Expand description
Format-neutral duplicate grouping over ArtifactIr.
Exact and normalized equality are equivalence relations, so their groups are keyed directly by content rather than by a transitive similarity graph. Near-match grouping is deliberately a later operation: it must use the source engine’s complete-linkage policy instead of union-find.
Structs§
- Dead
Code Report - Reachability result derived only from resolved local call edges.
- Duplicate
Group - One equality class of duplicate artifact symbols.
- Duplicate
Member - One occurrence in a
DuplicateGroup. - Duplicate
Report - Duplicate groups found in one artifact.
- Estimated
Refactor Savings Bytes - A model-derived estimate of a refactoring’s byte impact.
- Retained
Size - The bytes exclusively retained by one reachable symbol’s dominator region.
- Size
Classification - Size categories kept separate in artifact reports.
- Verified
Savings Bytes - A before/after reduction verified for one controlled refactoring.
Enums§
- Evidence
Confidence - Evidence strength reported without turning an observation into a promise.
Constants§
- DEFAULT_
MIN_ DUPLICATE_ DATA_ BYTES - The smallest data region that duplicate-data analysis reports by default.
Functions§
- classify_
sizes - Derive the size categories supported by the currently observed IR.
- classify_
sizes_ from_ duplicates - Derive size categories while reusing duplicate groups already calculated for another report surface.
- dead_
code_ candidates - Find symbols not reachable from parser-established exports.
- find_
duplicate_ data - Find exact duplicate data regions at or above
min_bytes. - find_
duplicates - Find exact and normalized duplicate groups in
artifact. - retained_
sizes - Calculate retained code sizes from a complete, unambiguous local call graph.