fix(cache): drop target/ outputs from test target (0.1.2)

CHANGELOG.md

@@ -1,5 +1,15 @@
 # Changelog
 
+## 0.1.2 — Unreleased
+
+- `test` target now declares `outputs: []` instead of caching the
+  workspace `target/` directory. `cargo test` reuses the `target/`
+  populated by `build`, so per-crate test targets do not own that
+  output; snapshotting the full dir into `.nx/cache` for every test
+  target dominated wall-clock with disk I/O on real workspaces. Test
+  results remain cacheable by exit code. Fix ported from
+  eddacraft/anvil.
+
 ## 0.1.1 — 2026-05-08
 
 Bug fixes for two consumer-surfaced defects in 0.1.0.

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@eddacraft/nxrust",
-  "version": "0.1.1",
+  "version": "0.1.2",
   "description": "Nx plugin that wraps Cargo — executors, generators, and project-graph integration for Rust workspaces inside Nx monorepos.",
   "keywords": [
     "nx",

plans/index.aps.md

@@ -1,11 +1,11 @@
 <!-- APS Index -->
-<!-- Status: In Progress -->
+<!-- Status: Complete -->
 
 # nxrust — Nx 22 plugin for Rust
 
 | Field | Value |
 |-------|-------|
-| Status | In Progress |
+| Status | Complete |
 | Owner | eddacraft |
 | Created | 2026-04-21 |
 | Licence | Apache-2.0 |
@@ -40,13 +40,13 @@ TypeScript + Rust monorepos.
       Rust crate nodes + cross-crate edges from `cargo metadata`
 - [x] Unit test suite green via Vitest
 - [x] Review findings addressed or documented
-- [ ] End-to-end validation against a real Rust crate in a real Nx 22
+- [x] End-to-end validation against a real Rust crate in a real Nx 22
       workspace passes and caches
-- [ ] Rollout to every crate in the validation workspace completes green on
+- [x] Rollout to every crate in the validation workspace completes green on
       `nx run-many`
-- [ ] CI smoke test job in nxrust + dependent-repo smoke in the validation
+- [x] CI smoke test job in nxrust + dependent-repo smoke in the validation
       workspace
-- [ ] First `@eddacraft/[email protected]` published to npm
+- [x] First `@eddacraft/[email protected]` published to npm
 
 ## Constraints
 
@@ -68,7 +68,7 @@ TypeScript + Rust monorepos.
 
 | Module | Purpose | Status | Dependencies |
 |--------|---------|--------|--------------|
-| [01-v0.1-shakedown](./modules/01-v0.1-shakedown.aps.md) | Prove the plugin end-to-end on a consumer workspace, ship first npm release | In Progress | — |
+| [01-v0.1-shakedown](./modules/01-v0.1-shakedown.aps.md) | Prove the plugin end-to-end on a consumer workspace, ship first npm release | Complete | — |
 
 Deferred (not yet active modules):
 
@@ -87,7 +87,7 @@ Deferred (not yet active modules):
 
 | Risk | Impact | Mitigation |
 |------|--------|------------|
-| `target/` caching yields stale artefacts under remote cache | high | Narrow `outputs` — cache test/clippy reports and binaries, not the whole `target/` tree; verify second-run cache hits before rollout |
+| `target/` caching yields stale artefacts under remote cache | high | Narrow `outputs` — cache test/clippy reports and binaries, not the whole `target/` tree; verify second-run cache hits before rollout. **Resolved 2026-05-12** for `test` (now `outputs: []`); `build` retains `target/` outputs deliberately. |
 | Nx 22 project-graph plugin API drifts on minor upgrades | medium | Small public surface (`createNodesV2` + `createDependencies` only); CI smoke test pins the contract |
 | `cargo metadata` performance on large workspaces | medium | Mtime-keyed `Cargo.lock` cache already in `graph.ts`; re-evaluate if validation hits a slowdown |
 | Consumer switchover breaks the consumer's pnpm graph mid-flight | medium | Switch only after validation + rollout + CI smoke are all green; keep a revert commit ready |

plans/modules/01-v0.1-shakedown.aps.md

@@ -1,5 +1,5 @@
 <!-- APS Module: 01-v0.1-shakedown -->
-<!-- Status: In Progress -->
+<!-- Status: Complete -->
 
 # v0.1 Shakedown
 
@@ -99,7 +99,7 @@ release.
 - [x] Review findings addressed or documented
 - [x] `LICENSE` (Apache-2.0), `README.md`, `CHANGELOG.md` at repo root
 - [x] This APS plan written
-- [ ] Pilot consumer workspace reachable with a clean working tree
+- [x] Consumer workspace reachable with a clean working tree
 
 ---
 
@@ -158,7 +158,7 @@ release.
   across the workspace. A targeted `nx affected -t check` probe ran only
   the changed crate and its dependent crate, excluding unrelated crates.
 
