summaryrefslogtreecommitdiff
path: root/helix-core/src/graphemes.rs
blob: c639887576288474d3a15bc020d091f9bb37b5d0 (plain) (blame)
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
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
//! Utility functions to traverse the unicode graphemes of a `Rope`'s text contents.
//!
//! Based on <https://github.com/cessen/led/blob/c4fa72405f510b7fd16052f90a598c429b3104a6/src/graphemes.rs>
use ropey::{iter::Chunks, str_utils::byte_to_char_idx, RopeSlice};
use unicode_segmentation::{GraphemeCursor, GraphemeIncomplete};
use unicode_width::UnicodeWidthStr;

use std::fmt;

#[must_use]
pub fn grapheme_width(g: &str) -> usize {
    if g.as_bytes()[0] <= 127 {
        // Fast-path ascii.
        // Point 1: theoretically, ascii control characters should have zero
        // width, but in our case we actually want them to have width: if they
        // show up in text, we want to treat them as textual elements that can
        // be editied.  So we can get away with making all ascii single width
        // here.
        // Point 2: we're only examining the first codepoint here, which means
        // we're ignoring graphemes formed with combining characters.  However,
        // if it starts with ascii, it's going to be a single-width grapeheme
        // regardless, so, again, we can get away with that here.
        // Point 3: we're only examining the first _byte_.  But for utf8, when
        // checking for ascii range values only, that works.
        1
    } else {
        // We use max(1) here because all grapeheme clusters--even illformed
        // ones--should have at least some width so they can be edited
        // properly.
        UnicodeWidthStr::width(g).max(1)
    }
}

#[must_use]
pub fn nth_prev_grapheme_boundary(slice: RopeSlice, char_idx: usize, n: usize) -> usize {
    // Bounds check
    debug_assert!(char_idx <= slice.len_chars());

    // We work with bytes for this, so convert.
    let mut byte_idx = slice.char_to_byte(char_idx);

    // Get the chunk with our byte index in it.
    let (mut chunk, mut chunk_byte_idx, mut chunk_char_idx, _) = slice.chunk_at_byte(byte_idx);

    // Set up the grapheme cursor.
    let mut gc = GraphemeCursor::new(byte_idx, slice.len_bytes(), true);

    // Find the previous grapheme cluster boundary.
    for _ in 0..n {
        loop {
            match gc.prev_boundary(chunk, chunk_byte_idx) {
                Ok(None) => return 0,
                Ok(Some(n)) => {
                    byte_idx = n;
                    break;
                }
                Err(GraphemeIncomplete::PrevChunk) => {
                    let (a, b, c, _) = slice.chunk_at_byte(chunk_byte_idx - 1);
                    chunk = a;
                    chunk_byte_idx = b;
                    chunk_char_idx = c;
                }
                Err(GraphemeIncomplete::PreContext(n)) => {
                    let ctx_chunk = slice.chunk_at_byte(n - 1).0;
                    gc.provide_context(ctx_chunk, n - ctx_chunk.len());
                }
                _ => unreachable!(),
            }
        }
    }
    let tmp = byte_to_char_idx(chunk, byte_idx - chunk_byte_idx);
    chunk_char_idx + tmp
}

/// Finds the previous grapheme boundary before the given char position.
#[must_use]
#[inline(always)]
pub fn prev_grapheme_boundary(slice: RopeSlice, char_idx: usize) -> usize {
    nth_prev_grapheme_boundary(slice, char_idx, 1)
}

#[must_use]
pub fn nth_next_grapheme_boundary(slice: RopeSlice, char_idx: usize, n: usize) -> usize {
    // Bounds check
    debug_assert!(char_idx <= slice.len_chars());

    // We work with bytes for this, so convert.
    let mut byte_idx = slice.char_to_byte(char_idx);

    // Get the chunk with our byte index in it.
    let (mut chunk, mut chunk_byte_idx, mut chunk_char_idx, _) = slice.chunk_at_byte(byte_idx);

    // Set up the grapheme cursor.
    let mut gc = GraphemeCursor::new(byte_idx, slice.len_bytes(), true);

    // Find the nth next grapheme cluster boundary.
    for _ in 0..n {
        loop {
            match gc.next_boundary(chunk, chunk_byte_idx) {
                Ok(None) => return slice.len_chars(),
                Ok(Some(n)) => {
                    byte_idx = n;
                    break;
                }
                Err(GraphemeIncomplete::NextChunk) => {
                    chunk_byte_idx += chunk.len();
                    let (a, _, c, _) = slice.chunk_at_byte(chunk_byte_idx);
                    chunk = a;
                    chunk_char_idx = c;
                }
                Err(GraphemeIncomplete::PreContext(n)) => {
                    let ctx_chunk = slice.chunk_at_byte(n - 1).0;
                    gc.provide_context(ctx_chunk, n - ctx_chunk.len());
                }
                _ => unreachable!(),
            }
        }
    }
    let tmp = byte_to_char_idx(chunk, byte_idx - chunk_byte_idx);
    chunk_char_idx + tmp
}

