Cadence Parser Optimizations

[j1010001]

This Epic tracks tasks/stories related to evaluation of the Cadence parser performance, such as using panics vs replacing with error values, using generics, and storing the AST so that it can be reused without re-parsing.

[turbolent]

PRs:

• Optimize location range #2036
• Make the lexer operate on []byte instead of string for source-code #1850
• Remove values from tokens, make all code byte slices #1930
• Improve the AST visitor further  #1919
• Optimize and make visitor generic #1907
• Pool lexers to reduce allocations and improve performance #1610