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
use memchr::{memchr, memrchr};
use regex::bytes::{Regex, RegexBuilder};
use syntax;

use literals::LiteralSets;
use nonl;
use Result;

/// A matched line.
#[derive(Clone, Debug, Default, Eq, PartialEq)]
pub struct Match {
    start: usize,
    end: usize,
}

impl Match {
    /// Create a new empty match value.
    pub fn new() -> Match {
        Match::default()
    }

    /// Return the starting byte offset of the line that matched.
    #[inline]
    pub fn start(&self) -> usize {
        self.start
    }

    /// Return the ending byte offset of the line that matched.
    #[inline]
    pub fn end(&self) -> usize {
        self.end
    }
}

/// A fast line oriented regex searcher.
#[derive(Clone, Debug)]
pub struct Grep {
    re: Regex,
    required: Option<Regex>,
    opts: Options,
}

/// A builder for a grep searcher.
#[derive(Clone, Debug)]
pub struct GrepBuilder {
    pattern: String,
    opts: Options,
}

#[derive(Clone, Debug)]
struct Options {
    case_insensitive: bool,
    line_terminator: u8,
    size_limit: usize,
    dfa_size_limit: usize,
}

impl Default for Options {
    fn default() -> Options {
        Options {
            case_insensitive: false,
            line_terminator: b'\n',
            size_limit: 10 * (1 << 20),
            dfa_size_limit: 10 * (1 << 20),
        }
    }
}

impl GrepBuilder {
    /// Create a new builder for line searching.
    ///
    /// The pattern given should be a regular expression. The precise syntax
    /// supported is documented on the regex crate.
    pub fn new(pattern: &str) -> GrepBuilder {
        GrepBuilder {
            pattern: pattern.to_string(),
            opts: Options::default(),
        }
    }

    /// Set the line terminator.
    ///
    /// The line terminator can be any ASCII character and serves to delineate
    /// the match boundaries in the text searched.
    ///
    /// This panics if `ascii_byte` is greater than `0x7F` (i.e., not ASCII).
    pub fn line_terminator(mut self, ascii_byte: u8) -> GrepBuilder {
        assert!(ascii_byte <= 0x7F);
        self.opts.line_terminator = ascii_byte;
        self
    }

    /// Set the case sensitive flag (`i`) on the regex.
    pub fn case_insensitive(mut self, yes: bool) -> GrepBuilder {
        self.opts.case_insensitive = yes;
        self
    }

    /// Set the approximate size limit of the compiled regular expression.
    ///
    /// This roughly corresponds to the number of bytes occupied by a
    /// single compiled program. If the program exceeds this number, then a
    /// compilation error is returned.
    pub fn size_limit(mut self, limit: usize) -> GrepBuilder {
        self.opts.size_limit = limit;
        self
    }

    /// Set the approximate size of the cache used by the DFA.
    ///
    /// This roughly corresponds to the number of bytes that the DFA will use
    /// while searching.
    ///
    /// Note that this is a per thread limit. There is no way to set a global
    /// limit. In particular, if a regex is used from multiple threads
    /// simulanteously, then each thread may use up to the number of bytes
    /// specified here.
    pub fn dfa_size_limit(mut self, limit: usize) -> GrepBuilder {
        self.opts.dfa_size_limit = limit;
        self
    }

    /// Create a line searcher.
    ///
    /// If there was a problem parsing or compiling the regex with the given
    /// options, then an error is returned.
    pub fn build(self) -> Result<Grep> {
        let expr = try!(self.parse());
        let literals = LiteralSets::create(&expr);
        let re = try!(
            RegexBuilder::new(&expr.to_string())
                .case_insensitive(self.opts.case_insensitive)
                .multi_line(true)
                .unicode(true)
                .size_limit(self.opts.size_limit)
                .dfa_size_limit(self.opts.dfa_size_limit)
                .compile()
        );
        Ok(Grep {
            re: re,
            required: literals.to_regex(),
            opts: self.opts,
        })
    }

    /// Parses the underlying pattern and ensures the pattern can never match
    /// the line terminator.
    fn parse(&self) -> Result<syntax::Expr> {
        let expr =
            try!(syntax::ExprBuilder::new()
                 .allow_bytes(true)
                 .unicode(true)
                 .case_insensitive(self.opts.case_insensitive)
                 .parse(&self.pattern));
        debug!("regex ast:\n{:#?}", expr);
        Ok(try!(nonl::remove(expr, self.opts.line_terminator)))
    }
}

impl Grep {
    /// Returns a reference to the underlying regex used by the searcher.
    pub fn regex(&self) -> &Regex {
        &self.re
    }

