Optimize pattern matching

[intsuc]

• Maranget, L. (2008). Compiling pattern matching to good decision trees. ML Workshop.
• Linear vs decision tree vs backtracking automaton?
• Due to the overhead of command execution, decision tree may not be always the most performant
  • We need to determine the threshold for switching between linear and decision tree based on benchmarks
• Due to the huge memory footprint of commands, we also need to take care of code size

Examples

if x [
  {a: 0, b: 0} -> 1,
  {a: 0, b: 1} -> 2,
  {a: 1}       -> 3,
]

#> linear
execute if data storage mcx: {scrutinee: {a: 0, b: 0}} run 1
execute if data storage mcx: {scrutinee: {a: 0, b: 1}} run 2
execute if data storage mcx: {scrutinee: {a: 1}} run 3

#> decision tree
execute if data storage mcx: {scrutinee: {a: 0}} run
  execute if data storage mcx: {scrutinee: {b: 0}} run 1
  execute if data storage mcx: {scrutinee: {b: 1}} run 2
execute if data storage mcx: {scrutinee: {a: 1}} run 3

#> backtracking automaton
# todo