feat(impact): codemap impact — symbol/file blast-radius walker (impl of PR #49 plan)#50
Merged
SutuSebastian merged 7 commits intomainfrom May 3, 2026
Merged
feat(impact): codemap impact — symbol/file blast-radius walker (impl of PR #49 plan)#50SutuSebastian merged 7 commits intomainfrom
SutuSebastian merged 7 commits intomainfrom
Conversation
application/impact-engine.ts — single WITH RECURSIVE per (direction, backend), JS-side merge + dedup + summary. Backends: - calls (symbol-level, up=callers / down=callees) - dependencies (file-level) - imports (file-level, resolved_path only) Plan §D2 backend split, §D6 cycle detection (path-string + instr), §D8 edge tags. 27 unit tests cover: walks (up/down/both), depth caps, limit caps, self-loop + 3-cycle, file-vs-symbol resolution, --via skipped_backends.
…cer 2) - src/cli/cmd-impact.ts — parser + runner mirroring cmd-show.ts shape - main.ts dispatch, bootstrap.ts validateIndexModeArgs + printCliUsage - 14 parser tests cover happy paths + 9 error cases (--depth/limit/direction/via validation) Smoke: `bun src/index.ts impact src/db.ts --direction up --depth 1` lists the real fan-in on this repo (every file in src/ that depends on db.ts). `bun src/index.ts impact handleQuery --json --summary` returns the summary shape ready for jq pipelines.
- impactArgsSchema + handleImpact in tool-handlers.ts (Zod + ToolResult shape) - registerImpactTool in mcp-server.ts (one-line wrapper, like show/snippet) - 5 MCP integration tests (tools/list, symbol target, file target, summary trims matches but keeps summary.nodes count, Zod rejects float depth) HTTP POST /tool/impact lights up automatically next tracer via the existing http-server.ts dispatcher (one switch arm).
- TOOL_NAMES gains 'impact' (auto-listed in /tools) - dispatchTool switch arm with Zod validation (400 on bad body) - 4 integration tests: happy path, missing target, non-int depth, bad direction enum Both transports (MCP + HTTP) now expose impact via the same handleImpact pure function.
…l + changeset (Tracer 5) - README.md: new 'Impact analysis' code block + impact in MCP tools list - docs/architecture.md: 'Impact wiring' section + impact-engine in application/ inventory - docs/glossary.md: 'codemap impact / impact tool' entry under I (full envelope spec) - docs/roadmap.md: drop the impact backlog item (it shipped) - docs/research/fallow.md: Adjacent-shipped bullet referencing PR #49 (plan) + #50 (impl) - .agents/rules/codemap.md + templates/agents/rules/codemap.md: new table row + 'Impact' paragraph + impact in MCP tools list (Rule 10 lockstep) - .agents/skills/codemap/SKILL.md + templates/agents/skills/codemap/SKILL.md: impact tool description (Rule 10 lockstep) - .changeset/codemap-impact.md: minor — full design rationale + decisions - docs/plans/impact.md: deleted per docs-governance Rule 3 (canonical descriptions now live in architecture / glossary / README)
🦋 Changeset detectedLatest commit: 8762100 The changes in this PR will be included in the next version bump. This PR includes changesets to release 1 package
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
|
Warning Rate limit exceeded
To keep reviews running without waiting, you can enable usage-based add-on for your organization. This allows additional reviews beyond the hourly cap. Account admins can enable it under billing. ⌛ How to resolve this issue?After the wait time has elapsed, a review can be triggered using the We recommend that you space out your commits to avoid hitting the rate limit. 🚦 How do rate limits work?CodeRabbit enforces hourly rate limits for each developer per organization. Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout. Please see our FAQ for further information. ℹ️ Review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (2)
