fix(peer): finalize hibernation iterators instead of throwing on cleanup

[dinwwwh]

Cleaning up a HibernationAsyncIteratorClass (via return, throw, or async dispose) no longer throws, and the server now finalizes the iterator right after the hibernation callback runs, so it always ends up properly closed.

Fixes

• HibernationAsyncIteratorClass cleanup was throwing "Cannot use hibernating iterator directly" on cancellation, so any generic code path that closed the iterator would blow up. Only direct iteration via next() still throws.
• ServerPeer now calls return() on the iterator in a finally after invoking the hibernation callback, so the iterator is finalized even when the callback rejects; the rejection still propagates and cancels the request as before.

Testing

• Server tests assert the iterator reports { done: true } after handling, for both successful and rejecting callbacks.
• All peer package tests pass and typecheck is clean.

[pkg-pr-new]

• @standardserver/hono-node-playground
• @standardserver/node-http-playground

@standardserver/aws-lambda

npm i https://pkg.pr.new/@standardserver/aws-lambda@49

@standardserver/bun

npm i https://pkg.pr.new/@standardserver/bun@49

@standardserver/core

npm i https://pkg.pr.new/@standardserver/core@49

@standardserver/deno

npm i https://pkg.pr.new/@standardserver/deno@49

@standardserver/fastify

npm i https://pkg.pr.new/@standardserver/fastify@49

@standardserver/fetch

npm i https://pkg.pr.new/@standardserver/fetch@49

@standardserver/node

npm i https://pkg.pr.new/@standardserver/node@49

@standardserver/peer

npm i https://pkg.pr.new/@standardserver/peer@49

@standardserver/shared

npm i https://pkg.pr.new/@standardserver/shared@49

commit: 471ee2a

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[codspeed-hq]

Merging this PR will not alter performance

✅ 26 untouched benchmarks
⏩ 108 skipped benchmarks¹

---

_{Comparing claude/hibernation-async-iterator-cleanup-8374c8 (471ee2a) with main (fada571)}

Footnotes

1. 108 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