From 9cbad2c3aa2fee30ad0f0616decaca6fe95f0ada Mon Sep 17 00:00:00 2001 From: zzstoatzz Date: Thu, 26 Feb 2026 20:47:36 -0600 Subject: [PATCH] fix: correct claim about Go arena allocator Go has an experimental arena package (GOEXPERIMENT=arenas), but it's on hold indefinitely and not recommended for production. Co-Authored-By: Claude Opus 4.6 --- devlog/002-firehose-and-benchmarks.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/devlog/002-firehose-and-benchmarks.md b/devlog/002-firehose-and-benchmarks.md index 20ea909..089a5b1 100644 --- a/devlog/002-firehose-and-benchmarks.md +++ b/devlog/002-firehose-and-benchmarks.md @@ -101,7 +101,7 @@ we include two go implementations: **go (indigo)** uses cbor-gen (code-generated, already reflection-free at the frame level) but pays for go-car's per-block SHA-256 CID verification and cbornode's reflection-based DAG-CBOR decode via the unmaintained refmt library. result: ~15k fps. -the go-raw improvement comes from two things: a faster per-block CBOR library (fxamacker vs refmt) and skipping CID hashing. GC pressure is the fundamental ceiling in Go — every string, byte slice, and decoded value is heap-allocated, and Go has no arena allocator. +the go-raw improvement comes from two things: a faster per-block CBOR library (fxamacker vs refmt) and skipping CID hashing. GC pressure is the fundamental ceiling in Go — every string, byte slice, and decoded value is heap-allocated, and Go's experimental arena package is on hold and not recommended for production. ### python -- 2.51.2