Scope
Umbrella tracker for graph query & trace_path resolution — cases where the graph contains the right nodes/edges but the query-side traversal returns empty or incomplete caller/callee sets. The edges exist; the answer is wrong.
Sub-issues
Acceptance
Per item: a reproduce-first test on a public fixture showing the expected caller/callee set, then the traversal returns it. #480 in particular needs a root-cause test (edges present, traversal yields nothing) before any fix.
Why one task
These all live in the traversal/query layer (BFS over CALLS, symbol identity / dedup, data-flow arg resolution) rather than extraction. Grouping keeps the query-correctness work coherent.
Scope
Umbrella tracker for graph query &
trace_pathresolution — cases where the graph contains the right nodes/edges but the query-side traversal returns empty or incomplete caller/callee sets. The edges exist; the answer is wrong.Sub-issues
trace_pathreturns empty for all functions despite traversable CALLS edges (v0.8.1, macOS arm64)trace_pathdata_flow mode doesn't surface arg expressions; NestJS DI patterns defeat ~70% of caller resolution.d.tsdeclaration (callers silently split by import style) #546 —trace_path/ reverse-dependency returns an INCOMPLETE caller set when a symbol is duplicated by an ambient.d.tsdeclaration (callers silently split by import style)Acceptance
Per item: a reproduce-first test on a public fixture showing the expected caller/callee set, then the traversal returns it. #480 in particular needs a root-cause test (edges present, traversal yields nothing) before any fix.
Why one task
These all live in the traversal/query layer (BFS over CALLS, symbol identity / dedup, data-flow arg resolution) rather than extraction. Grouping keeps the query-correctness work coherent.