📝 WalkthroughWalkthroughThe pull request implements and ships the ChangesImpact Feature Launch
Sequence DiagramsequenceDiagram
participant Client as Client (CLI/MCP/HTTP)
participant Handler as Tool Handler
participant Engine as findImpact Engine
participant DB as CodemapDatabase
Client->>Handler: impact(target, direction, via, depth, limit)
Handler->>Engine: findImpact(db, opts)
Engine->>DB: resolveTarget(target)
DB-->>Engine: ImpactTarget (kind, name, file_path[])
Engine->>Engine: resolveBackends(via, target.kind)
Engine-->>Engine: backend[], skipped[]
loop for each (direction, backend)
Engine->>DB: walk() recursive CTE<br/>(calls|dependencies|imports)
DB-->>Engine: ImpactNode[] (with depth, edge_type)
end
Engine->>Engine: deduplicate by<br/>(direction, kind, name, file_path)
Engine->>Engine: compute terminated_by<br/>(limit vs depth vs exhausted)
Engine->>Engine: summarize by_kind counts
Engine-->>Handler: ImpactResult
alt summary mode
Handler->>Handler: clear matches, keep summary
end
Handler-->>Client: JSON envelope<br/>(target, matches, summary)
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Possibly related PRs
Suggested labels
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Review rate limit: 0/1 reviews remaining, refill in 42 minutes and 17 seconds.Comment |
Re-read every comment authored in this PR; cut/slim ones a teammate could re-derive in <30s. Targets: - impact-engine.ts: drop trivial type doc, slim findImpact JSDoc, slim termination heuristic from 6 lines → 2, drop redundant inline notes - cmd-impact.ts + tool-handlers.ts: shorten the --summary trim explanation No behaviour change; checks green (89 tests + lint + typecheck).
Contributor
Author
|
@coderabbitai full review |
✅ Actions performedFull review triggered. |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
🧹 Nitpick comments (5)
src/application/impact-engine.ts (1)
281-331: 💤 Low valueConsider documenting the empty
file_pathfallback behavior.The
lookupSymbolFilefallback (line 320) is a good addition for cross-module callees. However, when the lookup also fails (line 327), the fallback to an empty string""could be unexpected for consumers. While this is better thannull/undefinedfor type safety, consider adding a brief comment explaining this edge case.💡 Suggested comment
const symFile = r.file_path ?? lookupSymbolFile(db, r.node); const node: ImpactNode = { depth: r.depth, direction: opts.direction, edge, kind: "symbol", name: r.node, + // Empty string when symbol not found in index (external/dynamic calls) file_path: symFile ?? "", };🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/application/impact-engine.ts` around lines 281 - 331, Add a brief inline comment in the walkCalls function near the fallback that sets file_path to an empty string (where symFile ?? "" is used) explaining that lookupSymbolFile(db, r.node) may return null for cross-module symbols and we intentionally use an empty string for type-safety and predictable consumer behavior (navigation absent), so callers should treat "" as "no known file path". Reference symbols: walkCalls, lookupSymbolFile, ImpactNode.file_path.src/cli/cmd-impact.ts (1)
268-275: 💤 Low valueRedundant
skipped_backends !== undefinedguard — field is required inImpactResult.
