AST Data Structures#

The Abstract Syntax Tree (AST) is the core representation of magic rules in libmagic-rs. This chapter provides detailed documentation of the fully implemented AST data structures with comprehensive test coverage (29 unit tests) and their usage patterns.

Overview#

The AST consists of several key types that work together to represent magic rules:

MagicRule: The main rule structure containing all components
OffsetSpec: Specifies where to read data in files
TypeKind: Defines how to interpret bytes
Operator: Comparison and bitwise operations
Value: Expected values for matching
Endianness: Byte order specifications

MagicRule Structure#

The MagicRule struct is the primary AST node representing a complete magic rule:

pub struct MagicRule {
    pub offset: OffsetSpec, // Where to read data
    pub typ: TypeKind, // How to interpret bytes
    pub op: Operator, // Comparison operation
    pub value: Value, // Expected value
    pub message: String, // Human-readable description
    pub children: Vec<MagicRule>, // Nested rules
    pub level: u32, // Indentation level
}

Example Usage#

use libmagic_rs::parser::ast::*;

// ELF magic number rule
let elf_rule = MagicRule {
    offset: OffsetSpec::Absolute(0),
    typ: TypeKind::Long {
        endian: Endianness::Little,
        signed: false
    },
    op: Operator::Equal,
    value: Value::Bytes(vec![0x7f, 0x45, 0x4c, 0x46]), // "\x7fELF"
    message: "ELF executable".to_string(),
    children: vec![],
    level: 0,
};

Hierarchical Rules#

Magic rules can contain child rules that are evaluated when the parent matches:

let parent_rule = MagicRule {
    offset: OffsetSpec::Absolute(0),
    typ: TypeKind::Byte { signed: false },
    op: Operator::Equal,
    value: Value::Uint(0x7f),
    message: "ELF".to_string(),
    children: vec![
        MagicRule {
            offset: OffsetSpec::Absolute(4),
            typ: TypeKind::Byte { signed: false },
            op: Operator::Equal,
            value: Value::Uint(1),
            message: "32-bit".to_string(),
            children: vec![],
            level: 1,
        },
        MagicRule {
            offset: OffsetSpec::Absolute(4),
            typ: TypeKind::Byte { signed: false },
            op: Operator::Equal,
            value: Value::Uint(2),
            message: "64-bit".to_string(),
            children: vec![],
            level: 1,
        },
    ],
    level: 0,
};

OffsetSpec Variants#

The OffsetSpec enum defines where to read data within a file:

Absolute Offsets#

pub enum OffsetSpec {
    /// Absolute offset from file start
    Absolute(i64),
    // ... other variants
}

Examples:

// Read at byte 0 (file start)
let start = OffsetSpec::Absolute(0);

// Read at byte 16
let offset_16 = OffsetSpec::Absolute(16);

// Read 4 bytes before current position (negative offset)
let relative_back = OffsetSpec::Absolute(-4);

Indirect Offsets#

Indirect offsets read a pointer value and use it as the actual offset:

Indirect {
    base_offset: i64, // Where to read the pointer
    pointer_type: TypeKind, // How to interpret the pointer
    adjustment: i64, // Value to add to pointer
    endian: Endianness, // Byte order for pointer
}

Example:

// Read a 32-bit little-endian pointer at offset 0x20,
// then read data at (pointer_value + 4)
let indirect = OffsetSpec::Indirect {
    base_offset: 0x20,
    pointer_type: TypeKind::Long {
        endian: Endianness::Little,
        signed: false
    },
    adjustment: 4,
    endian: Endianness::Little,
};

Relative and FromEnd Offsets#

// Relative to previous match position
Relative(i64),

// Relative to end of file
FromEnd(i64),

Examples:

// 8 bytes after previous match
let relative = OffsetSpec::Relative(8);

// 16 bytes before end of file
let from_end = OffsetSpec::FromEnd(-16);

TypeKind Variants#

The TypeKind enum specifies how to interpret bytes at the given offset:

Breaking Change in v0.2.0: The Byte variant changed from a unit variant (Byte) to a struct variant (Byte { signed: bool }). Code that pattern-matches exhaustively on TypeKind requires updates.

