cfgrammar/yacc/grammar.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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518
#![allow(clippy::derive_partial_eq_without_eq)]
use std::{cell::RefCell, collections::HashMap, fmt::Write};
use num_traits::{AsPrimitive, PrimInt, Unsigned};
#[cfg(feature = "serde")]
use serde::{Deserialize, Serialize};
use vob::Vob;
use super::{ast, firsts::YaccFirsts, follows::YaccFollows, parser::YaccGrammarResult, YaccKind};
use crate::{PIdx, RIdx, SIdx, Span, Symbol, TIdx};
const START_RULE: &str = "^";
const IMPLICIT_RULE: &str = "~";
const IMPLICIT_START_RULE: &str = "^~";
pub type PrecedenceLevel = u64;
#[derive(Clone, Copy, Debug, PartialEq)]
#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
pub struct Precedence {
pub level: PrecedenceLevel,
pub kind: AssocKind,
}
#[derive(Clone, Copy, Debug, PartialEq)]
#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
pub enum AssocKind {
Left,
Right,
Nonassoc,
}
/// Representation of a `YaccGrammar`. See the [top-level documentation](../../index.html) for the
/// guarantees this struct makes about rules, tokens, productions, and symbols.
#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
pub struct YaccGrammar<StorageT = u32> {
/// How many rules does this grammar have?
rules_len: RIdx<StorageT>,
/// A mapping from `RIdx` -> `(Span, String)`.
rule_names: Box<[(String, Span)]>,
/// A mapping from `TIdx` -> `Option<(Span, String)>`. Every user-specified token will have a name,
/// but tokens inserted by cfgrammar (e.g. the EOF token) won't.
token_names: Box<[Option<(Span, String)>]>,
/// A mapping from `TIdx` -> `Option<Precedence>`
token_precs: Box<[Option<Precedence>]>,
/// A mapping from `TIdx` -> `Option<String>` for the %epp declaration, giving pretty-printed
/// versions of token names that can be presented to the user in case of an error. Every
/// user-specified token will have a name that can be presented to the user (if a token doesn't
/// have an %epp entry, the token name will be used in lieu), but tokens inserted by cfgrammar
/// (e.g. the EOF token) won't.
token_epp: Box<[Option<String>]>,
/// How many tokens does this grammar have?
tokens_len: TIdx<StorageT>,
/// The offset of the EOF token.
eof_token_idx: TIdx<StorageT>,
/// How many productions does this grammar have?
prods_len: PIdx<StorageT>,
/// Which production is the sole production of the start rule?
start_prod: PIdx<StorageT>,
/// A list of all productions.
prods: Box<[Box<[Symbol<StorageT>]>]>,
/// A mapping from rules to their productions. Note that 1) the order of rules is identical to
/// that of `rule_names` 2) every rule will have at least 1 production 3) productions
/// are not necessarily stored sequentially.
rules_prods: Box<[Box<[PIdx<StorageT>]>]>,
/// A mapping from productions to their corresponding rule indexes.
prods_rules: Box<[RIdx<StorageT>]>,
/// The precedence of each production.
prod_precs: Box<[Option<Precedence>]>,
/// The index of the rule added for implicit tokens, if they were specified; otherwise
/// `None`.
implicit_rule: Option<RIdx<StorageT>>,
/// User defined Rust programs which can be called within actions
actions: Box<[Option<String>]>,
/// A `(name, type)` pair defining an extra parameter to pass to action functions.
parse_param: Option<(String, String)>,
/// Lifetimes for `param_args`
programs: Option<String>,
/// The actiontypes of rules (one per rule).
actiontypes: Box<[Option<String>]>,
/// Tokens marked as %avoid_insert (if any).
avoid_insert: Option<Vob>,
/// How many shift/reduce conflicts the grammar author expected (if any).
expect: Option<usize>,
/// How many reduce/reduce conflicts the grammar author expected (if any).
expectrr: Option<usize>,
}
// Internally, we assume that a grammar's start rule has a single production. Since we manually
// create the start rule ourselves (without relying on user input), this is a safe assumption.
impl YaccGrammar<u32> {
pub fn new(yacc_kind: YaccKind, s: &str) -> YaccGrammarResult<Self> {
YaccGrammar::new_with_storaget(yacc_kind, s)
}
}
impl<StorageT: 'static + PrimInt + Unsigned> YaccGrammar<StorageT>
where
usize: AsPrimitive<StorageT>,
{
/// Takes as input a Yacc grammar of [`YaccKind`](enum.YaccKind.html) as a `String` `s` and returns a
/// [`YaccGrammar`](grammar/struct.YaccGrammar.html) (or
/// ([`YaccGrammarError`](grammar/enum.YaccGrammarError.html) on error).
