Documentation
¶
Overview ¶
Package textutil holds small, security-relevant string/byte helpers shared across packages that must never disagree on how they cut text. The package exists so a rune-boundary cut has exactly ONE implementation in the tree (T-01-37): before this package existed, internal/mcp/session_line.go's truncateOnRuneBoundary was unexported and therefore not importable, and a second, hand-copied implementation in internal/uiserver would have been a second security-relevant copy of a cut that must never split a rune.
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func TruncateOnRuneBoundary ¶
TruncateOnRuneBoundary returns the prefix of s that is at most maxBytes long, walking backward from maxBytes to the nearest rune boundary so a multi-byte UTF-8 sequence is never split. Moved here, unchanged, from internal/mcp/session_line.go (originally unexported as truncateOnRuneBoundary) so internal/mcp and internal/uiserver share one implementation rather than each carrying their own copy of a cut that is security-relevant precisely because it must never split a rune.
The guarantee this function makes is SCOPED, not universal: if s was valid UTF-8, the returned prefix is valid UTF-8 too, because the cut never lands mid-rune. If s was not valid UTF-8 to begin with, the returned prefix is neither repaired nor rejected — it is simply cut at the nearest byte that looks like a rune-start byte (or an ASCII byte), which is the best a byte-oriented cut can promise over already-invalid input.
Types ¶
This section is empty.
Source Files