ImpactResult.skipped_backendsis typed as a required (non-optional) array, so the!== undefinedcheck is alwaystrue. The code is functionally correct (an empty array produces no output), but the guard implies optionality that doesn't exist and may confuse future readers.♻️ Suggested fix
- if (result.skipped_backends !== undefined) { - for (const s of result.skipped_backends) { + for (const s of result.skipped_backends) { console.error(`# skipped backend "${s.backend}": ${s.reason}`); - } }🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/cli/cmd-impact.ts` around lines 268 - 275, The check `result.skipped_backends !== undefined` is redundant because ImpactResult.skipped_backends is a required array; remove the guard and iterate directly over `result.skipped_backends` (e.g., `for (const s of result.skipped_backends) { ... }`). Update the block around the `result` usage so it assumes `skipped_backends` is always an array and let an empty array naturally produce no output; reference `ImpactResult.skipped_backends` and the `result` variable when making the change.src/application/tool-handlers.ts (1)
663-679: 💤 Low value
impactArgsSchemaandImpactArgslack documentation per coding guidelines.As per coding guidelines, "All public APIs must have accompanying documentation." Both
impactArgsSchemaandImpactArgsare newly exported public symbols without JSDoc.📝 Suggested doc additions
+/** Zod schema for the `impact` MCP/HTTP tool — mirrors `FindImpactOpts` plus the + * transport-level `summary` trim flag. */ export const impactArgsSchema = { target: z.string().min(1, "target must be a non-empty string"), ... }; +/** Validated args passed to {`@link` handleImpact}. */ export interface ImpactArgs {🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/application/tool-handlers.ts` around lines 663 - 679, Add JSDoc for the exported impactArgsSchema and ImpactArgs to satisfy public API documentation rules: for impactArgsSchema, document that it validates impact query options (describe each key: target - non-empty string, direction - "up" | "down" | "both" optional, via - "dependencies"|"calls"|"imports"|"all" optional, depth - non-negative int optional, limit - positive int optional, summary - boolean optional) and mention it uses Zod; for the ImpactArgs interface, add a JSDoc comment describing the shape and purpose of each property and reference the related types ImpactDirection and ImpactBackend where applicable, including any default behaviors if used elsewhere.src/application/impact-engine.test.ts (2)
203-210: ⚡ Quick winStrengthen the exact-fit limit test with a positive
terminated_byassertion.The test asserts only that
terminated_byis not"limit". If a future regression caused the engine to return"depth"instead of"exhausted"for a fully-explored graph, this test would silently pass.✅ Proposed improvement
expect(r.matches).toHaveLength(3); - expect(r.summary.terminated_by).not.toBe("limit"); + expect(r.summary.terminated_by).toBe("exhausted");🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/application/impact-engine.test.ts` around lines 203 - 210, The test for findImpact currently only asserts that r.summary.terminated_by is not "limit", which is weak; instead update the assertion to check the exact expected termination reason (e.g., expect(r.summary.terminated_by).toBe("exhausted")) to ensure the engine reports the correct completion state for a fully-explored graph; locate the test using the findImpact call and its r.summary.terminated_by assertion and replace the negative check with a positive equality check to the intended terminal value.
280-288: Prefer order-insensitive assertion for theviaarray since backend order is an implementation detail.
expect(r.via).toEqual(["dependencies", "imports"])is order-sensitive but the engine only guarantees which backends apply to file targets, not their order. The hardcodedfileBackendsarray at line 221 ofimpact-engine.tsdetermines order, which is an implementation choice rather than a contractual requirement. Consider using:- expect(r.via).toEqual(["dependencies", "imports"]); + expect(r.via.slice().sort()).toEqual(["dependencies", "imports"]);This test will remain stable if the backend enumeration order is ever refactored without changing behavior.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/application/impact-engine.test.ts` around lines 280 - 288, The assertion on r.via is order-sensitive but backend order is an implementation detail; update the test case "via=all on file target uses dependencies + imports (skips calls)" to assert membership rather than exact order by using an order-insensitive check (e.g., expect(r.via).toEqual(expect.arrayContaining(["dependencies","imports"])) plus an assertion on length or use a Set comparison) and keep the skipped_backends assertion as-is; locate the test using findImpact and the fileBackends/impact behavior references to make this change.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Nitpick comments:
In `@src/application/impact-engine.test.ts`:
- Around line 203-210: The test for findImpact currently only asserts that
r.summary.terminated_by is not "limit", which is weak; instead update the
assertion to check the exact expected termination reason (e.g.,
expect(r.summary.terminated_by).toBe("exhausted")) to ensure the engine reports
the correct completion state for a fully-explored graph; locate the test using
the findImpact call and its r.summary.terminated_by assertion and replace the
negative check with a positive equality check to the intended terminal value.
- Around line 280-288: The assertion on r.via is order-sensitive but backend
order is an implementation detail; update the test case "via=all on file target
uses dependencies + imports (skips calls)" to assert membership rather than
exact order by using an order-insensitive check (e.g.,
expect(r.via).toEqual(expect.arrayContaining(["dependencies","imports"])) plus
an assertion on length or use a Set comparison) and keep the skipped_backends
assertion as-is; locate the test using findImpact and the fileBackends/impact
behavior references to make this change.