    /// Returns an iterator over all matches in the given buffer.
    pub fn iter<'b, 's>(&'s self, buf: &'b [u8]) -> Iter<'b, 's> {
        Iter {
            searcher: self,
            buf: buf,
            start: 0,
        }
    }

    /// Fills in the next line that matches in the given buffer starting at
    /// the position given.
    ///
    /// If no match could be found, `false` is returned, otherwise, `true` is
    /// returned.
    pub fn read_match(
        &self,
        mat: &mut Match,
        buf: &[u8],
        mut start: usize,
    ) -> bool {
        if start >= buf.len() {
            return false;
        }
        if let Some(ref req) = self.required {
            while start < buf.len() {
                let e = match req.shortest_match(&buf[start..]) {
                    None => return false,
                    Some(e) => start + e,
                };
                let (prevnl, nextnl) = self.find_line(buf, e, e);
                match self.re.shortest_match(&buf[prevnl..nextnl]) {
                    None => {
                        start = nextnl;
                        continue;
                    }
                    Some(_) => {
                        self.fill_match(mat, prevnl, nextnl);
                        return true;
                    }
                }
            }
            false
        } else {
            let e = match self.re.shortest_match(&buf[start..]) {
                None => return false,
                Some(e) => start + e,
            };
            let (s, e) = self.find_line(buf, e, e);
            self.fill_match(mat, s, e);
            true
        }
    }

    fn fill_match(&self, mat: &mut Match, start: usize, end: usize) {
        mat.start = start;
        mat.end = end;
    }

    fn find_line(&self, buf: &[u8], s: usize, e: usize) -> (usize, usize) {
        (self.find_line_start(buf, s), self.find_line_end(buf, e))
    }

    fn find_line_start(&self, buf: &[u8], pos: usize) -> usize {
        memrchr(self.opts.line_terminator, &buf[0..pos]).map_or(0, |i| i + 1)
    }

    fn find_line_end(&self, buf: &[u8], pos: usize) -> usize {
        memchr(self.opts.line_terminator, &buf[pos..])
            .map_or(buf.len(), |i| pos + i + 1)
    }
}

/// An iterator over all matches in a particular buffer.
///
/// `'b` refers to the lifetime of the buffer, and `'s` refers to the lifetime
/// of the searcher.
pub struct Iter<'b, 's> {
    searcher: &'s Grep,
    buf: &'b [u8],
    start: usize,
}

impl<'b, 's> Iterator for Iter<'b, 's> {
    type Item = Match;

    fn next(&mut self) -> Option<Match> {
        let mut mat = Match::default();
        if !self.searcher.read_match(&mut mat, self.buf, self.start) {
            self.start = self.buf.len();
            return None;
        }
        self.start = mat.end;
        Some(mat)
    }
}

#[cfg(test)]
mod tests {
    #![allow(unused_imports)]

    use memchr::{memchr, memrchr};
    use regex::bytes::Regex;

    use super::{GrepBuilder, Match};

    static SHERLOCK: &'static [u8] = include_bytes!("./data/sherlock.txt");

    #[allow(dead_code)]
    fn s(bytes: &[u8]) -> String {
        String::from_utf8(bytes.to_vec()).unwrap()
    }

    fn find_lines(pat: &str, haystack: &[u8]) -> Vec<Match> {
        let re = Regex::new(pat).unwrap();
        let mut lines = vec![];
        for (s, e) in re.find_iter(haystack) {
            let start = memrchr(b'\n', &haystack[..s])
                        .map_or(0, |i| i + 1);
            let end = memchr(b'\n', &haystack[e..])
                      .map_or(haystack.len(), |i| e + i + 1);
            lines.push(Match {
                start: start,
                end: end,
            });
        }
        lines
    }

    fn grep_lines(pat: &str, haystack: &[u8]) -> Vec<Match> {
        let g = GrepBuilder::new(pat).build().unwrap();
        g.iter(haystack).collect()
    }

    #[test]
    fn buffered_literal() {
        let expected = find_lines("Sherlock Holmes", SHERLOCK);
        let got = grep_lines("Sherlock Holmes", SHERLOCK);
        assert_eq!(expected.len(), got.len());
        assert_eq!(expected, got);
    }
}