aboutsummaryrefslogtreecommitdiff
path: root/helix-core/src/selection.rs
blob: e452c2e28efa958b35cddf5c189f4112b74283dc (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
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
//! Selections are the primary editing construct. Even a single cursor is defined as an empty
//! single selection range.
//!
//! All positioning is done via `char` offsets into the buffer.
use crate::{Assoc, ChangeSet, Rope, RopeSlice};
use smallvec::{smallvec, SmallVec};
use std::borrow::Cow;

#[inline]
fn abs_difference(x: usize, y: usize) -> usize {
    if x < y {
        y - x
    } else {
        x - y
    }
}

/// A single selection range. Anchor-inclusive, head-exclusive.
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub struct Range {
    // TODO: optimize into u32
    /// The anchor of the range: the side that doesn't move when extending.
    pub anchor: usize,
    /// The head of the range, moved when extending.
    pub head: usize,
    pub horiz: Option<u32>,
} // TODO: might be cheaper to store normalized as from/to and an inverted flag

impl Range {
    pub fn new(anchor: usize, head: usize) -> Self {
        Self {
            anchor,
            head,
            horiz: None,
        }
    }

    pub fn point(head: usize) -> Self {
        Self::new(head, head)
    }

    /// Start of the range.
    #[inline]
    #[must_use]
    pub fn from(&self) -> usize {
        std::cmp::min(self.anchor, self.head)
    }

    /// End of the range.
    #[inline]
    #[must_use]
    pub fn to(&self) -> usize {
        std::cmp::max(self.anchor, self.head)
    }

    /// `true` when head and anchor are at the same position.
    #[inline]
    pub fn is_empty(&self) -> bool {
        self.anchor == self.head
    }

    /// Check two ranges for overlap.
    #[must_use]
    pub fn overlaps(&self, other: &Self) -> bool {
        // cursor overlap is checked differently
        if self.is_empty() {
            let pos = self.head;
            pos >= other.from() && other.to() >= pos
        } else {
            self.to() > other.from() && other.to() > self.from()
        }
    }

    pub fn contains(&self, pos: usize) -> bool {
        if self.is_empty() {
            return false;
        }

        if self.anchor < self.head {
            self.anchor <= pos && pos < self.head
        } else {
            self.head < pos && pos <= self.anchor
        }
    }

    /// Map a range through a set of changes. Returns a new range representing the same position
    /// after the changes are applied.
    pub fn map(self, changes: &ChangeSet) -> Self {
        let anchor = changes.map_pos(self.anchor, Assoc::After);
        let head = changes.map_pos(self.head, Assoc::After);

        // TODO: possibly unnecessary
        if self.anchor == anchor && self.head == head {
            return self;
        }
        Self {
            anchor,
            head,
            horiz: None,
        }
    }

    /// Extend the range to cover at least `from` `to`.
    #[must_use]
    pub fn extend(&self, from: usize, to: usize) -> Self {
        if from <= self.anchor && to >= self.anchor {
            return Self {
                anchor: from,
                head: to,
                horiz: None,
            };
        }

        Self {
            anchor: self.anchor,
            head: if abs_difference(from, self.anchor) > abs_difference(to, self.anchor) {
                from
            } else {
                to
            },
            horiz: None,
        }
    }

    // groupAt

    #[inline]
    pub fn fragment<'a, 'b: 'a>(&'a self, text: RopeSlice<'b>) -> Cow<'b, str> {
        Cow::from(text.slice(self.from()..self.to() + 1))
    }
}

/// A selection consists of one or more selection ranges.
/// invariant: A selection can never be empty (always contains at least primary range).
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct Selection {
    ranges: SmallVec<[Range; 1]>,
    primary_index: usize,
}

#[allow(clippy::len_without_is_empty)] // a Selection is never empty
impl Selection {
    // eq

    #[must_use]
    pub fn primary(&self) -> Range {
        self.ranges[self.primary_index]
    }

    #[must_use]
    pub fn cursor(&self) -> usize {
        self.primary().head
    }

    /// Ensure selection containing only the primary selection.
    pub fn into_single(self) -> Self {
        if self.ranges.len() == 1 {
            self
        } else {
            Self {
                ranges: smallvec![self.ranges[self.primary_index]],
                primary_index: 0,
            }
        }
    }

    pub fn push(mut self, range: Range) -> Self {
        let index = self.ranges.len();
        self.ranges.push(range);

