createTaskDir(t *testing.T, root, wsID, dirName string, meta *execenv.GCMeta) string {
t.Helper()
taskDir := filepath.Join(root, wsID, dirName)
if err := os.MkdirAll(taskDir, 0o755); err != nil {
t.Fatal(err)
}
if meta != nil {
data, _ := json.Marshal(meta)
if err := os.WriteFile(filepath.Join(taskDir, ".gc_meta.json"), data, 0o644); err != nil {
t.Fatal(err)
}
}
return taskDir
}
func TestShouldCleanTaskDir_DoneIssueOverTTL(t *testing.T) {
t.Parallel()
issueID := "11111111-1111-1111-1111-111111111111"
mux := http.NewServeMux()
mux.HandleFunc(fmt.Sprintf("/api/daemon/issues/%s/gc-check", issueID), func(w http.ResponseWriter, r *http.Request) {
w.Header().Set("Content-Type", "application/json")
json.NewEncoder(w).Encode(map[string]any{
"status": "done",
"updated_at": time.Now().Add(-10 * 24 * time.Hour), // 10 days ago
})
})
d := newGCTestDaemon(t, mux)
taskDir := createTaskDir(t, d.cfg.WorkspacesRoot, "ws1", "task1", &execenv.GCMeta{
IssueID: issueID,
WorkspaceID: "ws1",
CompletedAt: time.Now().Add(-10 * 24 * time.Hour),
})
action := d.shouldCleanTaskDir(context.Background(), taskDir)
if action != gcActionClean {
t.Fatalf("expected gcActionClean, got %d", action)
}
}
func TestShouldCleanTaskDir_CanceledIssueOverTTL(t *testing.T) {
t.Parallel()
issueID := "22222222-2222-2222-2222-222222222222"
mux := http.NewServeMux()
mux.HandleFunc(fmt.Sprintf("/api/daemon/issues/%s/gc-check", issueID), func(w http.ResponseWriter, r *http.Request) {
w.Header().Set("Content-Type", "application/json")
json.NewEncoder(w).Encode(map[string]any{
"status": "canceled",
"updated_at": time.Now().Add(-6 * 24 * time.Hour),
})
})
d := newGCTestDaemon(t, mux)
taskDir := createTaskDir(t, d.cfg.WorkspacesRoot, "ws1", "task2", &execenv.GCMeta{
IssueID: issueID,
WorkspaceID: "ws1",
CompletedAt: time.Now(),
})
action := d.shouldCleanTaskDir(context.Background(), taskDir)
if action != gcActionClean {
t.Fatalf("expected gcActionClean, got %d", action)
}
}
func TestShouldCleanTaskDir_OpenIssueSkipped(t *testing.T) {
t.Parallel()
issueID := "33333333-3333-3333-3333-333333333333"
mux := http.NewServeMux()
mux.HandleFunc(fmt.Sprintf("/api/daemon/issues/%s/gc-check", issueID), func(w http.ResponseWriter, r *http.Request) {
w.Header().Set("Content-Type", "application/json")
json.NewEncoder(w).Encode(map[string]any{
"status": "in_progress",
"updated_at": time.Now().Add(-30 * 24 * time.Hour),
})
})
d := newGCTestDaemon(t, mux)
taskDir := createTaskDir(t, d.cfg.WorkspacesRoot, "ws1", "task3", &execenv.GCMeta{
IssueID: issueID,
WorkspaceID: "ws1",
CompletedAt: time.Now(),
})
action := d.shouldCleanTaskDir(context.Background(), taskDir)
if action != gcActionSkip {
t.Fatalf("expected gcActionSkip for open issue, got %d", action)
}
}
func TestShouldCleanTaskDir_DoneButRecentSkipped(t *testing.T) {
t.Parallel()
issueID := "44444444-4444-4444-4444-444444444444"
mux := http.NewServeMux()
mux.HandleFunc(fmt.Sprintf("/api/daemon/issues/%s/gc-check", issueID), func(w http.ResponseWriter, r *http.Request) {
w.Header().Set("Content-Type", "application/json")
json.NewEncoder(w).Encode(map[string]any{
"status": "done",
"updated_at": time.Now().Add(-1 * 24 * time.Hour), // 1 day ago, within TTL
})
})
d := newGCTestDaemon(t, mux)
taskDir := createTaskDir(t, d.cfg.WorkspacesRoot, "ws1", "task4", &execenv.GCMeta{
IssueID: issueID,
WorkspaceID: "ws1",
CompletedAt: time.Now(),
})
action := d.shouldCleanTaskDir(context.Background(), taskDir)
if action != gcActionSkip {
t.Fatalf("expected gcActionSkip for recently-done issue, got %d", action)
}
}
func TestShouldCleanTaskDir_NoMetaRecentSkipped(t *testing.T) {
t.Parallel()
d := newGCTestDaemon(t, http.NewServeMux())
// No meta, fresh directory — should skip.
