fix(client): reject message port link calls after the port closes

[dinwwwh]

Calls made through the message port RPCLink after the port closes now reject with AbortError('MessagePort closed') instead of hanging forever. In-flight calls cut off by the close now reject with the same reason instead of the generic "Peer was closed". This is the message port counterpart of #2081.

Fixes

• Web MessagePort, Node worker_threads ports, and Electron MessagePortMain silently drop postMessage once closed, so any later call stayed pending unless the caller passed an abort signal. Those calls now reject right away and nothing is posted to the dead port.
• Browser extension ports already rejected, because postMessage throws on a disconnected port in both Chrome and Firefox. A remote disconnect now surfaces the same AbortError as the other port kinds.

Testing

• New test: a call made after close rejects with AbortError('MessagePort closed') and never posts a message. It timed out before the fix.
• The existing close test now asserts the exact abort reason.
• Checked Node 24 directly: after close, postMessage drops messages without throwing and close fires on both ends. Electron, Chrome, and Firefox behavior was checked against their source.

A client that calls close() on its own browser MessagePort can still hang, because the HTML spec fires close only at the other end of the channel. Node and Electron fire it on both ends.

[pkg-pr-new]

More templates

• @orpc/bun-playground
• @orpc/cloudflare-playground
• @orpc/nest-playground
• @orpc/next-playground

@orpc/ai-sdk

npm i https://pkg.pr.new/@orpc/ai-sdk@2082

@orpc/arktype

npm i https://pkg.pr.new/@orpc/arktype@2082

@orpc/bun

npm i https://pkg.pr.new/@orpc/bun@2082

@orpc/client

npm i https://pkg.pr.new/@orpc/client@2082

@orpc/cloudflare

npm i https://pkg.pr.new/@orpc/cloudflare@2082

@orpc/contract

npm i https://pkg.pr.new/@orpc/contract@2082

@orpc/experimental-effect

npm i https://pkg.pr.new/@orpc/experimental-effect@2082

@orpc/evlog

npm i https://pkg.pr.new/@orpc/evlog@2082

@orpc/hibernation

npm i https://pkg.pr.new/@orpc/hibernation@2082

@orpc/json-schema

npm i https://pkg.pr.new/@orpc/json-schema@2082

@orpc/experimental-lock

npm i https://pkg.pr.new/@orpc/experimental-lock@2082

@orpc/experimental-msw

npm i https://pkg.pr.new/@orpc/experimental-msw@2082

@orpc/nest

npm i https://pkg.pr.new/@orpc/nest@2082

@orpc/next

npm i https://pkg.pr.new/@orpc/next@2082

@orpc/node

npm i https://pkg.pr.new/@orpc/node@2082

@orpc/openapi

npm i https://pkg.pr.new/@orpc/openapi@2082

@orpc/opentelemetry

npm i https://pkg.pr.new/@orpc/opentelemetry@2082

@orpc/pinia-colada

npm i https://pkg.pr.new/@orpc/pinia-colada@2082

@orpc/pino

npm i https://pkg.pr.new/@orpc/pino@2082

@orpc/publisher

npm i https://pkg.pr.new/@orpc/publisher@2082

@orpc/ratelimit

npm i https://pkg.pr.new/@orpc/ratelimit@2082

@orpc/server

npm i https://pkg.pr.new/@orpc/server@2082

@orpc/shared

npm i https://pkg.pr.new/@orpc/shared@2082

@orpc/swr

npm i https://pkg.pr.new/@orpc/swr@2082

@orpc/tanstack-query

npm i https://pkg.pr.new/@orpc/tanstack-query@2082

@orpc/trpc

npm i https://pkg.pr.new/@orpc/trpc@2082

@orpc/valibot

npm i https://pkg.pr.new/@orpc/valibot@2082

@orpc/zod

npm i https://pkg.pr.new/@orpc/zod@2082

commit: 6465c5f

[codspeed]

Merging this PR will not alter performance

✅ 30 untouched benchmarks

---

_{Comparing dinwwwh:claude/hopeful-benz-efe7b5 (6465c5f) with main (29285f6)}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[pullfrog]

✅ No new issues found.

Reviewed changes

• Reject calls after port close — MessagePortLinkTransport now records an AbortError('MessagePort closed') when the port's close/disconnect event fires and throws it from the ClientPeer send callback, so post-close calls settle immediately instead of posting to a dead port and hanging.
• In-flight reason — the close handler passes that reason into peer.close(reason), so in-flight requests are cut off with MessagePort closed rather than the generic Peer was closed, matching the WebSocket transport.
• Tests — new rejects calls made after close asserts the exact error and that postMessage is never called; the existing pending-request test now pins the exact reason. I verified the new test hangs (times out) with the guard removed, and that the full message-port suite (19 tests), eslint, and tsc -b packages/client all pass.

Known limitation carried forward (already documented in the PR body): a browser MessagePort closed by the client itself never fires a local close event, so that specific path can still hang; Node/Electron fire on both ends.

  ^{｜ View workflow run ｜ Using DeepSeek Flash (free via Pullfrog for OSS) ｜ 𝕏}