Numeric Types#

pub enum TypeKind {
    /// Single byte (8-bit)
    Byte { signed: bool },

    /// 16-bit integer
    Short { endian: Endianness, signed: bool },

    /// 32-bit integer
    Long { endian: Endianness, signed: bool },

    /// 64-bit integer
    Quad { endian: Endianness, signed: bool },

    /// String data
    String { max_length: Option<usize> },

    /// Pascal string (length-prefixed)
    PString {
        max_length: Option<usize>,
        length_width: PStringLengthWidth,
        length_includes_itself: bool,
    },

    /// Regular expression pattern matching
    Regex {
        flags: RegexFlags,
        count: RegexCount,
    },

    /// Bounded literal byte sequence search
    Search {
        range: NonZeroUsize,
    },
}

Examples:

// Single unsigned byte
let byte_type = TypeKind::Byte { signed: false };

// Single signed byte
let signed_byte_type = TypeKind::Byte { signed: true };

// 16-bit little-endian unsigned integer
let short_le = TypeKind::Short {
    endian: Endianness::Little,
    signed: false
};

// 32-bit big-endian signed integer
let long_be = TypeKind::Long {
    endian: Endianness::Big,
    signed: true
};

// 64-bit little-endian unsigned integer
let quad_le = TypeKind::Quad {
    endian: Endianness::Little,
    signed: false
};

// 64-bit big-endian signed integer
let quad_be = TypeKind::Quad {
    endian: Endianness::Big,
    signed: true
};

// Null-terminated string, max 256 bytes
let string_type = TypeKind::String {
    max_length: Some(256)
};

PString (Pascal String)#

Pascal-style length-prefixed strings where the length prefix can be 1, 2, or 4 bytes depending on the length_width field.

Structure:

Length prefix: 1, 2, or 4 bytes indicating string length, with configurable endianness
String data: The number of bytes specified by the length prefix

Example:

0 pstring JPEG
0 pstring/H JPEG

The first line reads a 1-byte length prefix (default), then reads that many bytes as a string. The second line reads a 2-byte big-endian length prefix.

Behavior:

Returns Value::String containing the string data (without the length prefix)
Performs bounds checking on both the length prefix and the string data
Supports all string comparison operators
Length prefix width controlled by PStringLengthWidth enum
Optional /J flag indicates JPEG-style self-inclusive length (stored length includes the prefix itself)

PStringLengthWidth Enum#

The PStringLengthWidth enum specifies the width and endianness of the length prefix:

pub enum PStringLengthWidth {
    /// 1-byte length prefix (default, `/B` suffix)
    OneByte,
    /// 2-byte big-endian length prefix (`/H` suffix)
    TwoByteBE,
    /// 2-byte little-endian length prefix (`/h` suffix)
    TwoByteLE,
    /// 4-byte big-endian length prefix (`/L` suffix)
    FourByteBE,
    /// 4-byte little-endian length prefix (`/l` suffix)
    FourByteLE,
}

Suffix conventions:

/B - 1-byte length prefix (default if no suffix specified)
/H - 2-byte big-endian length prefix
/h - 2-byte little-endian length prefix
/L - 4-byte big-endian length prefix
/l - 4-byte little-endian length prefix
/J - Length includes the prefix width itself (combinable: /HJ, /lJ, etc.)

Examples:

use libmagic_rs::parser::ast::{TypeKind, PStringLengthWidth};

// 1-byte length prefix (default)
let pstring_default = TypeKind::PString {
    max_length: None,
    length_width: PStringLengthWidth::OneByte,
    length_includes_itself: false,
};

// 2-byte big-endian length prefix
let pstring_be = TypeKind::PString {
    max_length: None,
    length_width: PStringLengthWidth::TwoByteBE,
    length_includes_itself: false,
};

// 4-byte little-endian length prefix
let pstring_le = TypeKind::PString {
    max_length: None,
    length_width: PStringLengthWidth::FourByteLE,
    length_includes_itself: false,
};

// 2-byte big-endian with /J flag (JPEG-style self-inclusive length)
let pstring_jpeg = TypeKind::PString {
    max_length: None,
    length_width: PStringLengthWidth::TwoByteBE,
    length_includes_itself: true,
};

// Maximum 64-byte limit with 1-byte prefix
let limited_pstring = TypeKind::PString {
    max_length: Some(64),
    length_width: PStringLengthWidth::OneByte,
    length_includes_itself: false,
};

Regex (Regular Expression Pattern Matching)#

The Regex variant matches POSIX-extended regular expression patterns against file buffers. Patterns are binary-safe and always compiled with multi-line mode enabled (matching ^ and $ at line boundaries). The scan window is capped at 8192 bytes (FILE_REGEX_MAX) regardless of the count variant.