taskDir := createTaskDir(t, d.cfg.WorkspacesRoot, "ws1", "task5", nil)
action := d.shouldCleanTaskDir(context.Background(), taskDir)
if action != gcActionSkip {
t.Fatalf("expected gcActionSkip for recent orphan, got %d", action)
}
}
func TestShouldCleanTaskDir_NoMetaOldOrphan(t *testing.T) {
t.Parallel()
d := newGCTestDaemon(t, http.NewServeMux())
d.cfg.GCOrphanTTL = 0 // treat all orphans as expired
taskDir := createTaskDir(t, d.cfg.WorkspacesRoot, "ws1", "task6", nil)
action := d.shouldCleanTaskDir(context.Background(), taskDir)
if action != gcActionOrphan {
t.Fatalf("expected gcActionOrphan, got %d", action)
}
}
func TestShouldCleanTaskDir_APIErrorSkipped(t *testing.T) {
t.Parallel()
issueID := "55555555-5555-5555-5555-555555555555"
mux := http.NewServeMux()
mux.HandleFunc(fmt.Sprintf("/api/daemon/issues/%s/gc-check", issueID), func(w http.ResponseWriter, r *http.Request) {
w.WriteHeader(http.StatusInternalServerError)
})
d := newGCTestDaemon(t, mux)
taskDir := createTaskDir(t, d.cfg.WorkspacesRoot, "ws1", "task7", &execenv.GCMeta{
IssueID: issueID,
WorkspaceID: "ws1",
CompletedAt: time.Now(),
})
action := d.shouldCleanTaskDir(context.Background(), taskDir)
if action != gcActionSkip {
t.Fatalf("expected gcActionSkip on API error, got %d", action)
}
}
func TestShouldCleanTaskDir_Issue404OldOrphan(t *testing.T) {
t.Parallel()
issueID := "66666666-6666-6666-6666-666666666666"
mux := http.NewServeMux()
mux.HandleFunc(fmt.Sprintf("/api/daemon/issues/%s/gc-check", issueID), func(w http.ResponseWriter, r *http.Request) {
w.WriteHeader(http.StatusNotFound)
w.Write([]byte(`{"error":"issue not found"}`))
})
d := newGCTestDaemon(t, mux)
d.cfg.GCOrphanTTL = 0 // treat orphans as immediately eligible
taskDir := createTaskDir(t, d.cfg.WorkspacesRoot, "ws1", "task8", &execenv.GCMeta{
IssueID: issueID,
WorkspaceID: "ws1",
CompletedAt: time.Now(),
})
action := d.shouldCleanTaskDir(context.Background(), taskDir)
if action != gcActionOrphan {
t.Fatalf("expected gcActionOrphan for unreachable issue past TTL, got %d", action)
}
}
// TestShouldCleanTaskDir_Issue404RecentSkipped locks in the cross-workspace
// safety: the server returns 404 both for deleted issues and for workspaces
// the daemon token can't see, so a recent 404 must NOT trigger immediate
// cleanup — otherwise a token re-scope could wipe dirs whose issues are live.
