Module toml 

Simple TOML parser for kernel boot configuration Minimal TOML subset parser for kernel boot configuration.

Zero-copy design: keys, section names and unescaped string values borrow directly from the source buffer (Cow::Borrowed); only values requiring escape-sequence expansion allocate. Entries are stored in one flat, sorted Vec (binary-search lookups) instead of nested maps : the config is parsed once at boot and read-only afterwards.

Lifetime contract

[TomlConfig<'a>] borrows the input buffer: the caller must keep the TOML text alive for as long as the config is used. This is natural for boot configs read from persistent regions (initfs / mapped file).

Supported subset

• [section] headers (no nested tables, non-empty names)
• key = value inside a section (keys before any header are rejected)
• Value types: integers (decimal + 0x hex), booleans, basic strings
• # comments, full-line or trailing, respecting quoted strings
• Basic-string escapes: \n \t \r \" \\

Not supported by design: floats, arrays, inline tables, multi-line strings, dotted keys, duplicate keys (rejected at parse time).

Structs

TomlConfig

Parsed TOML configuration: flat (section, key, value) triples kept sorted for O(log n) binary-search lookups with minimal memory overhead.

Enums

ConfigValue

A single configuration value.

Functions

parse_toml

Parse a TOML-subset document. The returned config borrows data.

strip_comment 🔒

Strip a trailing # comment, ignoring # inside quoted strings.

unescape_cow 🔒

Expand basic-string escapes, borrowing when there is nothing to expand.