fix(cache): don't cache tmp copies after write failure (fix #11452)

[github-actions[bot]]

Originally proposed by @arjunharshana in #11462. Their description is reproduced below.

Pull requests from the community are converted to discussions automatically, where they can be triaged and prioritized. See the pull request policy.

Description

Resolves #11452

When writing a temporary module copy fails, cacheResult() returns the original module result, but the caller was still setting transformResult.__vitestTmp to the failed path. A subsequent fetch could then treat the nonexistent file as cached and fail with ENOENT.

This change only stores the tmp path when the cache operation actually returns a filesystem-backed result.

Added a regression test that simulates an ENOSPC write failure, fetches the same module again, and verifies that the module continues to be served inline without retaining the failed tmp path.

AI-assisted - ChatGPT assisted with the investigation and draft. I reviewed and tested the changes.

Please don't delete this checklist! Before submitting the PR, please make sure you do the following:

• It's really useful if your PR references an issue where it is discussed ahead of time. If the feature is substantial or introduces breaking changes without a discussion, PR might be closed.
• Ideally, include a test that fails without this PR but passes with it.
• Please, don't make changes to pnpm-lock.yaml unless you introduce a new test example.
• Please check Allow edits by maintainers to make review process faster. Note that this option is not available for repositories that are owned by Github organizations.

Tests

• Run the tests with pnpm test:ci.

Documentation

• If you introduce new functionality, document it. You can run documentation with pnpm run docs.

Changesets

• Changes in changelog are generated from PR name. Please, make sure that it explains your changes in an understandable manner. Please, prefix changeset messages with feat:, fix:, perf:, docs:, or chore:.

---

Original implementation from #11462 by @arjunharshana