Ask HN: How do you read and explore an undocumented codebase?

When working with a large third-party library which is open source or a large undocumented codebase left over by a previous developer, how do you go about understanding the codebase?