Structure:

Regex {
    flags: RegexFlags,
    count: RegexCount,
}

Fields:

flags: Modifier flags from the /c and /s suffixes (case-insensitive, start-offset). The /l suffix is NOT a flag — it selects the RegexCount::Lines variant below.
count: Scan window specifier, one of three variants (see RegexCount below).

Example:

0 regex [0-9]+ Numeric content
0 regex/1l ^#!/ Shebang on first line
0 regex/cs json Case-insensitive "json", anchor at match-start

Behavior:

Returns Value::String containing the matched text
Scan window capped at 8192 bytes (GNU file FILE_REGEX_MAX)
Multi-line mode unconditional (^/$ match line boundaries, . does not match newlines)
Zero-width matches (e.g., ^, a*) return Value::String("") and are distinguished from no-match
Only supports Equal and NotEqual operators; other comparison operators return TypeReadError::UnsupportedType

RegexFlags Struct#

The RegexFlags struct specifies regex behavior modifiers. All flags default to false via RegexFlags::default.

pub struct RegexFlags {
    /// `/c` - case-insensitive matching
    pub case_insensitive: bool,
    /// `/s` - advance anchor to match-start instead of match-end
    pub start_offset: bool,
}

The /l modifier is encoded by the RegexCount::Lines variant rather than a flag field, so "line count" and "byte count" are mutually exclusive at the type level.

RegexCount Enum#

The RegexCount enum specifies the scan window for a regex rule:

pub enum RegexCount {
    /// Plain `regex` with no suffix: full 8192-byte default window.
    Default,
    /// `regex/N`: scan at most `N` bytes, capped at 8192.
    Bytes(NonZeroU32),
    /// `regex/Nl` or `regex/l`: scan N line terminators (or the full
    /// capped window if `None`), capped at 8192 bytes.
    Lines(Option<NonZeroU32>),
}

Variant mapping:

regex → RegexCount::Default
regex/N → RegexCount::Bytes(N)
regex/Nl → RegexCount::Lines(Some(N))
regex/l → RegexCount::Lines(None) (behaviorally equivalent to Default: both walk the full 8192-byte capped window)

Examples:

use libmagic_rs::parser::ast::{RegexCount, RegexFlags, TypeKind};
use std::num::NonZeroU32;

// Plain regex with 8192-byte default scan window
let plain_regex = TypeKind::Regex {
    flags: RegexFlags::default(),
    count: RegexCount::Default,
};

// First line only (1 line, capped at 8192 bytes)
let first_line = TypeKind::Regex {
    flags: RegexFlags::default(),
    count: RegexCount::Lines(NonZeroU32::new(1)),
};

// Case-insensitive with anchor at match-start
let case_start = TypeKind::Regex {
    flags: RegexFlags {
        case_insensitive: true,
        start_offset: true,
    },
    count: RegexCount::Default,
};

Meta-types (Control Directives)#

The Meta variant represents pseudo-types that do not read bytes from the buffer. They encode control-flow directives inherited from the libmagic magic(5) format.

Structure:

/// Control-flow directives that do not read bytes from the buffer.
Meta(MetaType),

TypeKind::Meta(_) returns None from bit_width() because meta-types consume zero on-disk bytes.

