Support for some conditional call patterns + cloning improvement

[sifislag]

This PR includes a couple of improvements, vastly improving our function inference output:

• Add statement insertion round before the heuristic cloning round. Identifies labels pushed once but DUPed into multiple stack indexes by the end of the block's execution. This enables the HeuristicBlockCloner to clone them.
• Support for 0.8 JUMPI call patterns like the following:

  Block label_a:
    push continuation
    ...
    push label_b
    jumpi
  fallthrough:
    ...
    revert
  Block label_b:
    push func_start
    jump

  One downside is it takes 10s more to run the decompiler in interpreted mode. May cause some tests to fail from time to time.

---

Results

solc 0.8 contracts see a big improvement, as can be seen in the metadata-dataset1

1996 contracts decompiled/analyzed by some config
1996 contracts decompiled/analyzed by all configs (common)

ANALYTIC: decomp_time
jan24-metadata-master-200 (common): 6692.843376874924
jan24-metadata-cloneimprovement-200 (common): 6981.639788150787 (+4.315%)
jan24-metadata-bothimprovements-200 (common): 6955.371892213821 (+3.923%)

ANALYTIC: Analytics_JumpToMany
jan24-metadata-master-200 (common): 355 (+179.5%)
jan24-metadata-cloneimprovement-200 (common): 294 (+131.5%)
jan24-metadata-bothimprovements-200 (common): 127

ANALYTIC: Analytics_PublicFunction
jan24-metadata-master-200 (common): 78126
jan24-metadata-cloneimprovement-200 (common): 78126
jan24-metadata-bothimprovements-200 (common): 78126

ANALYTIC: Analytics_ReachableBlocks
jan24-metadata-master-200 (common): 1235649
jan24-metadata-cloneimprovement-200 (common): 1235649
jan24-metadata-bothimprovements-200 (common): 1235649

ANALYTIC: Analytics_BlockHasNoTACBlock
jan24-metadata-master-200 (common): 8
jan24-metadata-cloneimprovement-200 (common): 8
jan24-metadata-bothimprovements-200 (common): 11 (+37.5%)

ANALYTIC: Analytics_PolymorphicTargetSameCtx
jan24-metadata-master-200 (common): 37
jan24-metadata-cloneimprovement-200 (common): 37
jan24-metadata-bothimprovements-200 (common): 37

ANALYTIC: Analytics_LocalBlockEdge
jan24-metadata-master-200 (common): 1430818
jan24-metadata-cloneimprovement-200 (common): 1430743 (-0.005242%)
jan24-metadata-bothimprovements-200 (common): 1423748 (-0.4941%)

ANALYTIC: Analytics_StmtMissingOperand
jan24-metadata-master-200 (common): 87 (+128.9%)
jan24-metadata-cloneimprovement-200 (common): 55 (+44.74%)
jan24-metadata-bothimprovements-200 (common): 38

ANALYTIC: Analytics_PrivateFunctionMatchesMetadata
jan24-metadata-master-200 (common): 104741 (-0.163%)
jan24-metadata-cloneimprovement-200 (common): 104741 (-0.163%)
jan24-metadata-bothimprovements-200 (common): 104912

ANALYTIC: Analytics_PrivateFunctionMatchesMetadataIncorrectArgs
jan24-metadata-master-200 (common): 477 (+207.7%)
jan24-metadata-cloneimprovement-200 (common): 182 (+17.42%)
jan24-metadata-bothimprovements-200 (common): 155

ANALYTIC: Analytics_PrivateFunctionMatchesMetadataIncorrectReturnArgs
jan24-metadata-master-200 (common): 679 (+37.45%)
jan24-metadata-cloneimprovement-200 (common): 494
jan24-metadata-bothimprovements-200 (common): 494

for ir contracts the results are almost unchanged.

for contracts produced by the legacy pipeline in older compiler versions we still see an improvement, but way smaller than with 0.8 contracts

[github-actions[bot]]

Test Results (Souffle 2.3)

48 tests  +2   48 :white_check_mark: +2   28m 55s :stopwatch: + 4m 54s  1 suites ±0    0 :zzz: ±0   1 files   ±0    0 :x: ±0 

Results for commit f7efba92. ± Comparison against base commit 20df5a75.

:recycle: This comment has been updated with latest results.

[github-actions[bot]]

Test Results (Souffle 2.4)

48 tests  +2   48 :white_check_mark: +2   29m 16s :stopwatch: + 4m 22s  1 suites ±0    0 :zzz: ±0   1 files   ±0    0 :x: ±0 

Results for commit f7efba92. ± Comparison against base commit 20df5a75.

:recycle: This comment has been updated with latest results.