Pin the close before the reopen in the archive sweep (closes #103)

Add a test that keeps the archive's connection from before a sweep
and checks that the sweep closed it. Without the close before the
reopen, the reopen replaced the handle without closing it and one
connection leaked per archive per sweep, yet every test passed.

The sweeper's listing query now takes the sweep's context; a sweep
cancelled by the app stopping returns from that query without an
error line. A comment on the sweeper's cancel function says why it
needs no lock.

The handlers' type-filtered count of a webhook's remaining targets
no longer exists: since each database target has its own archive
file, deleting a target evicts that target's writer alone.

Model: opus-5-5

internal/delivery/archive_sweeper.go

@@ -45,8 +45,13 @@ type ArchiveSweeper struct {
 	eng      *Engine
 	log      *slog.Logger
 	interval time.Duration
-	cancel   context.CancelFunc
-	wg       sync.WaitGroup
+
+	// cancel needs no lock: fx calls the stop hook only after the
+	// start hook has returned, so stop never reads it while start
+	// is still setting it.
+	cancel context.CancelFunc
+
+	wg sync.WaitGroup
 }
 
 // NewArchiveSweeper creates the archive sweeper and registers
@@ -163,10 +168,18 @@ func (s *ArchiveSweeper) sweep(ctx context.Context) {
 	var targets []database.Target
 
 	err := s.db.DB().
+		WithContext(ctx).
 		Model(&database.Target{}).
 		Where("type = ?", database.TargetTypeDatabase).
 		Find(&targets).Error
 	if err != nil {
+		// The app stopping as a sweep starts cancels the listing.
+		// Stopping is not a failure, so it must not produce an
+		// error line.
+		if ctx.Err() != nil {
+			return
+		}
+
 		s.log.Error(
 			"archive sweep: failed to list database targets",
 			"error", err,