refactor(native): extract emit_pts_alias_edges params into struct

[carlos-alm]

Groups the 14 parameters of emit_pts_alias_edges in build_edges.rs into a PtsAliasCtx struct, removing the #[allow(clippy::too_many_arguments)] suppression. No behavior change.

Closes #1475

[claude]

Claude finished @carlos-alm's task —— View job

---

PR Review

• Gather context
• Read changed file
• Analyze root cause and code quality
• Post final review

---

Root Cause Analysis

Problem Being Solved: emit_pts_alias_edges previously took 14 parameters, triggering Clippy's too_many_arguments lint, which was suppressed with #[allow(clippy::too_many_arguments)]. This PR groups the per-call-site lookup parameters into a PtsAliasCtx<'a> struct, removing the suppression legitimately.

Why This Approach: Correct approach — the lint was pointing at a real design smell. A context struct is idiomatic Rust for this pattern. The parameters grouped (pts, lookup_name, call_line, caller_id, caller_name, is_dynamic, rel_path, imported_names, type_map) all genuinely belong to "what are we resolving right now," which is a coherent concept.

Risk Assessment: Low. Pure structural refactor; no logic paths changed. The struct is pub(crate) scoped (stack-allocated, no heap cost), lifetimes are carried correctly on the struct, and both call sites (process_file lines 600–616 and 629–645) are updated symmetrically.

---

Code Quality Observations

Positive:

• PtsAliasCtx<'a> is defined at lines 365–375 with a clear doc comment explaining its purpose
• Lifetimes are applied correctly: the struct borrows from the same 'a lifetime as EdgeContext<'a> where appropriate, and uses '_ for the function parameter
• Both call sites (emit_pts_alias_edges at lines 600 and 629) are updated consistently
• The #[allow(clippy::too_many_arguments)] suppression is cleanly removed

Minor observations (no blocking issues):

1. is_dynamic: u32 in the struct (line 371): The field carries u32 (a DB encoding) rather than bool, which is a pre-existing leaky abstraction — not introduced by this PR, but the struct wrapping now makes it more visible. Not a blocker.

2. Struct field ordering (lines 365–375): The struct interleaves lookup scalars (call_line, caller_id, caller_name, is_dynamic) with reference fields (pts, rel_path, imported_names, type_map). Grouping by conceptual role (identity vs. maps) would improve readability, but this is subjective and not a blocker.

3. Doc comment on emit_pts_alias_edges (line 377–379): The doc still says "Shared by the no-receiver gate and the receiver-key (rest.prop()) fallback" — this is accurate and unchanged. Good.

---

Backlog Compliance

• Zero-dep: ✓ — no new dependencies
• Foundation-aligned: ✓ — pure Rust internal refactor, no API/CLI/DB changes
• Problem-fit: 1 — this is internal code hygiene; it does not affect AI agent accuracy, token use, or blast-radius detection
• Breaking: No — PtsAliasCtx is not pub, emit_pts_alias_edges is not pub, and build_call_edges signature is unchanged
• Tier: 1 (zero-dep ✓ + foundation-aligned ✓)

Problem-fit = 1 note: Per the review rubric, this warrants flagging as low-priority / tangential. However, removing a #[allow(...)] suppression is a valid maintenance task and the PR is correctly scoped to a single concern. The low problem-fit score is not a rejection reason here — it's a classification, not a critique.

---

Final Recommendation

• Rating: ⭐⭐⭐⭐ (4/5)
• Action: APPROVE
• Reasoning: The refactor is correct, well-scoped, and legitimately removes a Clippy suppression rather than hiding a problem. Lifetimes are handled properly. No behavior change. The only reason it doesn't reach 5 stars is that the struct field ordering is slightly inconsistent and is_dynamic: u32 leaks a storage concern into the context abstraction — both pre-existing issues not introduced here.

---

[greptile-apps]

Greptile Summary

This PR refactors emit_pts_alias_edges in build_edges.rs by bundling 9 of its 14 parameters into a new PtsAliasCtx<'a> struct, allowing the #[allow(clippy::too_many_arguments)] suppression to be removed. The change is purely structural — no logic, algorithm, or output is modified.

• PtsAliasCtx<'a> is introduced as a private, module-scoped struct holding the per-call-site lookup inputs (pts, lookup_name, call_line, caller_id, caller_name, is_dynamic, rel_path, imported_names, type_map).
• Both call sites in process_file are updated to construct PtsAliasCtx inline and pass it by reference; the remaining mutable output parameters (seen_edges, pts_edge_map, edges) are kept as direct arguments.

Confidence Score: 5/5

Safe to merge — the change is a pure parameter-grouping refactor with no logic modifications.

Every access of the original positional parameters is mechanically replaced by the corresponding alias_ctx field, both call sites are updated, and the '_ lifetime annotation on alias_ctx correctly lets the borrow checker infer a lifetime independent of 'a. No algorithmic, output, or control-flow changes were made.

No files require special attention.

Important Files Changed

Filename	Overview
crates/codegraph-core/src/domain/graph/builder/stages/build_edges.rs	Adds PtsAliasCtx struct and rewrites emit_pts_alias_edges to accept it; both call sites updated. Pure structural refactor — no behavior change.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A["process_file()"] -->|"builds PtsAliasCtx { pts, lookup_name, call_line,\ncaller_id, caller_name, is_dynamic,\nrel_path, imported_names, type_map }"| B["emit_pts_alias_edges(ctx, &alias_ctx, seen_edges, pts_edge_map, edges)"]
    B --> C["resolve_via_points_to(alias_ctx.lookup_name, alias_ctx.pts)"]
    C --> D["for each alias: resolve_call_targets()"]
    D --> E["sort_targets_by_confidence()"]
    E --> F["compute edge_key\n(alias_ctx.caller_id << 32 | t.id)"]
    F --> G{{"t.id != caller_id\n&& not in seen_edges\n&& not in pts_edge_map"}}
    G -->|yes| H["compute_confidence() - HOP_PENALTY\nif > 0.0 → push ComputedEdge"]
    G -->|no| I["skip"]

Loading

_{Reviews (4): Last reviewed commit: "fix: resolve merge conflicts with main" | Re-trigger Greptile}

[github-actions]

Codegraph Impact Analysis

2 functions changed → 16 callers affected across 1 files

• emit_pts_alias_edges in crates/codegraph-core/src/domain/graph/builder/stages/build_edges.rs:380 (16 transitive callers)
• process_file in crates/codegraph-core/src/domain/graph/builder/stages/build_edges.rs:440 (15 transitive callers)

[carlos-alm]

@greptileai

[carlos-alm]

Acknowledged Claude's review: the struct field ordering (is_dynamic: u32 leaking storage concern) and the interleaved field order are pre-existing decisions. Both are accepted as-is — the struct is internal (pub(crate)) and refactoring the u32 encoding to bool would be a separate concern (no issue filed since it's in an existing stable pattern). No blocking items.

[carlos-alm]

@greptileai