Point out the correct statement.

Every Lucene segment now stores a unique id per-segment and per-commit to aid in accurate replication of index files

The default norms format now uses sparse encoding when appropriate

Tokenizers and Analyzers no longer require Reader on init