cfgrammar/yacc/
firsts.rs

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
use std::marker::PhantomData;

use num_traits::{AsPrimitive, PrimInt, Unsigned};
use vob::Vob;

use super::YaccGrammar;
use crate::{RIdx, Symbol, TIdx};

/// `Firsts` stores all the first sets for a given grammar. For example, given this code and
/// grammar:
/// ```text
///   let grm = YaccGrammar::new(YaccKind::Original(YaccOriginalActionKind::GenericParseTree), "
///     S: A 'b';
///     A: 'a'
///      | ;").unwrap();
///   let firsts = Firsts::new(&grm);
/// ```
/// then the following assertions (and only the following assertions) about the firsts set are
/// correct:
/// ```text
///   assert!(firsts.is_set(grm.rule_idx("S").unwrap(), grm.token_idx("a").unwrap()));
///   assert!(firsts.is_set(grm.rule_idx("S").unwrap(), grm.token_idx("b").unwrap()));
///   assert!(firsts.is_set(grm.rule_idx("A").unwrap(), grm.token_idx("a").unwrap()));
///   assert!(firsts.is_epsilon_set(grm.rule_idx("A").unwrap()));
/// ```
#[derive(Debug)]
pub struct YaccFirsts<StorageT> {
    firsts: Vec<Vob>,
    epsilons: Vob,
    phantom: PhantomData<StorageT>,
}

impl<StorageT: 'static + PrimInt + Unsigned> YaccFirsts<StorageT>
where
    usize: AsPrimitive<StorageT>,
{
    /// Generates and returns the firsts set for the given grammar.
    pub fn new(grm: &YaccGrammar<StorageT>) -> Self {
        let mut firsts = YaccFirsts {
            firsts: vec![
                Vob::from_elem(false, usize::from(grm.tokens_len()));
                usize::from(grm.rules_len())
            ],
            epsilons: Vob::from_elem(false, usize::from(grm.rules_len())),
            phantom: PhantomData,
        };

        // Loop looking for changes to the firsts set, until we reach a fixed point. In essence, we
        // look at each rule E, and see if any of the rules at the start of its productions
        // have new elements in since we last looked. If they do, we'll have to do another round.
        loop {
            let mut changed = false;
            for ridx in grm.iter_rules() {
                // For each rule E
                for &pidx in grm.rule_to_prods(ridx).iter() {
                    // ...and each production A
                    let prod = grm.prod(pidx);
                    if prod.is_empty() {
                        // if it's an empty production, ensure this rule's epsilon bit is
                        // set.
                        if !firsts.is_epsilon_set(ridx) {
                            firsts.epsilons.set(usize::from(ridx), true);
                            changed = true;
                        }
                        continue;
                    }
                    for (sidx, sym) in prod.iter().enumerate() {
                        match *sym {
                            Symbol::Token(s_tidx) => {
                                // if symbol is a token, add to FIRSTS
                                if !firsts.set(ridx, s_tidx) {
                                    changed = true;
                                }
                                break;
                            }
                            Symbol::Rule(s_ridx) => {
                                // if we're dealing with another rule, union its FIRSTs
                                // together with the current rules FIRSTs. Note this is
                                // (intentionally) a no-op if the two tokens are one and the
                                // same.
                                for tidx in grm.iter_tidxs() {
                                    if firsts.is_set(s_ridx, tidx) && !firsts.set(ridx, tidx) {
                                        changed = true;
                                    }
                                }

                                // If the epsilon bit in the rule being referenced is set,
                                // and if its the last symbol in the production, then add epsilon
                                // to FIRSTs.
                                if firsts.is_epsilon_set(s_ridx) && sidx == prod.len() - 1 {
                                    // Only add epsilon if the symbol is the last in the production
                                    if !firsts.epsilons[usize::from(ridx)] {
                                        firsts.epsilons.set(usize::from(ridx), true);
                                        changed = true;
                                    }
                                }

                                // If FIRST(X) of production R : X Y2 Y3 doesn't contain epsilon,
                                // then don't try and calculate the FIRSTS of Y2 or Y3 (i.e. stop
                                // now).
                                if !firsts.is_epsilon_set(s_ridx) {
                                    break;
                                }
                            }
                        }
                    }
                }
            }
            if !changed {
                return firsts;
            }
        }
    }

    /// Return all the firsts for rule `ridx`.
    pub fn firsts(&self, ridx: RIdx<StorageT>) -> &Vob {
        &self.firsts[usize::from(ridx)]
    }

    /// Returns true if the token `tidx` is in the first set for rule `ridx`.
    pub fn is_set(&self, ridx: RIdx<StorageT>, tidx: TIdx<StorageT>) -> bool {
        self.firsts[usize::from(ridx)][usize::from(tidx)]
    }

    /// Returns true if the rule `ridx` has epsilon in its first set.
    pub fn is_epsilon_set(&self, ridx: RIdx<StorageT>) -> bool {
        self.epsilons[usize::from(ridx)]
    }

    /// Ensures that the firsts bit for token `tidx` rule `ridx` is set. Returns true if
    /// it was already set, or false otherwise.
    pub fn set(&mut self, ridx: RIdx<StorageT>, tidx: TIdx<StorageT>) -> bool {
        let r = &mut self.firsts[usize::from(ridx)];
        if r[usize::from(tidx)] {
            true
        } else {
            r.set(usize::from(tidx), true);
            false
        }
    }
}