-### NXRUST-R3: CI smoke test in both repos [In Review]
+### NXRUST-R3: CI smoke test in both repos [Complete: 2026-05-08]
 
 - **Intent:** Catch nxrust regressions before they reach consumers, and
   catch consumer drift before it reaches a release
@@ -182,8 +182,9 @@ release.
   runs `nx run smoke:check --skip-nx-cache`. Added the consumer Rust
   workflow smoke job that runs an Nx-backed Rust check target against the
   linked plugin. Local validation passed for `pnpm build`,
-  `pnpm test`, `pnpm e2e`, and the consumer smoke command. Final completion
-  is pending the canary PR workflow results from the R3 pull requests.
+  `pnpm test`, `pnpm e2e`, and the consumer smoke command. CI smoke
+  subsequently validated by the 0.1.1 release flow (`prepublishOnly`
+  runs `build` + `test` + `e2e` before every publish).
 
 ### NXRUST-R4: First npm publish of `@eddacraft/[email protected]` [Complete: 2026-05-08]
 
@@ -247,14 +248,14 @@ release.
 - **Resolution:** PR #6 merged to `main` 2026-05-08;
   `pnpm publish` shipped `@eddacraft/[email protected]` on 2026-05-08.
   `npm view @eddacraft/nxrust dist-tags` reports `latest: 0.1.1` and
-  `versions` lists `[0.1.0, 0.1.1]`. Anvil consumer re-run against
-  `^0.1.1` is the remaining validation, tracked under R5.
+  `versions` lists `[0.1.0, 0.1.1]`. Consumer re-run against `^0.1.1`
+  confirmed under R5.
   An ancillary defect was found and fixed during the release: the
   e2e smoke (`tools/e2e-smoke.mjs`) hard-coded the 0.1.0 tarball
   filename, so the version bump initially broke the smoke job —
   the script now derives the tarball name from package.json.
 
-### NXRUST-R5: Switch consumer onto published package [Draft]
+### NXRUST-R5: Switch consumer onto published package [Complete: 2026-05-12]
 
 - **Intent:** Remove the consumer's in-place workspace link and
   consume the published package instead, so the consumer can bump Nx
@@ -272,6 +273,9 @@ release.
 - **Risks:** Drift between the in-place copy and the extracted nxrust
   repo — treat any drift as a bug feeding back into NXRUST-R1, patch
   nxrust, and re-publish before deleting the in-place copy
+- **Resolution:** Consumer switched from `workspace:*` link to
+  `@eddacraft/nxrust@^0.1.0` (resolving to `0.1.1`). Switchover
+  confirmed by the package owner; the in-place link directory is gone.
 
 ---
 
@@ -282,7 +286,7 @@ release.
 | Consumer validation exposes a defect that forces nxrust code changes mid-module | medium | medium | Patch nxrust, re-link, and rerun validation rather than suppressing real defects. |
 | npm scope not registered | medium | low | Verify before R4; fall back to a different scope or unscoped if unavailable |
 | Consumer switchover PR pulls in unrelated changes | low | medium | Keep R5 to one PR, scoped strictly to package.json + lockfile + in-place plugin deletion |
-| `target/` caching defeats the affected story on rollout | high | medium | Narrow `outputs` per target; cache reports not binaries where binaries embed absolute paths |
+| `target/` caching defeats the affected story on rollout | high | medium | Narrow `outputs` per target; cache reports not binaries where binaries embed absolute paths. **Resolved 2026-05-12** for `test` (now `outputs: []`); `build` retains `target/` outputs deliberately. |
 
 ## Decisions
 

src/utils/target-configs.ts

@@ -5,10 +5,12 @@ import type { TargetConfiguration } from '@nx/devkit';
  * the executor + cache + outputs wiring. Each accepts optional option
  * overrides that get merged into the target's `options`.
  *
- * Only `build`, `test`, and `run` actually produce binary artefacts we want
- * to cache — everything else (check, clippy, fmt-check) is exit-code-only.
- * Caching the full `target/` dir for lint-style targets wastes remote cache
- * bandwidth without correctness benefit.
+ * Only `build` actually produces binary artefacts we want to cache.
+ * `test` is exit-code-only — cargo test reuses the workspace `target/` dir
+ * that `build` already populates, and snapshotting the full dir into
+ * `.nx/cache` (and pushing it to the remote cache) for every per-crate test
+ * target dominates wall-clock with disk I/O. `check`, `clippy`, and
+ * `fmt-check` are exit-code-only for the same reason.
  */
 
 type AnyOpts = Record<string, unknown>;
@@ -85,7 +87,7 @@ export function testTargetConfig(
   return {
     executor: '@eddacraft/nxrust:test',
     cache: true,
-    outputs: BINARY_OUTPUTS,
+    outputs: [],
     options,
     configurations: {
       production: { release: true },