Fix truncated container file reads

[Brandon Waterloo [MSFT] (bwateratmsft)]

🤖

Summary

• Honor backpressure and wait for archive consumption and extracted output to finish before returning container file bytes.
• Parse tar entries in memory instead of using disk extraction, and propagate failed or incomplete reads without returning partial content.
• Preserve the existing VS Code filesystem-copy download path and Windows raw-byte reads.

Fixes #631.

Validation

• 22 focused filesystem-provider and tar-extraction regressions pass, including 30 MB size/SHA-256 comparisons, slow destination draining, filesystem copy, opened files, and failure cases.
• Extension build/type check and touched-file lint pass.
• Manual testing by the maintainer is pending; this PR is intentionally a draft.

UPDATE: I did the manual testing, all green.

[Copilot]

🔵 Needs a closer look

The stream lifecycle and failure-propagation changes are well tested but warrant final maintainer review.

0 open findings

What changed in this PR

Fixes truncated container downloads by correctly awaiting stream completion and tar extraction.

Changes:

• Adds backpressure-aware in-memory tar extraction.
• Propagates incomplete and failed reads.
• Adds comprehensive Linux, Windows, large-file, and failure regression tests.

File	Description
src/​utils/​tarUtils.ts	Implements awaited in-memory tar parsing.
src/​runtimes/​files/​ContainerFilesProvider.ts	Pipelines container output through extraction or raw accumulation.
src/​test/​utils/​tarUtils.test.ts	Tests extraction lifecycle and failures.
src/​test/​runtimes/​files/​ContainerFilesProvider.test.ts	Tests complete reads and filesystem operations.

🧠 Review effort: Balanced

---

💡 Configure MCP servers for context-aware, tailored reviews. Learn more in the docs.