fix(client): exclude aborted subrequests from batch messages

[dinwwwh]

Fixes #2172

Summary

Fixes batch link plugin to exclude subrequests that are aborted before the batch is sent to the transport. This prevents the server from counting cancelled requests against its batch size limit and avoids executing calls the client has already rejected.

Changes

• Filter pending messages before sending: Only include messages for subrequests that are still open (not aborted) when the batch is handed to the transport
• Apply to both GET and POST: The filtering works for both HTTP methods — GET encodes messages in the URL query parameter, POST in the request body
• Preserve batch integrity after send: Added a new test case to ensure that subrequests aborted after the batch is sent don't affect the already-sent batch messages

Implementation Details

• Introduced outgoingMessages variable that filters pendingMessages to only include those with IDs in openRequestIds
• This filtering happens right before the batch request is constructed, ensuring we only send messages for active subrequests
• The openRequestIds Set is already maintained by the plugin to track which subrequests are still pending, so this reuses existing tracking logic

https://claude.ai/code/session_01KbSdQ6v2aYW1fT6FmbkS7g

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Preview URL	Updated (UTC)
✅ Deployment successful! View logs	orpc	0e1fc5d	Commit Preview URL Branch Preview URL	Oct 04 2026, 07:45 AM

[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

Reviewed the full diff (2 files, 2 commits) and traced the abort lifecycle through ClientPeer (@standard-server/peer) and BatchHandlerPlugin.

• Filtered batch payload — executeBatch now snapshots outgoingMessages = pendingMessages.filter(m => openRequestIds.has(m.id)) after isBatchSent = true, so a subrequest aborted (and its cancel message) before send drops out of both the GET data query param and the POST/QUERY body.
• Boundary is isBatchSent — cancels arriving later are still tracked via cancelledRunningRequestIds and only abort the whole transport request when every open subrequest is cancelled, so an in-flight batch is never rewritten.
• Tests — the pre-send abort case is now parameterized over POST/GET and asserts the batch carries only the open request; a new test aborts inside transport.send and pins that the sent messages stay unchanged.

The openRequestIds-keyed filter is sound: every open id is guaranteed to have a matching request message (the id is added before markRequestSent, removed only by a cancel), and cancels for open requests cannot exist, so filtering is equivalent to keeping exactly the still-open request messages in order. pendingMessages is no longer referenced after the snapshot, so later pushes cannot mutate the payload.

Verified locally: 59/59 batch.test.ts tests pass, pnpm --filter @orpc/client type:check passes, eslint clean.

The pre-send abort test genuinely fails against the old payload ([['request','/b']] vs. ['request','cancel','request']), so the coverage is real. QUERY is exercised via the shared else body branch even though the parameterized test only names POST/GET.

  ^{｜ View workflow run ｜ Using deepseek-v4.1-flash (free via Pullfrog for OSS) ｜ 𝕏}

[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@2176

@orpc/arktype

npm i https://pkg.pr.new/@orpc/arktype@2176

@orpc/bun

npm i https://pkg.pr.new/@orpc/bun@2176

@orpc/client

npm i https://pkg.pr.new/@orpc/client@2176

@orpc/cloudflare

npm i https://pkg.pr.new/@orpc/cloudflare@2176

@orpc/contract

npm i https://pkg.pr.new/@orpc/contract@2176

@orpc/experimental-effect

npm i https://pkg.pr.new/@orpc/experimental-effect@2176

@orpc/evlog

npm i https://pkg.pr.new/@orpc/evlog@2176

@orpc/hibernation

npm i https://pkg.pr.new/@orpc/hibernation@2176

@orpc/json-schema

npm i https://pkg.pr.new/@orpc/json-schema@2176

@orpc/experimental-lock

npm i https://pkg.pr.new/@orpc/experimental-lock@2176

@orpc/experimental-msw

npm i https://pkg.pr.new/@orpc/experimental-msw@2176

@orpc/nest

npm i https://pkg.pr.new/@orpc/nest@2176

@orpc/next

npm i https://pkg.pr.new/@orpc/next@2176

@orpc/node

npm i https://pkg.pr.new/@orpc/node@2176

@orpc/openapi

npm i https://pkg.pr.new/@orpc/openapi@2176

@orpc/opentelemetry

npm i https://pkg.pr.new/@orpc/opentelemetry@2176

@orpc/pinia-colada

npm i https://pkg.pr.new/@orpc/pinia-colada@2176

@orpc/pino

npm i https://pkg.pr.new/@orpc/pino@2176

@orpc/publisher

npm i https://pkg.pr.new/@orpc/publisher@2176

@orpc/ratelimit

npm i https://pkg.pr.new/@orpc/ratelimit@2176

@orpc/server

npm i https://pkg.pr.new/@orpc/server@2176

@orpc/shared

npm i https://pkg.pr.new/@orpc/shared@2176

@orpc/swr

npm i https://pkg.pr.new/@orpc/swr@2176

@orpc/tanstack-query

npm i https://pkg.pr.new/@orpc/tanstack-query@2176

@orpc/trpc

npm i https://pkg.pr.new/@orpc/trpc@2176

@orpc/valibot

npm i https://pkg.pr.new/@orpc/valibot@2176

@orpc/zod

npm i https://pkg.pr.new/@orpc/zod@2176

commit: 0e1fc5d

[codspeed]

Merging this PR will improve performance by 12.02%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 1 improved benchmark
✅ 29 untouched benchmarks

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	serve deeply nested encoded path	1.2 ms	1 ms	+12.02%

Tip

Curious why performance improved? Comment @codspeedbot explain why performance improved on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing claude/dazzling-gates-jdyqmh (0e1fc5d) with main (161cdc0)}