In `@src/application/impact-engine.ts`:
- Around line 281-331: Add a brief inline comment in the walkCalls function near
the fallback that sets file_path to an empty string (where symFile ?? "" is
used) explaining that lookupSymbolFile(db, r.node) may return null for
cross-module symbols and we intentionally use an empty string for type-safety
and predictable consumer behavior (navigation absent), so callers should treat
"" as "no known file path". Reference symbols: walkCalls, lookupSymbolFile,
ImpactNode.file_path.
In `@src/application/tool-handlers.ts`:
- Around line 663-679: Add JSDoc for the exported impactArgsSchema and
ImpactArgs to satisfy public API documentation rules: for impactArgsSchema,
document that it validates impact query options (describe each key: target -
non-empty string, direction - "up" | "down" | "both" optional, via -
"dependencies"|"calls"|"imports"|"all" optional, depth - non-negative int
optional, limit - positive int optional, summary - boolean optional) and mention
it uses Zod; for the ImpactArgs interface, add a JSDoc comment describing the
shape and purpose of each property and reference the related types
ImpactDirection and ImpactBackend where applicable, including any default
behaviors if used elsewhere.
In `@src/cli/cmd-impact.ts`:
- Around line 268-275: The check `result.skipped_backends !== undefined` is
redundant because ImpactResult.skipped_backends is a required array; remove the
guard and iterate directly over `result.skipped_backends` (e.g., `for (const s
of result.skipped_backends) { ... }`). Update the block around the `result`
usage so it assumes `skipped_backends` is always an array and let an empty array
naturally produce no output; reference `ImpactResult.skipped_backends` and the
`result` variable when making the change.
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: bc53c33a-49b8-4d0d-a649-197be78e3748
📒 Files selected for processing (22)
.agents/rules/codemap.md.agents/skills/codemap/SKILL.md.changeset/codemap-impact.mdREADME.mddocs/architecture.mddocs/glossary.mddocs/plans/impact.mddocs/research/fallow.mddocs/roadmap.mdsrc/application/http-server.test.tssrc/application/http-server.tssrc/application/impact-engine.test.tssrc/application/impact-engine.tssrc/application/mcp-server.test.tssrc/application/mcp-server.tssrc/application/tool-handlers.tssrc/cli/bootstrap.tssrc/cli/cmd-impact.test.tssrc/cli/cmd-impact.tssrc/cli/main.tstemplates/agents/rules/codemap.mdtemplates/agents/skills/codemap/SKILL.md
💤 Files with no reviewable changes (2)
- docs/roadmap.md
- docs/plans/impact.md
- impact-engine.ts:333 — one-line comment explaining empty-string file_path
fallback (external/dynamic call sites)
- impact-engine.test.ts:209 — strengthen exact-fit limit test from
not.toBe('limit') to toBe('exhausted')
Pushed back on nitpicks 2+3 (hallucinated — see PR comment) and 5 (style;
deterministic order is fine to lock).
