review column: 'approve above' on hover, wrap long audio titles

each top-level card now shows a secondary button on hover ('↑ approve
above') that approves every card listed above this one in one
round-trip. uses a new POST /api/review/approve-batch { itemIds } that
ignores non-pending items so stale client state can't 409. series cards
get the same affordance scoped via a named tailwind group so it
doesn't collide with the inner episode cards' own hover state.

fix the horizontal-scroll glitch: long unbreakable audio titles (e.g.
the raw release filename) now line-wrap inside the card via
[overflow-wrap:anywhere] + min-w-0 on the span. previously
break-words was a no-op since there were no whitespace break points
in the release string.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

server/api/review.ts

@@ -575,6 +575,44 @@ app.post("/approve-all", (c) => {
 	return c.json({ ok: true, count: pending.length });
 });
 
+// ─── Batch approve (by item id list) ─────────────────────────────────────────
+// Used by the "approve up to here" affordance in the review column. The
+// client knows the visible order (movies + series sort-key) and passes in
+// the prefix of item ids it wants approved in one round-trip. Items that
+// aren't pending (already approved / skipped / done) are silently ignored
+// so the endpoint is idempotent against stale client state.
+app.post("/approve-batch", async (c) => {
+	const db = getDb();
+	const body = await c.req.json<{ itemIds?: unknown }>().catch(() => ({ itemIds: undefined }));
+	if (
+		!Array.isArray(body.itemIds) ||
+		!body.itemIds.every((v) => typeof v === "number" && Number.isInteger(v) && v > 0)
+	) {
+		return c.json({ ok: false, error: "itemIds must be an array of positive integers" }, 400);
+	}
+	const ids = body.itemIds as number[];
+	if (ids.length === 0) return c.json({ ok: true, count: 0 });
+
+	const placeholders = ids.map(() => "?").join(",");
+	const pending = db
+		.prepare(
+			`SELECT rp.*, mi.id as item_id FROM review_plans rp JOIN media_items mi ON mi.id = rp.item_id
+			 WHERE rp.status = 'pending' AND rp.is_noop = 0 AND mi.id IN (${placeholders})`,
+		)
+		.all(...ids) as (ReviewPlan & { item_id: number })[];
+
+	let count = 0;
+	for (const plan of pending) {
+		db.prepare("UPDATE review_plans SET status = 'approved', reviewed_at = datetime('now') WHERE id = ?").run(plan.id);
+		const { item, streams, decisions } = loadItemDetail(db, plan.item_id);
+		if (item) {
+			enqueueAudioJob(db, plan.item_id, buildCommand(item, streams, decisions));
+			count++;
+		}
+	}
+	return c.json({ ok: true, count });
+});
+
 // ─── Auto-approve high-confidence ────────────────────────────────────────────
 // Approves every pending plan whose original language came from an authoritative
 // source (radarr/sonarr). Anything with low confidence keeps needing a human.