arity 0.3.0

An LSP, formatter, and linter for R
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
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186

//! Byte-offset → (line, column) conversion.
//!
//! Two coordinate systems share the same line-start table:
//! - 1-indexed (line, column) in **code points** for CLI diagnostics.
//! - 0-indexed (line, character) in **UTF-16 units** for LSP positions.

use lsp_types::Position;

#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub struct LineCol {
    /// 1-indexed line number.
    pub line: usize,
    /// 1-indexed column in code points (not bytes, not UTF-16 units).
    pub column: usize,
}

/// Precomputed line-start byte offsets for a text buffer.
///
/// `line_starts[i]` is the byte offset of the first character of line `i`
/// (0-indexed). `line_starts` always starts with `0`.
#[derive(Debug, Clone)]
pub struct LineIndex<'a> {
    text: &'a str,
    line_starts: Vec<usize>,
}

impl<'a> LineIndex<'a> {
    pub fn new(text: &'a str) -> Self {
        let mut line_starts = Vec::with_capacity(text.len() / 40 + 1);
        line_starts.push(0);
        for (offset, byte) in text.bytes().enumerate() {
            if byte == b'\n' {
                line_starts.push(offset + 1);
            }
        }
        Self { text, line_starts }
    }

    /// 1-indexed (line, column-in-code-points). Suitable for CLI diagnostics.
    pub fn byte_to_lc(&self, offset: usize) -> LineCol {
        let clamped = offset.min(self.text.len());
        let line_idx = self.line_index_for(clamped);
        let line_start = self.line_starts[line_idx];
        let column = self.text[line_start..clamped].chars().count() + 1;
        LineCol {
            line: line_idx + 1,
            column,
        }
    }

    /// 0-indexed LSP `Position` (UTF-16 character offsets).
    pub fn byte_to_position(&self, offset: usize) -> Position {
        let clamped = offset.min(self.text.len());
        let line_idx = self.line_index_for(clamped);
        let line_start = self.line_starts[line_idx];
        let character = self.text[line_start..clamped].encode_utf16().count() as u32;
        Position::new(line_idx as u32, character)
    }

    /// Inverse of [`byte_to_position`](Self::byte_to_position): a 0-indexed LSP
    /// `Position` (UTF-16 character offset) back to a byte offset. A line or
    /// character past the end clamps to the end of the line / buffer.
    pub fn position_to_byte(&self, position: Position) -> usize {
        let line = position.line as usize;
        let Some(&line_start) = self.line_starts.get(line) else {
            return self.text.len();
        };
        let line_end = self
            .line_starts
            .get(line + 1)
            .copied()
            .unwrap_or(self.text.len());
        let line_text = &self.text[line_start..line_end];
        let mut utf16 = 0u32;
        for (byte_off, ch) in line_text.char_indices() {
            if utf16 >= position.character {
                return line_start + byte_off;
            }
            utf16 += ch.len_utf16() as u32;
        }
        line_end
    }

    /// Total line count (1 even for empty text).
    pub fn line_count(&self) -> usize {
        self.line_starts.len()
    }

    fn line_index_for(&self, offset: usize) -> usize {
        match self.line_starts.binary_search(&offset) {
            Ok(idx) => idx,
            Err(idx) => idx.saturating_sub(1),
        }
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn empty_string() {
        let idx = LineIndex::new("");
        assert_eq!(idx.byte_to_lc(0), LineCol { line: 1, column: 1 });
        assert_eq!(idx.byte_to_position(0), Position::new(0, 0));
    }

    #[test]
    fn single_line() {
        let idx = LineIndex::new("abc");
        assert_eq!(idx.byte_to_lc(0).column, 1);
        assert_eq!(idx.byte_to_lc(2).column, 3);
        assert_eq!(idx.byte_to_lc(3).column, 4);
        assert_eq!(idx.byte_to_position(2), Position::new(0, 2));
    }

    #[test]
    fn multi_line() {
        let idx = LineIndex::new("ab\ncd\nef");
        assert_eq!(idx.byte_to_lc(0), LineCol { line: 1, column: 1 });
        assert_eq!(idx.byte_to_lc(2), LineCol { line: 1, column: 3 }); // the newline
        assert_eq!(idx.byte_to_lc(3), LineCol { line: 2, column: 1 });
        assert_eq!(idx.byte_to_lc(6), LineCol { line: 3, column: 1 });
        assert_eq!(idx.byte_to_position(6), Position::new(2, 0));
    }

    #[test]
    fn utf8_multibyte() {
        // Each non-ASCII char below is 2 bytes in UTF-8, 1 UTF-16 unit.
        let idx = LineIndex::new("\u{00e1}b\nc");
        // offset 2 = after á (2 bytes), column should be 2 (code points)
        assert_eq!(idx.byte_to_lc(2), LineCol { line: 1, column: 2 });
        assert_eq!(idx.byte_to_position(2), Position::new(0, 1));
        // offset 3 = at 'b'
        assert_eq!(idx.byte_to_lc(3), LineCol { line: 1, column: 3 });
        assert_eq!(idx.byte_to_position(3), Position::new(0, 2));
    }

    #[test]
    fn utf16_surrogate_pair() {
        // U+1F600 (emoji) is 4 bytes in UTF-8, 2 UTF-16 units (surrogate pair).
        let idx = LineIndex::new("\u{1F600}x");
        // offset 4 = after emoji
        assert_eq!(idx.byte_to_lc(4), LineCol { line: 1, column: 2 });
        assert_eq!(idx.byte_to_position(4), Position::new(0, 2));
    }

    #[test]
    fn offset_past_end_clamps() {
        let idx = LineIndex::new("abc");
        assert_eq!(idx.byte_to_lc(100), LineCol { line: 1, column: 4 });
    }

    #[test]
    fn trailing_newline() {
        let idx = LineIndex::new("ab\n");
        assert_eq!(idx.byte_to_lc(3), LineCol { line: 2, column: 1 });
    }

    #[test]
    fn position_to_byte_round_trips() {
        let text = "ab\ncde\nf";
        let idx = LineIndex::new(text);
        for offset in 0..=text.len() {
            // Only byte offsets on char boundaries round-trip exactly.
            if !text.is_char_boundary(offset) {
                continue;
            }
            let pos = idx.byte_to_position(offset);
            assert_eq!(idx.position_to_byte(pos), offset, "offset {offset}");
        }
    }

    #[test]
    fn position_to_byte_handles_utf16_and_overshoot() {
        // Emoji is 4 UTF-8 bytes, 2 UTF-16 units.
        let idx = LineIndex::new("\u{1F600}x\ny");
        assert_eq!(idx.position_to_byte(Position::new(0, 0)), 0);
        assert_eq!(idx.position_to_byte(Position::new(0, 2)), 4); // at 'x'
        assert_eq!(idx.position_to_byte(Position::new(1, 0)), 6); // at 'y'
        // A character past the line end clamps to the line end.
        assert_eq!(idx.position_to_byte(Position::new(0, 99)), 6);
        // A line past the end clamps to the buffer end.
        assert_eq!(idx.position_to_byte(Position::new(9, 0)), 7);
    }
}