Contributor
Author
|
Triaged the 5 nitpicks per Applied (2):
Pushed back (2 — hallucinated):
Declined (1 — style):
|
2 tasks
SutuSebastian
added a commit
that referenced
this pull request
May 3, 2026
…(v1.x backlog) (#51) * docs(plan): codemap audit --base <ref> — worktree + reindex strategy (v1.x backlog) Plan for the next-best agent-value loop: PR-review structural-diff. Replaces today's 3-step --baseline dance with one verb. Reuses 90% of the existing audit infrastructure (PR #33); only new piece is the worktree+reindex snapshot path. Cache-by-resolved-sha; LRU 5/500 MiB; mutual-exclusive with --baseline; per-delta override compatible. Hard error on non-git projects (no graceful fallback — there's no meaningful 'ref' without git). Plan only — implementation follows after CodeRabbit review per the impact (#49→#50) / watch (#46→#47) workflow. * docs(plan): address CodeRabbit findings — atomic cache populate (D11), worktree-as-cache lifecycle clarity, TS type widening callout - D1/D2/D8 rewritten: worktree IS the cache entry (kept until LRU evicts); cleanup runs only on reindex failure rollback OR LRU eviction. The earlier ambiguity (D2 said 'cache by sha' while D8 said 'remove in finally') is resolved. - D11 added: atomic cache populate via per-pid temp dir + POSIX rename → free single-flight semantics; no lock files needed. Same pattern for eviction. Closes the race CodeRabbit flagged on concurrent CI matrix runs against the same sha. - AuditBase TS type widening to discriminated union called out explicitly above the Decisions table (Tracer 1 ships it). - CODEMAP_AUDIT_CACHE_SIZE env var mention dropped — was promising an unimplemented config knob; v1 hardcodes the limits, defer to v1.x+.
SutuSebastian
added a commit
that referenced
this pull request
May 3, 2026
…f + 2 nitpicks)
Real bug:
- main.ts: runListBaselinesCmd was being called without stateDir — `codemap query --baselines` would fall back to the default DB instead of the caller-selected one. Fixed.
Stale doc refs:
- audit-worktree.ts: 3 JSDoc strings still said `.codemap.db` after Tracer 1's CACHE_ENTRY_DB_REL move; bumped to `.codemap/index.db`.
- bootstrap.ts: printCliUsage() had two `.codemap.db` refs + missing --state-dir/CODEMAP_STATE_DIR docs in Environment+Options. Fixed.
- config.ts: Zod databasePath.describe() said default was `<root>/.codemap.db`; corrected to `<state-dir>/index.db`.
- .agents/skills + templates skills: 2 hard-coded `.codemap/` refs reworded to `<state-dir>/` with `(default .codemap/)` callout (state-dir is configurable).
Nitpicks applied:
- state-dir.test.ts: dropped redundant `require('node:fs')` for mkdirSync (already imported).
- bootstrap-codemap.ts: consolidated two single-import lines from state-dir into one statement.
Nitpicks declined:
- changeset code-fence missing 'text' lang — purely cosmetic.
- cmd-index.ts JSDoc on runIndexCmd — 'all public APIs need JSDoc' is a fabricated rule (sibling cmds inconsistent; same hallucination rejected on PR #50).
Merged
2 tasks
SutuSebastian
added a commit
that referenced
this pull request
May 3, 2026
…es (impl of PR #53 plan) (#54) * feat(state-dir): resolveStateDir + DB at <state-dir>/index.db (Tracer 1) - application/state-dir.ts: resolveStateDir({root, cliFlag, env}) per plan §D7. Constants STATE_DIR_DEFAULT='.codemap', STATE_DB_NAME='index.db'. 12 unit tests cover precedence + relative/absolute paths. - config.ts: ResolvedCodemapConfig gains stateDir; resolveCodemapConfig 3rd arg opts.stateDir; databasePath defaults to <stateDir>/index.db. User-supplied databasePath wins (escape hatch). - config.ts: loadUserConfig reads <state-dir>/config.{ts,js,json} (D8); legacy <root>/codemap.config.* dropped (pre-v1). - runtime.ts: getStateDir() getter. - bootstrap.ts: --state-dir + CODEMAP_STATE_DIR; precedence per D7. - bootstrap-codemap.ts (new): single helper extracts the loadUserConfig+resolveCodemapConfig+initCodemap+configureResolver dance from 9 cmd-* files. Tracer 4's ensureStateDir attaches here. - All 9 cmd-* files refactored; stateDir threaded through interfaces + main.ts dispatch + ServerOpts (mcp/serve). - audit-worktree.ts: cached entries at <sha>/.codemap/index.db (recursive layout — each cached worktree is its own self-contained codemap project). - audit-engine.ts: makeWorktreeReindex stops hard-coding db path; openCodemapDatabase() reads from initialised runtime. - sqlite-db.ts: openCodemapDatabase mkdirs the parent (state-dir may not exist on fresh project). Dogfood: - .codemap/.gitignore (self-managed, blacklist) — codemap repo + fixtures/minimal/ - root .gitignore: dropped .codemap.