fix(MULTIVERSION-0001): fix multi-version indexing — jobs never created or triggered for secondary versions

Two bugs prevented secondary versions from ever being indexed:

1. JobQueue.enqueue() and RepositoryService.createIndexingJob() deduplication
   only checked repository_id, so a queued default-branch job blocked all
   version-specific jobs for the same repo. Fix: include version_id in the
   WHERE clause so only exact (repository_id, version_id) pairs are deduped.

2. POST /api/v1/libs/:id/versions used repoService.createIndexingJob() which
   inserts a job record but never triggers queue processing. Fix: use
   queue.enqueue() (same fallback pattern as the libs endpoint) so setImmediate
   fires processNext() after the job is inserted.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

src/lib/server/pipeline/job-queue.ts

@@ -36,14 +36,17 @@ export class JobQueue {
 	 * existing job instead of creating a duplicate.
 	 */
 	enqueue(repositoryId: string, versionId?: string): IndexingJob {
-		// Return early if there's already an active job for this repo.
+		// Return early if there's already an active job for this exact (repo, version) pair.
+		const resolvedVersionId = versionId ?? null;
 		const activeRaw = this.db
-			.prepare<[string], IndexingJobEntity>(
+			.prepare<[string, string | null, string | null], IndexingJobEntity>(
 				`${JOB_SELECT}
-         WHERE repository_id = ? AND status IN ('queued', 'running')
+         WHERE repository_id = ?
+           AND (version_id = ? OR (version_id IS NULL AND ? IS NULL))
+           AND status IN ('queued', 'running')
          ORDER BY created_at DESC LIMIT 1`
 			)
-			.get(repositoryId);
+			.get(repositoryId, resolvedVersionId, resolvedVersionId);
 
 		if (activeRaw) {
 			// Ensure the queue is draining even if enqueue was called concurrently.