        Self::normalize(self.ranges, index)
    }
    // replace_range

    /// Map selections over a set of changes. Useful for adjusting the selection position after
    /// applying changes to a document.
    pub fn map(self, changes: &ChangeSet) -> Self {
        if changes.is_empty() {
            return self;
        }

        Self::new(
            self.ranges
                .into_iter()
                .map(|range| range.map(changes))
                .collect(),
            self.primary_index,
        )
    }

    pub fn ranges(&self) -> &[Range] {
        &self.ranges
    }

    pub fn primary_index(&self) -> usize {
        self.primary_index
    }

    #[must_use]
    /// Constructs a selection holding a single range.
    pub fn single(anchor: usize, head: usize) -> Self {
        Self {
            ranges: smallvec![Range {
                anchor,
                head,
                horiz: None
            }],
            primary_index: 0,
        }
    }

    /// Constructs a selection holding a single cursor.
    pub fn point(pos: usize) -> Self {
        Self::single(pos, pos)
    }

    fn normalize(mut ranges: SmallVec<[Range; 1]>, mut primary_index: usize) -> Self {
        let primary = ranges[primary_index];
        ranges.sort_unstable_by_key(Range::from);
        primary_index = ranges.iter().position(|&range| range == primary).unwrap();

        let mut result = SmallVec::with_capacity(ranges.len()); // approx

        // TODO: we could do with one vec by removing elements as we mutate

        for (i, range) in ranges.into_iter().enumerate() {
            // if previous value exists
            if let Some(prev) = result.last_mut() {
                // and we overlap it

                // TODO: we used to simply check range.from() <(=) prev.to()
                // avoiding two comparisons
                if range.overlaps(prev) {
                    let from = prev.from();
                    let to = std::cmp::max(range.to(), prev.to());

                    if i <= primary_index {
                        primary_index -= 1
                    }

                    // merge into previous
                    if range.anchor > range.head {
                        prev.anchor = to;
                        prev.head = from;
                    } else {
                        prev.anchor = from;
                        prev.head = to;
                    }
                    continue;
                }
            }

            result.push(range)
        }

        Self {
            ranges: result,
            primary_index,
        }
    }

    // TODO: consume an iterator or a vec to reduce allocations?
    #[must_use]
    pub fn new(ranges: SmallVec<[Range; 1]>, primary_index: usize) -> Self {
        assert!(!ranges.is_empty());

        // fast path for a single selection (cursor)
        if ranges.len() == 1 {
            return Self {
                ranges,
                primary_index: 0,
            };
        }

        // TODO: only normalize if needed (any ranges out of order)
        Self::normalize(ranges, primary_index)
    }

    /// Takes a closure and maps each selection over the closure.
    pub fn transform<F>(&self, f: F) -> Self
    where
        F: Fn(Range) -> Range,
    {
        Self::new(
            self.ranges.iter().copied().map(f).collect(),
            self.primary_index,
        )
    }

    pub fn fragments<'a>(&'a self, text: RopeSlice<'a>) -> impl Iterator<Item = Cow<str>> + 'a {
        self.ranges.iter().map(move |range| range.fragment(text))
    }

    #[inline(always)]
    pub fn iter(&self) -> std::slice::Iter<'_, Range> {
        self.ranges.iter()
    }

    #[inline(always)]
    pub fn len(&self) -> usize {
        self.ranges.len()
    }
}

impl<'a> IntoIterator for &'a Selection {
    type Item = &'a Range;
    type IntoIter = std::slice::Iter<'a, Range>;

    fn into_iter(self) -> std::slice::Iter<'a, Range> {
        self.ranges().iter()
    }
}

// TODO: checkSelection -> check if valid for doc length && sorted

pub fn keep_matches(
    text: RopeSlice,
    selection: &Selection,
    regex: &crate::regex::Regex,
) -> Option<Selection> {
    let result: SmallVec<_> = selection
        .iter()
        .filter(|range| regex.is_match(&range.fragment(text)))
        .copied()
        .collect();

    // TODO: figure out a new primary index
    if !result.is_empty() {
        return Some(Selection::new(result, 0));
    }
    None
}

pub fn select_on_matches(
    text: RopeSlice,
    selection: &Selection,
    regex: &crate::regex::Regex,
) -> Option<Selection> {
    let mut result = SmallVec::with_capacity(selection.len());

    for sel in selection {
        // TODO: can't avoid occasional allocations since Regex can't operate on chunks yet
        let fragment = sel.fragment(text);

        let mut sel_start = sel.from();
        let sel_end = sel.to();

        let mut start_byte = text.char_to_byte(sel_start);

        for mat in regex.find_iter(&fragment) {
            // TODO: retain range direction

            let start = text.byte_to_char(start_byte + mat.start());
            let end = text.byte_to_char(start_byte + mat.end());
            result.push(Range::new(start, end - 1));
        }
    }

    // TODO: figure out a new primary index
    if !result.is_empty() {
        return Some(Selection::new(result, 0));
    }

    None
}

// TODO: support to split on capture #N instead of whole match
pub fn split_on_matches(
    text: RopeSlice,
    selection: &Selection,
    regex: &crate::regex::Regex,
) -> Selection {
    let mut result = SmallVec::with_capacity(selection.len());

    for sel in selection {
        // TODO: can't avoid occasional allocations since Regex can't operate on chunks yet
        let fragment = sel.fragment(text);

        let mut sel_start = sel.from();
        let sel_end = sel.to();

        let mut start_byte = text.char_to_byte(sel_start);

        let mut start = sel_start;

        for mat in regex.find_iter(&fragment) {
            // TODO: retain range direction

            let end = text.byte_to_char(start_byte + mat.start());
            result.push(Range::new(start, end.saturating_sub(1)));
            start = text.byte_to_char(start_byte + mat.end());
        }

        if start <= sel_end {
            result.push(Range::new(start, sel_end));
        }
    }

    // TODO: figure out a new primary index
    Selection::new(result, 0)
}

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

    #[test]
    #[should_panic]
    fn test_new_empty() {
        let sel = Selection::new(smallvec![], 0);
    }

    #[test]
    fn test_create_normalizes_and_merges() {
        let sel = Selection::new(
            smallvec![
                Range::new(10, 12),
                Range::new(6, 7),
                Range::new(4, 5),
                Range::new(3, 4),
                Range::new(0, 6),
                Range::new(7, 8),
                Range::new(9, 13),
                Range::new(13, 14),
            ],
            0,
        );

        let res = sel
            .ranges
            .into_iter()
            .map(|range| format!("{}/{}", range.anchor, range.head))
            .collect::<Vec<String>>()
            .join(",");

        assert_eq!(res, "0/6,6/7,7/8,9/13,13/14");
    }

    #[test]
    fn test_create_merges_adjacent_points() {
        let sel = Selection::new(
            smallvec![
                Range::new(10, 12),
                Range::new(12, 12),
                Range::new(12, 12),
                Range::new(10, 10),
                Range::new(8, 10),
            ],
            0,
        );

        let res = sel
            .ranges
            .into_iter()
            .map(|range| format!("{}/{}", range.anchor, range.head))
            .collect::<Vec<String>>()
            .join(",");

        assert_eq!(res, "8/10,10/12");
    }

    #[test]
    fn test_contains() {
        let range = Range::new(10, 12);

        assert_eq!(range.contains(9), false);
        assert_eq!(range.contains(10), true);
        assert_eq!(range.contains(11), true);
        assert_eq!(range.contains(12), false);
        assert_eq!(range.contains(13), false);

        let range = Range::new(9, 6);
        assert_eq!(range.contains(9), true);
        assert_eq!(range.contains(7), true);
        assert_eq!(range.contains(6), false);
    }

    #[test]
    fn test_split_on_matches() {
        use crate::regex::Regex;

        let text = Rope::from("abcd efg wrs   xyz 123 456");

        let selection = Selection::new(smallvec![Range::new(0, 8), Range::new(10, 19),], 0);

        let result = split_on_matches(text.slice(..), &selection, &Regex::new(r"\s+").unwrap());

        assert_eq!(
            result.ranges(),
            &[
                Range::new(0, 3),
                Range::new(5, 7),
                Range::new(10, 11),
                Range::new(15, 17),
                Range::new(19, 19),
            ]
        );

        assert_eq!(
            result.fragments(text.slice(..)).collect::<Vec<_>>(),
            &["abcd", "efg", "rs", "xyz", "1"]
        );
    }
}