Repository navigation
v2 Stage 3: L3 intraprocedural CFG/CDG/DDG (syntactic) #73
Description
Part of #69.
Stage 3 — L3 intraprocedural CFG/CDG/DDG (syntactic)
Branch: feat/v2-stage3-l3
Deliverables:
dataflow/identity.py— the(signature, int node_id) ↔ (can:// callable-id, ordinal-id)bijection; consumed by both emitters.- Populate
bodywith statement nodes +@entry/@exit; emitcfg/cdg/ddgedge lists on each callable, re-targeting the existingcfg/dominance/defusekernels through the identity map. - Split the current PDG rows into
cfg/cdg/ddg; add a syntactic-only DDG mode (textual interference, oracle bypassed,prov:["ssa"]). - CLI:
--graphs cfg,dfg,pdgscoping;--graph-field-depthat L3+;sdgrequires-a 4(flag error otherwise). - Neo4j:
PyCFGNodekeyed by ordinal id;PY_DDGgainsprov. - Migrate
test/test_dataflow_{cfg,dominance,defuse,pdg,scc,cpg,slicing}.pyto v2 ids.
Files: dataflow/{identity,builder,defuse}.py, core.py, __main__.py, neo4j/project.py, dataflow tests.
Gate: intraprocedural backward slice == hand-computed set exactly (over ordinal ids); assert_conformant(..., 3) incl. no dangling cfg/cdg/ddg; L2 ⊆ L3.
Expand into full bite-sized tasks with
superpowers:writing-plansat stage start.
Activity
rahlk commentedon Jul 7, 2026
Two forward-risks surfaced by the Stage 1 whole-branch review that Stage 3 must handle:
Analysis(bodies + edge lists) and_file_unchangedkeys only on content hash / mtime / size — not onanalysis_level, andpopulate_l1_body/assign_idsadd tobodybut never clear it. Harmless at L1 (onlycallnodes populatebody), but once Stage 3 fillsbodywith statements +@entry/@exit+cfg/cdg/ddg, a subsequent lower-level run reusing the cache would leak higher-level content into lower-level output — breaking level purity / theL1⊆…⊆L4monotonicity gate. Fix: key the cache byanalysis_level(or clear regeneratedbody/edge collections on reuse)._project_program_graphsis currently disabled + dead-references the dissolvedprogram_graphsfield) on the v2 tree, and close the latent dual:PySymbolconstraint (id + signature) before Neo4j is exercised.rahlk commentedon Jul 7, 2026
Additional Stage-3 tracking (from the Stage 2 whole-branch review): when re-enabling the Neo4j projection, reconcile a classification-precedence asymmetry —
reidentify_call_graph(JSON) resolves declared-first (sig_to_id wins) while_call_endpoint(Neo4j) resolves external-first (external_symbols wins). Inert today (a symbol is either project-declared or external, never both), but make them consistent so JSON and Neo4j can't disagree on an endpoint id if that invariant ever weakens. Track alongside the latent dual:PySymbol(id + signature) uniqueness constraint.rahlk commentedon Jul 15, 2026
Landed on main via the stacked v2 chain (PR #88 → merge d477207) and released as v1.0.0 (schema_version 2.0.0).