/// Finds the next grapheme boundary after the given char position.
#[must_use]
#[inline(always)]
pub fn next_grapheme_boundary(slice: RopeSlice, char_idx: usize) -> usize {
    nth_next_grapheme_boundary(slice, char_idx, 1)
}

/// Returns the passed char index if it's already a grapheme boundary,
/// or the next grapheme boundary char index if not.
#[must_use]
#[inline]
pub fn ensure_grapheme_boundary_next(slice: RopeSlice, char_idx: usize) -> usize {
    if char_idx == 0 {
        char_idx
    } else {
        next_grapheme_boundary(slice, char_idx - 1)
    }
}

/// Returns the passed char index if it's already a grapheme boundary,
/// or the prev grapheme boundary char index if not.
#[must_use]
#[inline]
pub fn ensure_grapheme_boundary_prev(slice: RopeSlice, char_idx: usize) -> usize {
    if char_idx == slice.len_chars() {
        char_idx
    } else {
        prev_grapheme_boundary(slice, char_idx + 1)
    }
}

/// Returns whether the given char position is a grapheme boundary.
#[must_use]
pub fn is_grapheme_boundary(slice: RopeSlice, char_idx: usize) -> bool {
    // Bounds check
    debug_assert!(char_idx <= slice.len_chars());

    // We work with bytes for this, so convert.
    let byte_idx = slice.char_to_byte(char_idx);

    // Get the chunk with our byte index in it.
    let (chunk, chunk_byte_idx, _, _) = slice.chunk_at_byte(byte_idx);

    // Set up the grapheme cursor.
    let mut gc = GraphemeCursor::new(byte_idx, slice.len_bytes(), true);

    // Determine if the given position is a grapheme cluster boundary.
    loop {
        match gc.is_boundary(chunk, chunk_byte_idx) {
            Ok(n) => return n,
            Err(GraphemeIncomplete::PreContext(n)) => {
                let (ctx_chunk, ctx_byte_start, _, _) = slice.chunk_at_byte(n - 1);
                gc.provide_context(ctx_chunk, ctx_byte_start);
            }
            Err(_) => unreachable!(),
        }
    }
}

/// An iterator over the graphemes of a `RopeSlice`.
#[derive(Clone)]
pub struct RopeGraphemes<'a> {
    text: RopeSlice<'a>,
    chunks: Chunks<'a>,
    cur_chunk: &'a str,
    cur_chunk_start: usize,
    cursor: GraphemeCursor,
}

impl<'a> fmt::Debug for RopeGraphemes<'a> {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        f.debug_struct("RopeGraphemes")
            .field("text", &self.text)
            .field("chunks", &self.chunks)
            .field("cur_chunk", &self.cur_chunk)
            .field("cur_chunk_start", &self.cur_chunk_start)
            // .field("cursor", &self.cursor)
            .finish()
    }
}

impl<'a> RopeGraphemes<'a> {
    #[must_use]
    pub fn new(slice: RopeSlice) -> RopeGraphemes {
        let mut chunks = slice.chunks();
        let first_chunk = chunks.next().unwrap_or("");
        RopeGraphemes {
            text: slice,
            chunks,
            cur_chunk: first_chunk,
            cur_chunk_start: 0,
            cursor: GraphemeCursor::new(0, slice.len_bytes(), true),
        }
    }
}

impl<'a> Iterator for RopeGraphemes<'a> {
    type Item = RopeSlice<'a>;

    fn next(&mut self) -> Option<RopeSlice<'a>> {
        let a = self.cursor.cur_cursor();
        let b;
        loop {
            match self
                .cursor
                .next_boundary(self.cur_chunk, self.cur_chunk_start)
            {
                Ok(None) => {
                    return None;
                }
                Ok(Some(n)) => {
                    b = n;
                    break;
                }
                Err(GraphemeIncomplete::NextChunk) => {
                    self.cur_chunk_start += self.cur_chunk.len();
                    self.cur_chunk = self.chunks.next().unwrap_or("");
                }
                Err(GraphemeIncomplete::PreContext(idx)) => {
                    let (chunk, byte_idx, _, _) = self.text.chunk_at_byte(idx.saturating_sub(1));
                    self.cursor.provide_context(chunk, byte_idx);
                }
                _ => unreachable!(),
            }
        }

        if a < self.cur_chunk_start {
            let a_char = self.text.byte_to_char(a);
            let b_char = self.text.byte_to_char(b);

            Some(self.text.slice(a_char..b_char))
        } else {
            let a2 = a - self.cur_chunk_start;
            let b2 = b - self.cur_chunk_start;
            Some((&self.cur_chunk[a2..b2]).into())
        }
    }
}