feat(rome_control_flow): introduce explicit exception control flow

[leops]

Summary

This PR replaces the (temporary) initial implementation of try-catch-finally statements in the control flow analysis (based on the concept of multiple entry points) with a more complete modelling solution for the exception control flow. The BasicBlock struct that make up the control flow graph now also contain a list of "exception handlers" and "cleanup handlers": these handlers point to a set of blocks to be executed in order when an exception is thrown (either with an explicit throw instruction or as a possible side-effect of a statement) or when the function returns respectively.

For JavaScript, the exceptions handlers of a block are all the finally clauses found going upward in the syntax tree until a catch clause is found (the block corresponding to this clause is the last handler in the list), while the cleanup handlers are made of all the finally clauses going up to the root level of the function (and ignoring all catch clauses)

The last change I've made is introducing a new finally_fallthrough flag on Jump instructions. This flag is used specifically to mark the implicit jump instruction inserted by the control flow generator at the end of a finally block towards the next statement: when the finally clause gets executed as an exception or cleanup handler instead of going through the normal control flow the target block of this jump gets modified to the next handler in the chain if it exists, or as a return instruction otherwise.

Test Plan

I've added additional test cases for JsTryStatement and JsTryFinallyStatement under the snapshots for the noDeadCode rule

[github-actions]

Playground for commit 0c6f162

[github-actions]

Parser conformance results on ubuntu-latest

js/262

Test result	main count	This PR count	Difference
Total	45878	45878	0
Passed	44938	44938	0
Failed	940	940	0
Panics	0	0	0
Coverage	97.95%	97.95%	0.00%

jsx/babel

Test result	main count	This PR count	Difference
Total	39	39	0
Passed	36	36	0
Failed	3	3	0
Panics	0	0	0
Coverage	92.31%	92.31%	0.00%

symbols/microsoft

Test result	main count	This PR count	Difference
Total	5946	5946	0
Passed	388	388	0
Failed	5558	5558	0
Panics	0	0	0
Coverage	6.53%	6.53%	0.00%

ts/babel

Test result	main count	This PR count	Difference
Total	588	588	0
Passed	519	519	0
Failed	69	69	0
Panics	0	0	0
Coverage	88.27%	88.27%	0.00%

ts/microsoft

Test result	main count	This PR count	Difference
Total	16257	16257	0
Passed	12393	12393	0
Failed	3864	3864	0
Panics	0	0	0
Coverage	76.23%	76.23%	0.00%

[IWANABETHATGUY]

I'm not sure if it's proper to put a language-specific property into "rome_control_flow", which seems to be a language-irrelevant crate

[leops]

finally clauses are a concept found in a few other languages than JavaScript (like Java, C#, Python, ...), languages that don't use it simply won't emit instructions with this flag (I specifically made it possible to create an instruction with this flag through the specialized append_finally_fallthrough method so it doesn't pollute all the other call sites of append_jump and languages that don't use it can transparently ignore it)

[IWANABETHATGUY]

Thanks for explaining