* and .codemap/audit-cache/ (nested .gitignore handles them) - fixtures/minimal/.codemap.db* removed (stale legacy) 703 tests pass. * feat(state-dir): ensureStateGitignore reconciler — self-healing (Tracer 2) - STATE_GITIGNORE_BODY constant — single source of truth for the canonical blacklist. - ensureStateGitignore(stateDir) — pure-shape return ({before, after, written}); idempotent (no write on steady state); auto-mkdir; user-edits rewritten back per D11 (file is codemap-managed; header line declares it). - 5 tests cover: fresh write, idempotent, user-modified, older-version self-heal, returned shape matches disk. Bumping STATE_GITIGNORE_BODY in a future PR is the entire migration — every consumer's project repairs itself on next codemap run. * feat(state-dir): ensureStateConfig reconciler (Tracer 3) * feat(state-dir): bootstrap orchestrator + agents-init delegation (Tracer 4) * docs(state-dir): sync README + architecture + glossary + agent rule/skill + changeset (Tracer 5) - README.md: --state-dir flag, config-file location, self-healing callout - docs/architecture.md: state-dir resolver in src/config.ts intro; bulk .codemap.db → .codemap/index.db; User config section rewritten with self-healing details (ensureStateGitignore + ensureStateConfig from src/application/, single attachment point in cli/bootstrap-codemap.ts) - docs/glossary.md: new entries for '.codemap/' / <state-dir> / CODEMAP_STATE_DIR; '.codemap/index.db'; '.codemap/.gitignore' / self-healing files - docs/roadmap.md: drop the consolidated-dir backlog item (it shipped) - docs/research/fallow.md: Adjacent-shipped bullet referencing PR #53 (plan) + #54 (impl) - .agents/rules/codemap.md + templates/agents/rules/codemap.md: bulk .codemap.db → .codemap/index.db; one-paragraph addition explaining state-dir + self-healing .gitignore (Rule 10 lockstep) - .agents/skills/codemap/SKILL.md + templates/agents/skills/codemap/SKILL.md: bulk path update (Rule 10 lockstep) - .changeset/codemap-dir-consolidation.md: minor — full design rationale + cleanup steps - docs/plans/codemap-dir-consolidation.md: deleted per docs-governance Rule 3 * docs(state-dir): refresh stale path refs across docs/ + slim self-authored comments Doc staleness sweep (after Tracer 5): - docs/glossary.md, docs/agents.md, docs/benchmark.md, docs/why-codemap.md, docs/research/competitive-scan-2026-04.md, docs/research/fallow.md (B.6 row): bulk `.codemap.db` → `.codemap/index.db` everywhere except the intentional 'old → new' migration callouts. - docs/architecture.md, docs/research/fallow.md, docs/packaging.md: `codemap.config.{ts,json}` → `<state-dir>/config.{ts,js,json}`. - docs/agents.md § Git: rewritten to describe the self-managed <state-dir>/.gitignore reconciler instead of root-.gitignore patching. - docs/benchmark.md: 'where the DB lives' updated; manual .gitignore note dropped (reconciler handles it). Concise-comments sweep on this turn's authored comments: - src/application/state-config.ts: 2 inline comments slimmed (TS/JS-validation-only and passthrough-rationale). * fix(state-dir): address CodeRabbit findings (1 inline + 4 outside-diff + 2 nitpicks) Real bug: - main.ts: runListBaselinesCmd was being called without stateDir — `codemap query --baselines` would fall back to the default DB instead of the caller-selected one. Fixed. Stale doc refs: - audit-worktree.ts: 3 JSDoc strings still said `.codemap.db` after Tracer 1's CACHE_ENTRY_DB_REL move; bumped to `.codemap/index.db`. - bootstrap.ts: printCliUsage() had two `.codemap.db` refs + missing --state-dir/CODEMAP_STATE_DIR docs in Environment+Options. Fixed. - config.ts: Zod databasePath.describe() said default was `<root>/.codemap.db`; corrected to `<state-dir>/index.db`. - .agents/skills + templates skills: 2 hard-coded `.codemap/` refs reworded to `<state-dir>/` with `(default .codemap/)` callout (state-dir is configurable). Nitpicks applied: - state-dir.test.ts: dropped redundant `require('node:fs')` for mkdirSync (already imported). - bootstrap-codemap.ts: consolidated two single-import lines from state-dir into one statement. Nitpicks declined: - changeset code-fence missing 'text' lang — purely cosmetic. - cmd-index.ts JSDoc on runIndexCmd — 'all public APIs need JSDoc' is a fabricated rule (sibling cmds inconsistent; same hallucination rejected on PR #50).
