feat(cache): integrate runner-aware IPC reports

[branchseer]

Replaces the previous PR's placeholder with the real
vite_task_server::serve(...) call: per-task IPC server, napi
addon embedded into the runner binary and materialized to disk on
first use, VP_RUN_NODE_CLIENT_PATH injected into the child so the
JS wrapper can require() it.

Cache integration: Reports collected from the IPC drive
PostRunFingerprint and the cache-update path —

• ignore_input reads → excluded from input fingerprint
• ignore_output writes → excluded from output archive
• tracked_envs (single name) + tracked_env_globs → folded into
  the post-run fingerprint so a value change misses the cache
• disable_cache → skips the cache-update path entirely
  (ToolRequested)
• IPC server bind/runtime failure → IpcServerError cache disable

End-to-end coverage via the ipc_client_test fixture set: one
fixture per API method, each exercising the real Rust ↔ JS path and
asserting the right cache behaviour. Adds vtt (test-only)
helpers grep_file and stat_file that the fixtures need.

Co-Authored-By: Claude Opus 4.7 (1M context) noreply@anthropic.com

[branchseer]

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

• test(cache): real-vite integration fixtures + playground #408
• feat(cache): integrate runner-aware IPC reports #407 👈 (View in Graphite)
• refactor(execute): reshape for IPC server lifecycle #406
• feat(client): napi binding + JS wrapper #405
• feat(ipc): protocol + Rust transport #404
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[chatgpt-codex-connector]

💡 Codex Review

vite-task/crates/vite_task/src/session/execute/mod.rs

Line 482 in b3a5cfa

if metadata.input_config.includes_auto || metadata.output_config.includes_auto {

Start IPC even when auto file tracking is disabled

For cached tasks whose input/output are fully explicit (no { auto: true }), this branch never starts the IPC server or injects VP_RUN_NODE_CLIENT_PATH. Runner-aware calls like getEnv(..., { tracked: true }) and disableCache() therefore silently no-op, even though tracked envs are independent cache inputs and do not require fspy. In that configuration a build can produce output from an env value obtained through the client, then incorrectly hit the cache after the env changes because no report was collected.

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Serve IPC env lookups from the spawned environment

This captures only the runner process environment, but the actual child environment is spawn_execution.spawn_command.all_envs plus prefix envs and runtime injections. When a task uses a command prefix or other planned env override (for example FOO=prod node ... while the parent has FOO=dev or no FOO), getEnv('FOO') returns the parent value/None instead of what the spawned tool sees, so runner-aware tools can compute outputs from the wrong value and record the wrong post-run fingerprint.

Useful? React with 👍 / 👎.

[branchseer]

Superseded by restructured 2-PR stack:

• feat(client): vite-task-client JS package #409 — vite-task-client JS package
• feat(ipc): protocol + Rust transport #410 — Rust crates + runner integration (5 commits, one per closed PR)