func TestShouldCleanTaskDir_Issue404RecentSkipped(t *testing.T) {
t.Parallel()
issueID := "66666666-6666-6666-6666-666666666667"
mux := http.NewServeMux()
mux.HandleFunc(fmt.Sprintf("/api/daemon/issues/%s/gc-check", issueID), func(w http.ResponseWriter, r *http.Request) {
w.WriteHeader(http.StatusNotFound)
w.Write([]byte(`{"error":"not found"}`))
})
d := newGCTestDaemon(t, mux)
// Default production OrphanTTL; taskDir mtime is now, so it's fresh.
taskDir := createTaskDir(t, d.cfg.WorkspacesRoot, "ws1", "fresh-404", &execenv.GCMeta{
IssueID: issueID,
WorkspaceID: "ws1",
CompletedAt: time.Now(),
})
action := d.shouldCleanTaskDir(context.Background(), taskDir)
if action != gcActionSkip {
t.Fatalf("expected gcActionSkip for recent 404 (cross-workspace safety), got %d", action)
}
}
func TestCleanTaskDir_RemovesDirectory(t *testing.T) {
t.Parallel()
d := newGCTestDaemon(t, http.NewServeMux())
taskDir := createTaskDir(t, d.cfg.WorkspacesRoot, "ws1", "doomed", nil)
if _, err := os.Stat(taskDir); err != nil {
t.Fatal("task dir should exist before cleanup")
}
d.cleanTaskDir(taskDir)
if _, err := os.Stat(taskDir); !os.IsNotExist(err) {
t.Fatal("task dir should be removed after cleanup")
}
}
func TestGcWorkspace_CleansEmptyWorkspaceDir(t *testing.T) {
t.Parallel()
issueID := "77777777-7777-7777-7777-777777777777"
mux := http.NewServeMux()
mux.HandleFunc(fmt.Sprintf("/api/daemon/issues/%s/gc-check", issueID), func(w http.ResponseWriter, r *http.Request) {
w.Header().Set("Content-Type", "application/json")
json.NewEncoder(w).Encode(map[string]any{
"status": "done",
"updated_at": time.Now().Add(-10 * 24 * time.Hour),
})
})
d := newGCTestDaemon(t, mux)
wsDir := filepath.Join(d.cfg.WorkspacesRoot, "ws-empty")
createTaskDir(t, d.cfg.WorkspacesRoot, "ws-empty", "only-task", &execenv.GCMeta{
IssueID: issueID,
WorkspaceID: "ws-empty",
CompletedAt: time.Now(),
})
d.gcWorkspace(context.Background(), wsDir)
if _, err := os.Stat(wsDir); !os.IsNotExist(err) {
t.Fatal("empty workspace dir should be removed after all tasks cleaned")
}
}
func TestIsBareRepo(t *testing.T) {
t.Parallel()
t.Run("valid bare repo", func(t *testing.T) {
dir := t.TempDir()
os.WriteFile(filepath.Join(dir, "HEAD"), []byte("ref: refs/heads/main"), 0o644)
os.MkdirAll(filepath.Join(dir, "objects"), 0o755)
if !isBareRepo(dir) {
t.Fatal("expected isBareRepo=true for dir with HEAD + objects/")
}
})
t.Run("HEAD only", func(t *testing.T) {
dir := t.TempDir()
os.WriteFile(filepath.Join(dir, "HEAD"), []byte("ref: refs/heads/main"), 0o644)
if isBareRepo(dir) {
t.Fatal("expected isBareRepo=false for dir with only HEAD")
}
})
t.Run("empty dir", func(t *testing.T) {
dir := t.TempDir()
if isBareRepo(dir) {
t.Fatal("expected isBareRepo=false for empty dir")
}
})
}