2 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Implementation of PR #49's
codemap impactplan. Single verb walks the calls / dependencies / imports graphs (callers, callees, dependents, dependencies) — replaces the "agent composesWITH RECURSIVEby hand" tax that's the highest-friction loop left after watch (#47) shipped.Three transports, one engine: CLI (
codemap impact <target>), MCP (impacttool), HTTP (POST /tool/impact). All three dispatch the same purefindImpactengine inapplication/impact-engine.tsper the post-PR #41 layering.Tracers
application/impact-engine.ts— pure walker. OneWITH RECURSIVEper (direction, backend) combo, JS-side merge + dedup + summary. Cycle detection via path-stringinstrcheck (per plan §D6). 27 unit tests cover symbol + file targets, up/down/both, depth caps, limit caps, self-loop + 3-cycle,--viaskipped_backends, case-sensitive name match.cmd-impact.ts— CLI verb mirroringcmd-show.tsshape. Parser for--direction/--via/--depth/--limit/--summary/--json. Wired intomain.ts+bootstrap.ts+printCliUsage. 14 parser tests.impacttool —impactArgsSchema+handleImpactintool-handlers.ts;registerImpactToolone-line wrapper inmcp-server.ts. 5 integration tests (tools/list, symbol target, file target, summary trim semantics, Zod rejects float depth).POST /tool/impact—TOOL_NAMESgainsimpact; one switch arm indispatchToolwith Zod validation. 4 integration tests (happy path, missing target, non-int depth, bad direction enum).docs/architecture.md(new "Impact wiring" section),docs/glossary.md,docs/research/fallow.mdadjacent-shipped bullet,.agents/+templates/agents/rule + skill (Rule 10 lockstep), changeset, plan deleted (docs-governance Rule 3).Smoke
```bash
$ bun src/index.ts impact handleQuery --json --summary | jq '.summary'
{
"nodes": 65,
"max_depth_reached": 3,
"by_kind": { "symbol": 65 },
"terminated_by": "depth"
}
$ bun src/index.ts impact src/db.ts --direction up --depth 1
impact file:src/db.ts direction=up via=dependencies+imports depth=1
[1] up depended_on_by src/api.ts
[1] up depended_on_by src/application/audit-engine.ts
... (real fan-in across the codebase)
```
Decisions worth knowing (full set in changeset)
/or matchesfiles.path→ file; otherwise symbol. Symbol → walkscalls; file → walksdependencies+imports. Mismatched explicit--via→skipped_backends(no error).instrcheck inside the recursive CTE. Bounded depth +--limit(default 500) keep cyclic graphs cheap.summary.terminated_by:limit>depth>exhausted.--summaryshape. Trimsmatchesfor transport bandwidth but preservessummary.nodescount — thejq '.summary.nodes'CI-gate pattern still works.Test plan
WITH RECURSIVESQL shape, dedup logic, termination classification heuristicSummary by CodeRabbit
codemap impactcommand for blast-radius analysis of symbols and files across dependency and call graphsPOST /tool/impact), and MCP tool interface