dfajit 0.1.0

JIT compilation of DFA transition tables to native x86_64 jump tables
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121

//! Match result types for DFA scanning.
//!
//! This module provides the [`Match`] struct, a lightweight `repr(C)` type
//! used as the common currency for match results throughout `dfajit`.
//! It is designed to be GPU-compatible (16 bytes, 4×`u32`) and can be
//! written directly into output buffers by the JIT-compiled scanner.

/// A match result from pattern scanning.
///
/// Uses `u32` offsets for GPU buffer compatibility. For inputs larger
/// than 4 GiB, scan in chunks and add the chunk base offset.
///
/// Uses `repr(C)` for GPU buffer compatibility. 16 bytes per match.
///
/// # Example
///
/// ```rust
/// use dfajit::Match;
///
/// let m = Match::from_parts(0, 5, 10);
/// assert_eq!(m.len(), 5);
/// assert!(m.contains(&Match::from_parts(0, 6, 8)));
/// ```
#[repr(C)]
#[derive(Debug, Clone, Copy, Default)]
pub struct Match {
    /// Index of the pattern that matched (0-based, in insertion order).
    pub pattern_id: u32,
    /// Byte offset where the match starts (inclusive).
    pub start: u32,
    /// Byte offset where the match ends (exclusive).
    pub end: u32,
    /// Padding for GPU alignment. Ignored in equality comparisons.
    pub padding: u32,
}

impl PartialEq for Match {
    fn eq(&self, other: &Self) -> bool {
        self.pattern_id == other.pattern_id
            && self.start == other.start
            && self.end == other.end
    }
}

impl Eq for Match {}

impl PartialOrd for Match {
    fn partial_cmp(&self, other: &Self) -> Option<std::cmp::Ordering> {
        Some(self.cmp(other))
    }
}

impl Ord for Match {
    fn cmp(&self, other: &Self) -> std::cmp::Ordering {
        self.start
            .cmp(&other.start)
            .then(self.pattern_id.cmp(&other.pattern_id))
            .then(self.end.cmp(&other.end))
    }
}

impl Match {
    /// Construct a match from its public fields.
    #[must_use]
    pub const fn from_parts(pattern_id: u32, start: u32, end: u32) -> Self {
        Self {
            pattern_id,
            start,
            end,
            padding: 0,
        }
    }

    /// Returns the padding field (reserved for future use / GPU flags).
    #[must_use]
    pub const fn padding(&self) -> u32 {
        self.padding
    }

    /// Create a match with explicit padding value.
    ///
    /// Used internally by GPU backends that pack flags into the padding field.
    #[must_use]
    pub const fn from_parts_with_padding(
        pattern_id: u32,
        start: u32,
        end: u32,
        padding: u32,
    ) -> Self {
        Self {
            pattern_id,
            start,
            end,
            padding,
        }
    }

    /// Returns `true` if this match's byte range fully contains `other`.
    #[must_use]
    pub const fn contains(&self, other: &Match) -> bool {
        self.start <= other.start && self.end >= other.end
    }

    /// Returns `true` if this match's byte range overlaps with `other`.
    #[must_use]
    pub const fn overlaps(&self, other: &Match) -> bool {
        self.start < other.end && other.start < self.end
    }

    /// Byte length of the matched region.
    #[must_use]
    pub const fn len(&self) -> u32 {
        self.end.saturating_sub(self.start)
    }

    /// Returns `true` if the match has zero length.
    #[must_use]
    pub const fn is_empty(&self) -> bool {
        self.start == self.end
    }
}