MetaType Enum:

pub enum MetaType {
    /// `default` — fires only when no sibling at the same level has matched.
    Default,
    /// `clear` — resets the per-level sibling-matched flag.
    Clear,
    /// `name <id>` — defines a named subroutine (hoisted out of the rule list at load time).
    Name(String),
    /// `use <id>` — invokes a named subroutine at the resolved offset.
    Use(String),
    /// `indirect` — re-applies the full rule database at the resolved offset.
    Indirect,
    /// `offset` — emits the resolved file position as `Value::Uint` for
    /// printf-style format substitution (e.g. `%lld`).
    Offset,
}

Examples:

use libmagic_rs::parser::ast::{TypeKind, MetaType};

// Default fallback rule
let default_rule = TypeKind::Meta(MetaType::Default);

// Clear sibling-matched flag
let clear_rule = TypeKind::Meta(MetaType::Clear);

// Named subroutine declaration
let name_rule = TypeKind::Meta(MetaType::Name("part2".to_string()));

// Subroutine invocation
let use_rule = TypeKind::Meta(MetaType::Use("part2".to_string()));

// Re-entry into root rules
let indirect_rule = TypeKind::Meta(MetaType::Indirect);

// Report the resolved file offset for format substitution
let offset_rule = TypeKind::Meta(MetaType::Offset);

Parse-time Name Extraction:

Top-level name <id> rules are hoisted out of ParsedMagic::rules by parser::name_table::extract_name_table and placed into the name_table: NameTable field of ParsedMagic keyed by identifier. As a result, MetaType::Name variants in the final parsed rule list are expected only as an internal intermediate representation — name rules do not survive past the load boundary in normal operation.

Search (Bounded Literal Byte Sequence Search)#

The Search variant scans for a literal byte pattern within a bounded range. Unlike String, which matches only at the exact offset, Search scans forward up to range bytes for the first occurrence.

Structure:

Search {
    range: NonZeroUsize,
}

Fields:

range: Mandatory scan window width in bytes (must be non-zero per GNU file magic(5) specification)

Example:

0 search/256 PK\003\004 ZIP archive within first 256 bytes

Behavior:

Returns Value::String containing the matched bytes if found within range
Anchor advances to match_idx + pattern.len() (the byte position just past the matched needle), matching GNU file's softmagic.c FILE_SEARCH path where ms->search.offset += idx and then moffset() adds vlen = m->vallen. An earlier implementation incorrectly advanced by the full window size (range), but this caused relative-offset children to land far past the intended byte.
Only supports Equal and NotEqual operators
Range is mandatory; search/0 or bare search are parse errors

Examples:

use libmagic_rs::parser::ast::TypeKind;
use std::num::NonZeroUsize;

// Scan up to 256 bytes for the pattern
let bounded_search = TypeKind::Search {
    range: NonZeroUsize::new(256).unwrap(),
};

// Scan up to 1024 bytes
let wide_search = TypeKind::Search {
    range: NonZeroUsize::new(1024).unwrap(),
};

Endianness Options#

pub enum Endianness {
    Little, // Little-endian (x86, ARM in little mode)
    Big, // Big-endian (network byte order, PowerPC)
    Native, // Host system byte order
}

Operator Types#

The Operator enum defines comparison and bitwise operations:

pub enum Operator {
    Equal, // == (equality comparison)
    NotEqual, // != (inequality comparison)
    LessThan, // < (less-than comparison)
    GreaterThan, // > (greater-than comparison)
    LessEqual, // <= (less-than-or-equal comparison)
    GreaterEqual, // >= (greater-than-or-equal comparison)
    BitwiseAnd, // & (bitwise AND for pattern matching)
    BitwiseAndMask(u64), // & (bitwise AND with mask value)
}

Added in v0.2.0: The comparison operators LessThan, GreaterThan, LessEqual, and GreaterEqual were added. This is a breaking change for exhaustive matches on Operator.

Usage Examples:

// Exact match
let equal_op = Operator::Equal;

// Not equal
let not_equal_op = Operator::NotEqual;

// Less than comparison
let less_op = Operator::LessThan;

// Greater than comparison
let greater_op = Operator::GreaterThan;

// Less than or equal
let less_equal_op = Operator::LessEqual;

// Greater than or equal
let greater_equal_op = Operator::GreaterEqual;

// Bitwise AND (useful for flag checking)
let bitwise_op = Operator::BitwiseAnd;

// Bitwise AND with mask
let bitwise_mask_op = Operator::BitwiseAndMask(0xFF00);

Value Types#

The Value enum represents expected values for comparison:

pub enum Value {
    Uint(u64), // Unsigned integer
    Int(i64), // Signed integer
    Bytes(Vec<u8>), // Byte sequence
    String(String), // String value
}

Examples:

// Unsigned integer value
let uint_val = Value::Uint(0x464c457f);

// Signed integer value
let int_val = Value::Int(-1);

// Byte sequence (magic numbers)
let bytes_val = Value::Bytes(vec![0x50, 0x4b, 0x03, 0x04]); // ZIP signature

// String value
let string_val = Value::String("#!/bin/sh".to_string());

Serialization Support#

All AST types implement Serialize and Deserialize for caching and interchange with comprehensive test coverage:

use serde_json;

// Serialize a rule to JSON (fully tested)
let rule = MagicRule { /* ... */ };
let json = serde_json::to_string(&rule)?;

// Deserialize from JSON (fully tested)
let rule: MagicRule = serde_json::from_str(&json)?;

// All edge cases are tested including:
// - Empty collections (Vec::new(), String::new())
// - Extreme values (u64::MAX, i64::MIN, i64::MAX)
// - Complex nested structures with multiple levels
// - All enum variants and their serialization round-trips

Implementation Status:

✅ Complete serialization for all AST types
✅ Comprehensive testing with edge cases and boundary values
✅ JSON compatibility for rule caching and interchange
✅ Round-trip validation ensuring data integrity

Common Patterns#

ELF File Detection#

let elf_rules = vec![
    MagicRule {
        offset: OffsetSpec::Absolute(0),
        typ: TypeKind::Long { endian: Endianness::Little, signed: false },
        op: Operator::Equal,
        value: Value::Bytes(vec![0x7f, 0x45, 0x4c, 0x46]),
        message: "ELF".to_string(),
        children: vec![
            MagicRule {
                offset: OffsetSpec::Absolute(4),
                typ: TypeKind::Byte { signed: false },
                op: Operator::Equal,
                value: Value::Uint(1),
                message: "32-bit".to_string(),
                children: vec![],
                level: 1,
            },
            MagicRule {
                offset: OffsetSpec::Absolute(4),
                typ: TypeKind::Byte { signed: false },
                op: Operator::Equal,
                value: Value::Uint(2),
                message: "64-bit".to_string(),
                children: vec![],
                level: 1,
            },
        ],
        level: 0,
    }
];

ZIP Archive Detection#

let zip_rule = MagicRule {
    offset: OffsetSpec::Absolute(0),
    typ: TypeKind::Long { endian: Endianness::Little, signed: false },
    op: Operator::Equal,
    value: Value::Bytes(vec![0x50, 0x4b, 0x03, 0x04]),
    message: "ZIP archive".to_string(),
    children: vec![],
    level: 0,
};

Script Detection with String Matching#

let script_rule = MagicRule {
    offset: OffsetSpec::Absolute(0),
    typ: TypeKind::String { max_length: Some(32) },
    op: Operator::Equal,
    value: Value::String("#!/bin/bash".to_string()),
    message: "Bash script".to_string(),
    children: vec![],
    level: 0,
};

Best Practices#

Rule Organization#

Start with broad patterns and use child rules for specifics
Order rules by probability of matching (most common first)
Use appropriate types for the data being checked
Minimize indirection for performance

Type Selection#

Use Byte { signed } for single-byte values and flags, specifying signedness
Use Short/Long/Quad with explicit endianness and signedness for multi-byte integers
Use String with length limits for text patterns at exact offsets
Use PString for Pascal-style length-prefixed strings
Use Regex for pattern matching (complex patterns, line-based checks, case-insensitive matching)
Use Search for simple substring matching within a bounded range (faster than regex for literal patterns)
Use Bytes for exact binary sequences

Performance Considerations#

Prefer absolute offsets over indirect when possible
Use bitwise AND for flag checking instead of multiple equality rules
Limit string lengths to prevent excessive reading
Structure hierarchies to fail fast on non-matches

The AST provides a flexible, type-safe foundation for representing magic rules while maintaining compatibility with existing magic file formats.