Skip to content

Release Notes - v0.9.2

Title: Harden parsing and data normalization

Version: v0.9.2

Release: v0.9.2

Comparison: v0.9.1…v0.9.2

Date: 2026-10-11

NetworkCompatible
mainnet⛔
preprod-testnet✅
preview-testnet✅
musashi✅

This release distinguishes embedded NUL characters from the end of input, prevents typed nil data containers from causing normalization panics, and adds licensing attribution to the Go source files.

  • Improved parser validation so syn.Parse and ParseWithContext reject valid prefixes followed by embedded U+0000 or by trailing text instead of accepting only the prefix. The parser preserves U+0000 in string literals, skips it in comments, and rejects it in byte strings. Applications that relied on U+0000 to truncate input must update their inputs.
  • Protected data normalization from nil pointer panics when it receives typed nil *Constr, *Map, or *List values. data.Normalize now returns those values unchanged, while nil *Value behavior remains unchanged.
  • Added the standard Blink Labs Apache 2.0 copyright header to 137 Go files, strengthening licensing attribution and compliance without changing runtime or API behavior.
  • Thanked @arepala-uml, who made their first contribution in PR #474.

Doc Holiday logo

Docs authored by Doc Holiday

Edit page

Last updated: