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
//! A fallback "parser" for plain text.

use lazy_static::lazy_static;
use line_numbers::LinePositions;
use regex::Regex;

use crate::words::split_words;
use crate::{
    diff::myers_diff,
    parse::syntax::{AtomKind, MatchKind, MatchedPos, TokenKind},
};

const MAX_WORDS_IN_LINE: usize = 1000;

fn split_lines_keep_newline(s: &str) -> Vec<&str> {
    lazy_static! {
        static ref NEWLINE_RE: Regex = Regex::new("\n").unwrap();
    }

    let mut offset = 0;
    let mut lines = vec![];
    for newline_match in NEWLINE_RE.find_iter(s) {
        lines.push(s[offset..newline_match.end()].into());
        offset = newline_match.end();
    }

    if offset < s.len() {
        lines.push(s[offset..].into());
    }

    lines
}

#[derive(Debug)]
enum TextChangeKind {
    Novel,
    Unchanged,
}

/// Merge contiguous sequences marked as novel.
fn merge_novel<'a>(
    lines: &[(TextChangeKind, Vec<&'a str>, Vec<&'a str>)],
) -> Vec<(TextChangeKind, Vec<&'a str>, Vec<&'a str>)> {
    let mut lhs_novel = vec![];
    let mut rhs_novel = vec![];

    let mut res: Vec<(TextChangeKind, Vec<_>, Vec<_>)> = vec![];
    for (kind, lhs_lines, rhs_lines) in lines {
        match kind {
            TextChangeKind::Novel => {
                lhs_novel.extend(lhs_lines.iter().copied());
                rhs_novel.extend(rhs_lines.iter().copied());
            }
            TextChangeKind::Unchanged => {
                if !lhs_novel.is_empty() || !rhs_novel.is_empty() {
                    res.push((TextChangeKind::Novel, lhs_novel, rhs_novel));
                    lhs_novel = vec![];
                    rhs_novel = vec![];
                }

                res.push((
                    TextChangeKind::Unchanged,
                    lhs_lines.clone(),
                    rhs_lines.clone(),
                ));
            }
        }
    }

    if !lhs_novel.is_empty() || !rhs_novel.is_empty() {
        res.push((TextChangeKind::Novel, lhs_novel, rhs_novel));
    }
    res
}

fn changed_parts<'a>(
    src: &'a str,
    opposite_src: &'a str,
) -> Vec<(TextChangeKind, Vec<&'a str>, Vec<&'a str>)> {
    let src_lines = split_lines_keep_newline(src);
    let opposite_src_lines = split_lines_keep_newline(opposite_src);

    let mut res: Vec<(TextChangeKind, Vec<&'a str>, Vec<&'a str>)> = vec![];
    for diff_res in myers_diff::slice_unique_by_hash(&src_lines, &opposite_src_lines) {
        match diff_res {
            myers_diff::DiffResult::Left(line) => {
                res.push((TextChangeKind::Novel, vec![line], vec![]));
            }
            myers_diff::DiffResult::Both(line, opposite_line) => {
                res.push((TextChangeKind::Unchanged, vec![line], vec![opposite_line]));
            }
            myers_diff::DiffResult::Right(opposite_line) => {
                res.push((TextChangeKind::Novel, vec![], vec![opposite_line]));
            }
        }
    }

    // Merge so we can try word highlighting on similar lines.
    merge_novel(&res)
}

fn line_len_in_bytes(line: &str) -> usize {
    if line.ends_with('\n') {
        line.len() - 1
    } else {
        line.len()
    }
}

// TODO: Prefer src/opposite_src nomenclature as this function is called from both sides.
pub(crate) fn change_positions(lhs_src: &str, rhs_src: &str) -> Vec<MatchedPos> {
    // TODO: If either side is "", don't split each line by words
    // pointlessly. This is common for file additions/removals.
    let lhs_lp = LinePositions::from(lhs_src);
    let rhs_lp = LinePositions::from(rhs_src);

    let mut lhs_offset = 0;
    let mut rhs_offset = 0;

    let mut mps = vec![];
    for (kind, lhs_lines, rhs_lines) in changed_parts(lhs_src, rhs_src) {
        match kind {
            TextChangeKind::Unchanged => {
                for (lhs_line, rhs_line) in lhs_lines.iter().zip(rhs_lines) {
                    let lhs_pos =
                        lhs_lp.from_region(lhs_offset, lhs_offset + line_len_in_bytes(lhs_line));
                    let rhs_pos =
                        rhs_lp.from_region(rhs_offset, rhs_offset + line_len_in_bytes(rhs_line));

                    mps.push(MatchedPos {
                        kind: MatchKind::UnchangedToken {
                            highlight: TokenKind::Atom(AtomKind::Normal),
                            self_pos: lhs_pos.clone(),
                            opposite_pos: rhs_pos,
                        },
                        pos: lhs_pos[0],
                    });

                    lhs_offset += lhs_line.len();
                    rhs_offset += rhs_line.len();
                }
            }
            TextChangeKind::Novel => {
                let lhs_part = lhs_lines.join("");
                let rhs_part = rhs_lines.join("");

                let lhs_words = split_words(&lhs_part);
                let rhs_words = split_words(&rhs_part);

                // Myers Diff scales badly on large inputs, and
                // word-level diffing is merely nice to have. If we
                // have a very large number of words, don't diff
                // individual words.
                if lhs_words.len() > MAX_WORDS_IN_LINE || rhs_words.len() > MAX_WORDS_IN_LINE {
                    let lhs_pos = lhs_lp.from_region(lhs_offset, lhs_offset + lhs_part.len());
                    mps.push(MatchedPos {
                        kind: MatchKind::NovelWord {
                            highlight: TokenKind::Atom(AtomKind::Normal),
                        },
                        pos: lhs_pos[0],
                    });

                    lhs_offset += lhs_part.len();
                    rhs_offset += rhs_part.len();
                    continue;
                }

                for diff_res in myers_diff::slice_unique_by_hash(&lhs_words, &rhs_words) {
                    match diff_res {
                        myers_diff::DiffResult::Left(lhs_word) => {
                            let lhs_pos =
                                lhs_lp.from_region(lhs_offset, lhs_offset + lhs_word.len());
                            mps.push(MatchedPos {
                                kind: MatchKind::NovelWord {
                                    highlight: TokenKind::Atom(AtomKind::Normal),
                                },
                                pos: lhs_pos[0],
                            });

                            lhs_offset += lhs_word.len();
                        }
                        myers_diff::DiffResult::Both(lhs_word, rhs_word) => {
                            if *lhs_word != "\n" {
                                let lhs_pos =
                                    lhs_lp.from_region(lhs_offset, lhs_offset + lhs_word.len());
                                let rhs_pos =
                                    rhs_lp.from_region(rhs_offset, rhs_offset + rhs_word.len());

                                mps.push(MatchedPos {
                                    kind: MatchKind::NovelLinePart {
                                        highlight: TokenKind::Atom(AtomKind::Normal),
                                        self_pos: lhs_pos[0],
                                        opposite_pos: rhs_pos,
                                    },
                                    pos: lhs_pos[0],
                                });
                            }

                            lhs_offset += lhs_word.len();
                            rhs_offset += rhs_word.len();
                        }
                        myers_diff::DiffResult::Right(rhs_word) => {
                            rhs_offset += rhs_word.len();
                        }
                    }
                }
            }
        }
    }

    mps
}

#[cfg(test)]
mod tests {
    use line_numbers::SingleLineSpan;
    use pretty_assertions::assert_eq;

    use super::*;

    #[test]
    fn test_split_newlines() {
        let s = "foo\nbar\nbaz";
        let res = split_lines_keep_newline(s);
        assert_eq!(res, vec!["foo\n", "bar\n", "baz"])
    }

    #[test]
    fn test_positions_no_changes() {
        let positions = change_positions("foo", "foo");

        assert_eq!(positions.len(), 1);
        assert!(!positions[0].kind.is_novel());
        assert_eq!(
            positions[0].pos,
            SingleLineSpan {
                line: 0.into(),
                start_col: 0,
                end_col: 3
            }
        );
    }

    #[test]
    fn test_positions_whitespace_is_change() {
        // Even though the word exists on both sides, it should still
        // be treated as a change. We're doing a line-based diff and
        // the lines are different.
        let positions = change_positions("foo", " foo");
        assert!(positions[0].kind.is_novel());
    }

    #[test]
    fn test_no_changes_trailing_newlines() {
        let positions = change_positions("foo\n", "foo\n");

        assert_eq!(positions.len(), 1);
        assert!(!positions[0].kind.is_novel());
        assert_eq!(
            positions[0].pos,
            SingleLineSpan {
                line: 0.into(),
                start_col: 0,
                end_col: 3
            }
        );
    }

    #[test]
    fn test_novel_lhs_trailing_newlines() {
        let positions = change_positions("foo\n", "");

        assert_eq!(positions.len(), 2);
        assert!(positions[0].kind.is_novel());
    }

    #[test]
    fn test_positions_novel_lhs() {
        let positions = change_positions("foo", "");

        assert_eq!(positions.len(), 1);
        assert!(positions[0].kind.is_novel());
    }
}