167 wires and counting

$ follow Go

Keep up with Go in about 3 minutes: what actually shipped — the commits, pull requests, releases, and security advisories that matter.

or

fair warning: these emails are deeply technical. diffs, version numbers, CVEs, benchmark deltas. if that's not your idea of a good read, this isn't your newsletter.

Folds into your digest — weekly by default, monthly if you prefer. Unsubscribe in one click.

$ status

wire 2026-10-30
stories 35

© 2026 RepoJournal Home Showcase How it works Privacy

$ the-wire · showcase

Windows SendFile was reporting partial 2 GiB+ transfers as success

By RepoJournal · Filed · About Go · Composed from the cited sources · methodology

The day's real bug is in internal/poll on Windows, where an inverted error check hid TransmitFile failures on large files; the rest of the queue is loong64 codegen work and tooling hygiene.

internal/poll: don't drop TransmitFile errors in Windows SendFile golang/go

by Tobias Klauser

After a partial transfer SendFile seeks the source file to the end of the written data, and the check meant to preserve the transfer error was inverted: a TransmitFile error was replaced by the usually-nil Seek error, while a Seek error after a successful transfer was dropped. So a failure while sending a later chunk of a file larger than 2 GiB was reported as success. Only the Seek error is re...

refactor/satisfy: handle multi-valued and spread args to built-ins golang/tools

by Alan Donovan

Spread calls to built-ins such as delete(f()) were handled wrongly, producing a panic or missing constraints; the fix factors out the tuple-unpacking logic and passes CallExpr.Ellipsis down instead of checking the argument. Matters if you run refactor/satisfy over code that spreads a multi-valued call into a builtin.

crypto/internal/fips140/aes: pre-swap enc[] round keys to state order on loong64 golang/go

by XiaolinZhao

expandKeyAsm now byte-swaps all enc[] round keys into state order once, in a new Phase 2.5, instead of having every consumer (encryptBlockAsm, ctrBlocks1/2/4/8Asm) redo the same VSHUFB swap at each AddRoundKey, once per round per block.

cmd/internal/obj/loong64: optimize the function stacksplit implementation golang/go

by Guoqi Chen

The loong64 stacksplit sequence for framesize <= StackSmall collapses from SGTU SP, stackguard, RT1 followed by a branch into a single BGEU stackguard, SP, morestack. Guoqi Chen's CL covers the same code path for every function call on the platform.

cmd/internal/obj/loong64: add unary destination instruction table golang/go

by Guoqi Chen

JIRL, previously a name with no concrete implementation, is replaced with JAL, the JMP and JAL OpCode definitions are corrected to match the actual loong64 machine-code formats, and JAL gets a plan9 form: JAL offset(RJ), RD. Any out-of-tree loong64 assembler work touching JIRL needs to move. The long tail is routine: a dead RET removed in the chacha8rand loong64 trampoline, a SIMD instruction-i...

Quick answers

What shipped in Go on October 30, 2026?
The day's real bug is in internal/poll on Windows, where an inverted error check hid TransmitFile failures on large files; the rest of the queue is loong64 codegen work and tooling hygiene. In total, 35 commits landed.
Who contributed to Go on October 30, 2026?
7 developers shipped this update, including XiaolinZhao, Guoqi Chen, Tobias Klauser, limeidan, Dmitri Shuralyov, Michael Matloob, and Alan Donovan.
What were the notable Go updates?
internal/poll: don't drop TransmitFile errors in Windows SendFile, refactor/satisfy: handle multi-valued and spread args to built-ins, and crypto/internal/fips140/aes: pre-swap enc[] round keys to state order on loong64.