1 //! `LineIndex` maps flat `TextSize` offsets into `(Line, Column)`
5 use rustc_hash::FxHashMap;
6 use syntax::{TextRange, TextSize};
8 #[derive(Clone, Debug, PartialEq, Eq)]
10 /// Offset the the beginning of each line, zero-based
11 pub(crate) newlines: Vec<TextSize>,
12 /// List of non-ASCII characters on each line
13 pub(crate) utf16_lines: FxHashMap<u32, Vec<Utf16Char>>,
16 #[derive(Clone, Copy, Debug, PartialEq, Eq, Hash)]
17 pub struct LineColUtf16 {
24 #[derive(Clone, Copy, Debug, PartialEq, Eq, Hash)]
28 /// Zero-based utf8 offset
32 #[derive(Clone, Debug, Hash, PartialEq, Eq)]
33 pub(crate) struct Utf16Char {
34 /// Start offset of a character inside a line, zero-based
35 pub(crate) start: TextSize,
36 /// End offset of a character inside a line, zero-based
37 pub(crate) end: TextSize,
41 /// Returns the length in 8-bit UTF-8 code units.
42 fn len(&self) -> TextSize {
46 /// Returns the length in 16-bit UTF-16 code units.
47 fn len_utf16(&self) -> usize {
48 if self.len() == TextSize::from(4) {
57 pub fn new(text: &str) -> LineIndex {
58 let mut utf16_lines = FxHashMap::default();
59 let mut utf16_chars = Vec::new();
61 let mut newlines = vec![0.into()];
62 let mut curr_row = 0.into();
63 let mut curr_col = 0.into();
65 for c in text.chars() {
66 let c_len = TextSize::of(c);
69 newlines.push(curr_row);
71 // Save any utf-16 characters seen in the previous line
72 if !utf16_chars.is_empty() {
73 utf16_lines.insert(line, utf16_chars);
74 utf16_chars = Vec::new();
77 // Prepare for processing the next line
84 utf16_chars.push(Utf16Char { start: curr_col, end: curr_col + c_len });
90 // Save any utf-16 characters seen in the last line
91 if !utf16_chars.is_empty() {
92 utf16_lines.insert(line, utf16_chars);
95 LineIndex { newlines, utf16_lines }
98 pub fn line_col(&self, offset: TextSize) -> LineCol {
99 let line = self.newlines.partition_point(|&it| it <= offset) - 1;
100 let line_start_offset = self.newlines[line];
101 let col = offset - line_start_offset;
102 LineCol { line: line as u32, col: col.into() }
105 pub fn offset(&self, line_col: LineCol) -> TextSize {
106 self.newlines[line_col.line as usize] + TextSize::from(line_col.col)
109 pub fn to_utf16(&self, line_col: LineCol) -> LineColUtf16 {
110 let col = self.utf8_to_utf16_col(line_col.line, line_col.col.into());
111 LineColUtf16 { line: line_col.line, col: col as u32 }
114 pub fn to_utf8(&self, line_col: LineColUtf16) -> LineCol {
115 let col = self.utf16_to_utf8_col(line_col.line, line_col.col);
116 LineCol { line: line_col.line, col: col.into() }
119 pub fn lines(&self, range: TextRange) -> impl Iterator<Item = TextRange> + '_ {
120 let lo = self.newlines.partition_point(|&it| it < range.start());
121 let hi = self.newlines.partition_point(|&it| it <= range.end());
122 let all = iter::once(range.start())
123 .chain(self.newlines[lo..hi].iter().copied())
124 .chain(iter::once(range.end()));
128 .map(|(lo, hi)| TextRange::new(lo, hi))
129 .filter(|it| !it.is_empty())
132 fn utf8_to_utf16_col(&self, line: u32, col: TextSize) -> usize {
133 let mut res: usize = col.into();
134 if let Some(utf16_chars) = self.utf16_lines.get(&line) {
135 for c in utf16_chars {
137 res -= usize::from(c.len()) - c.len_utf16();
139 // From here on, all utf16 characters come *after* the character we are mapping,
140 // so we don't need to take them into account
148 fn utf16_to_utf8_col(&self, line: u32, mut col: u32) -> TextSize {
149 if let Some(utf16_chars) = self.utf16_lines.get(&line) {
150 for c in utf16_chars {
151 if col > u32::from(c.start) {
152 col += u32::from(c.len()) - c.len_utf16() as u32;
154 // From here on, all utf16 characters come *after* the character we are mapping,
155 // so we don't need to take them into account