EXPRESSO: A Benchmark and Analysis of Discrete Expressive Speech Resynthesis