Extract direct recursive calls as calls self loops

[ayushcodes10]

What

_extract_generic's call resolution required tgt_nid != caller_nid before emitting a calls edge, so a function calling itself by name was silently dropped, no matter which language sharing this engine it was written in.

build_from_json already preserves a supplied recursive calls self loop rather than stripping it (test_recursive_call_self_loop_is_preserved), so the intent to keep genuine recursion visible already existed one layer up. Extraction just never produced the edge for that layer to preserve.

Repro (from the issue, reproduced before this fix)

def factorial(n):
    return 1 if n < 2 else n * factorial(n - 1)

def entry(n):
    return factorial(n)

Before: calls: [('repro_entry', 'repro_factorial')] (recursive call missing)
After: calls: [('repro_factorial', 'repro_factorial'), ('repro_entry', 'repro_factorial')]

Verified this also fixes recursion for Java, Ruby, and other _extract_generic-routed languages, since the guard removed was language agnostic.

Scope

Several dedicated per-language extractors (Go, Rust, CommonLisp, DreamMaker, Elixir, PowerShell, Zig) carry the same tgt_nid != caller_nid guard in their own separate files. Left those untouched here since the issue's repro and my verification are both Python/_extract_generic-scoped, and fixing those too would be a materially larger, separate change across files this report didn't exercise. Happy to open a follow-up if that's wanted.

Fixes #3350.

[ayushcodes10]

@safishamsi opened this against a report of direct Python recursion never producing a calls self-loop (#3350), traced to a shared tgt_nid != caller_nid guard in _extract_generic used by many languages, not just Python. Verified the same fix also restores recursion for Java and Ruby. Flagging the scope choice in the PR description: a few dedicated per-language extractors carry the same guard separately and weren't touched here. Happy to address any feedback.

[graphify-labs]

Graphify reviewed this change.

Looks safe to merge — no coupling regressions and no blocking issues, checked against the code graph (not a self-assessment).

Formal verification. No changes could be formally verified in this run.

---

Graphify review — findings

Preserves recursive call self-loops during extraction by dropping the tgt_nid != caller_nid guard in _extract_generic, so a function or method that calls itself now emits a calls edge to itself instead of having it silently discarded (#3350). Adds test_recursive_call_produces_self_loop and test_recursive_method_call_produces_self_loop covering plain and self.<method>() recursion, and documents in test_calls_no_self_loops that its fixture just has no recursive functions.

No blocking issues surfaced. 2 lower-confidence candidates did not survive cross-model review.

Analysis details — impact, health, verification

Impact & health

Graphify review

Impact — 1009 functions depend on the 608 functions this change touches.

Health — this change adds coupling hotspots:

• new: _extract_generic() — 18 callers, 26 callees
• new: extract_js() — 85 callers, 4 callees
• new: extract_xaml() — 19 callers, 17 callees
• new: extract_objc() — 27 callers, 9 callees
• new: extract_julia() — 17 callers, 7 callees
• new: extract_cpp() — 27 callers, 3 callees
• new: extract_vue() — 10 callers, 7 callees
• new: walk() — 1 callers, 58 callees
• …and 8 more — each is listed as a finding

Verification — 1009 functions in the blast radius were not formally verified this run (proofs are advisory here).

Gate & verification

graphify gate

PASS — objectively clean (no health regressions, tests not run — proofs not run this pass (advisory)). Grounded, not self-assessed.

Advisory (not blocking):

• verification_scope: 949 function(s) in the blast radius were not formally verified this run

Formal verification

Could not verify: Could not verify \_extract\_generic.

The verifier did not have enough to check \_extract\_generic, so it is saying so rather than guessing. No false assurance is the whole point.

Guarantee: No guarantee either way, this is an honest abstention, not a pass.

Note: Reason: parameter `path` is annotated `Path` — outside the synthesizable primitive/collection set

· 16 more finding(s) on lines outside this diff (see the check run).