#[cfg(test)]
mod test {
    use super::{
        super::{YaccGrammar, YaccKind, YaccOriginalActionKind},
        YaccFirsts,
    };
    use num_traits::{AsPrimitive, PrimInt, Unsigned};

    fn has<StorageT: 'static + PrimInt + Unsigned>(
        grm: &YaccGrammar<StorageT>,
        firsts: &YaccFirsts<StorageT>,
        rn: &str,
        should_be: Vec<&str>,
    ) where
        usize: AsPrimitive<StorageT>,
    {
        let ridx = grm.rule_idx(rn).unwrap();
        for tidx in grm.iter_tidxs() {
            let n = grm.token_name(tidx).unwrap_or("<no name>");
            match should_be.iter().position(|&x| x == n) {
                Some(_) => {
                    if !firsts.is_set(ridx, tidx) {
                        panic!("{} is not set in {}", n, rn);
                    }
                }
                None => {
                    if firsts.is_set(ridx, tidx) {
                        panic!("{} is incorrectly set in {}", n, rn);
                    }
                }
            }
        }
        if should_be.iter().any(|x| x == &"") {
            assert!(firsts.is_epsilon_set(ridx));
        }
    }

    #[test]
    fn test_first() {
        let grm = YaccGrammar::new(
            YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
            "
          %start C
          %token c d
          %%
          C: 'c';
          D: 'd';
          E: D | C;
          F: E;
          ",
        )
        .unwrap();
        let firsts = grm.firsts();
        has(&grm, &firsts, "^", vec!["c"]);
        has(&grm, &firsts, "D", vec!["d"]);
        has(&grm, &firsts, "E", vec!["d", "c"]);
        has(&grm, &firsts, "F", vec!["d", "c"]);
    }

    #[test]
    fn test_first_no_subsequent_rules() {
        let grm = YaccGrammar::new(
            YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
            "
          %start C
          %token c d
          %%
          C: 'c';
          D: 'd';
          E: D C;
          ",
        )
        .unwrap();
        let firsts = grm.firsts();
        has(&grm, &firsts, "E", vec!["d"]);
    }

    #[test]
    fn test_first_epsilon() {
        let grm = YaccGrammar::new(
            YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
            "
          %start A
          %token a b c
          %%
          A: B 'a';
          B: 'b' | ;
          C: 'c' | ;
          D: C;
          ",
        )
        .unwrap();
        let firsts = grm.firsts();
        has(&grm, &firsts, "A", vec!["b", "a"]);
        has(&grm, &firsts, "C", vec!["c", ""]);
        has(&grm, &firsts, "D", vec!["c", ""]);
    }

    #[test]
    fn test_last_epsilon() {
        let grm = YaccGrammar::new(
            YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
            "
          %start A
          %token b c
          %%
          A: B C;
          B: 'b' | ;
          C: B 'c' B;
          ",
        )
        .unwrap();
        let firsts = grm.firsts();
        has(&grm, &firsts, "A", vec!["b", "c"]);
        has(&grm, &firsts, "B", vec!["b", ""]);
        has(&grm, &firsts, "C", vec!["b", "c"]);
    }

    #[test]
    fn test_first_no_multiples() {
        let grm = YaccGrammar::new(
            YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
            "
          %start A
          %token b c
          %%
          A: B 'b';
          B: 'b' | ;
          ",
        )
        .unwrap();
        let firsts = grm.firsts();
        has(&grm, &firsts, "A", vec!["b"]);
    }

    fn eco_grammar() -> YaccGrammar {
        YaccGrammar::new(
            YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
            "
          %start S
          %token a b c d f
          %%
          S: S 'b' | 'b' A 'a' | 'a';
          A: 'a' S 'c' | 'a' | 'a' S 'b';
          B: A S;
          C: D A;
          D: 'd' | ;
          F: C D 'f';
          ",
        )
        .unwrap()
    }

    #[test]
    fn test_first_from_eco() {
        let grm = eco_grammar();
        let firsts = grm.firsts();
        has(&grm, &firsts, "S", vec!["a", "b"]);
        has(&grm, &firsts, "A", vec!["a"]);
        has(&grm, &firsts, "B", vec!["a"]);
        has(&grm, &firsts, "D", vec!["d", ""]);
        has(&grm, &firsts, "C", vec!["d", "a"]);
        has(&grm, &firsts, "F", vec!["d", "a"]);
    }

    #[test]
    fn test_first_from_eco_bug() {
        let grm = YaccGrammar::new(
            YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
            "
          %start E
          %token a b c d e f
          %%
          E : T | E 'b' T;
          T : P | T 'e' P;
          P : 'a';
          C: C 'c' | ;
          D: D 'd' | F;
          F: 'f' | ;
          G: C D;
          ",
        )
        .unwrap();
        let firsts = grm.firsts();
        has(&grm, &firsts, "E", vec!["a"]);
        has(&grm, &firsts, "T", vec!["a"]);
        has(&grm, &firsts, "P", vec!["a"]);
        has(&grm, &firsts, "C", vec!["c", ""]);
        has(&grm, &firsts, "D", vec!["f", "d", ""]);
        has(&grm, &firsts, "G", vec!["c", "d", "f", ""]);
    }
}