feat(server/m7-352): cron loop + main.go wiring for system playlist build

internal/playlists/system_cron.go

@@ -0,0 +1,61 @@
+package playlists
+
+import (
+	"context"
+	"log/slog"
+	"time"
+
+	"github.com/jackc/pgx/v5/pgxpool"
+
+	"git.fabledsword.com/bvandeusen/minstrel/internal/db/dbq"
+)
+
+// StartSystemPlaylistCron launches the in-process daily build loop in a
+// goroutine. Caller passes a context that's cancelled on graceful shutdown.
+//
+// Lifecycle:
+//   1. Clear any stale in_flight=true rows from a previously crashed process
+//      (safe at startup: by definition, nothing is running yet).
+//   2. Run once at startup so users catch up after a server restart without
+//      waiting up to 24h.
+//   3. Tick every 24h; runOnce on each tick.
+//   4. Exit on context cancellation.
+//
+// Per-user errors are logged but don't halt the loop. Per-cycle errors (e.g.,
+// "list active users" failed) are logged but the next tick still fires.
+func StartSystemPlaylistCron(ctx context.Context, pool *pgxpool.Pool, logger *slog.Logger) {
+	go func() {
+		q := dbq.New(pool)
+		if err := q.ClearStaleSystemPlaylistInFlight(ctx); err != nil {
+			logger.Warn("system playlist cron: startup recovery failed", "err", err)
+		}
+
+		runOnce(ctx, pool, logger)
+
+		ticker := time.NewTicker(24 * time.Hour)
+		defer ticker.Stop()
+		for {
+			select {
+			case <-ctx.Done():
+				return
+			case <-ticker.C:
+				runOnce(ctx, pool, logger)
+			}
+		}
+	}()
+}
+
+func runOnce(ctx context.Context, pool *pgxpool.Pool, logger *slog.Logger) {
+	q := dbq.New(pool)
+	users, err := q.ListActiveUsersForSystemPlaylists(ctx)
+	if err != nil {
+		logger.Error("system playlist cron: list active users failed", "err", err)
+		return
+	}
+	for _, userID := range users {
+		if err := BuildSystemPlaylists(ctx, pool, logger, userID, time.Now()); err != nil {
+			logger.Warn("system playlist build failed in cron",
+				"user_id", uuidStringPL(userID), "err", err)
+		}
+	}
+}