///
/// As we're compiling the `YaccGrammar`, we add a new start rule (which we'll refer to as `^`,
/// though the actual name is a fresh name that is guaranteed to be unique) that references the
/// user defined start rule.
pub fn new_with_storaget(yacc_kind: YaccKind, s: &str) -> YaccGrammarResult<Self> {
let ast_validation = ast::ASTWithValidityInfo::new(yacc_kind, s);
Self::new_from_ast_with_validity_info(yacc_kind, &ast_validation)
}
pub fn new_from_ast_with_validity_info(
yacc_kind: YaccKind,
ast_validation: &ast::ASTWithValidityInfo,
) -> YaccGrammarResult<Self> {
if !ast_validation.is_valid() {
return Err(ast_validation.errors().to_owned());
}
let ast = ast_validation.ast();
// Check that StorageT is big enough to hold RIdx/PIdx/SIdx/TIdx values; after these
// checks we can guarantee that things like RIdx(ast.rules.len().as_()) are safe.
if ast.rules.len() > num_traits::cast(StorageT::max_value()).unwrap() {
panic!("StorageT is not big enough to store this grammar's rules.");
}
if ast.tokens.len() > num_traits::cast(StorageT::max_value()).unwrap() {
panic!("StorageT is not big enough to store this grammar's tokens.");
}
if ast.prods.len() > num_traits::cast(StorageT::max_value()).unwrap() {
panic!("StorageT is not big enough to store this grammar's productions.");
}
for p in &ast.prods {
if p.symbols.len() > num_traits::cast(StorageT::max_value()).unwrap() {
panic!("StorageT is not big enough to store the symbols of at least one of this grammar's productions.");
}
}
let mut rule_names: Vec<(String, Span)> = Vec::with_capacity(ast.rules.len() + 1);
// Generate a guaranteed unique start rule name. We simply keep making the string longer
// until we've hit something unique (at the very worst, this will require looping for as
// many times as there are rules). We use the same technique later for unique end
// token and whitespace names.
let mut start_rule = START_RULE.to_string();
while ast.rules.get(&start_rule).is_some() {
start_rule += START_RULE;
}
rule_names.push((start_rule.clone(), Span::new(0, 0)));
let implicit_rule;
let implicit_start_rule;
match yacc_kind {
YaccKind::Original(_) | YaccKind::Grmtools => {
implicit_rule = None;
implicit_start_rule = None;
}
YaccKind::Eco => {
if ast.implicit_tokens.is_some() {
let mut n1 = IMPLICIT_RULE.to_string();
while ast.rules.get(&n1).is_some() {
n1 += IMPLICIT_RULE;
}
rule_names.push((n1.clone(), Span::new(0, 0)));
implicit_rule = Some(n1);
let mut n2 = IMPLICIT_START_RULE.to_string();
while ast.rules.get(&n2).is_some() {
n2 += IMPLICIT_START_RULE;
}
rule_names.push((n2.clone(), Span::new(0, 0)));
implicit_start_rule = Some(n2);
} else {
implicit_rule = None;
implicit_start_rule = None;
}
}
};
for (
k,
ast::Rule {
name: (_, name_span),
..
},
) in &ast.rules
{
rule_names.push((k.clone(), *name_span));
}
let mut rules_prods: Vec<Vec<PIdx<StorageT>>> = Vec::with_capacity(rule_names.len());
let mut rule_map = HashMap::<String, RIdx<StorageT>>::new();
for (i, (v, _)) in rule_names.iter().enumerate() {
rules_prods.push(Vec::new());
rule_map.insert(v.clone(), RIdx(i.as_()));
}
let mut token_names: Vec<Option<(Span, String)>> = Vec::with_capacity(ast.tokens.len() + 1);
let mut token_precs: Vec<Option<Precedence>> = Vec::with_capacity(ast.tokens.len() + 1);
let mut token_epp: Vec<Option<String>> = Vec::with_capacity(ast.tokens.len() + 1);
for (i, k) in ast.tokens.iter().enumerate() {
token_names.push(Some((ast.spans[i], k.clone())));
token_precs.push(ast.precs.get(k).map(|(prec, _)| prec).cloned());
token_epp.push(Some(
ast.epp.get(k).map(|(_, (s, _))| s).unwrap_or(k).clone(),
));
}
let eof_token_idx = TIdx(token_names.len().as_());
token_names.push(None);
token_precs.push(None);
token_epp.push(None);
let mut token_map = HashMap::<String, TIdx<StorageT>>::new();
for (i, v) in token_names.iter().enumerate() {
if let Some((_, n)) = v.as_ref() {
token_map.insert(n.clone(), TIdx(i.as_()));
}
}
// In order to avoid fiddling about with production indices from the AST, we simply map
// tem 1:1 to grammar indices. That means that any new productions are added to the *end*
// of the list of productions.
let mut prods = vec![None; ast.prods.len()];
let mut prod_precs: Vec<Option<Option<Precedence>>> = vec![None; ast.prods.len()];
let mut prods_rules = vec![None; ast.prods.len()];
let mut actions = vec![None; ast.prods.len()];
let mut actiontypes = vec![None; rule_names.len()];
let (start_name, _) = ast.start.as_ref().unwrap();
for (astrulename, _) in &rule_names {
let ridx = rule_map[astrulename];
if astrulename == &start_rule {
// Add the special start rule which has a single production which references a
// single rule.
rules_prods[usize::from(ridx)].push(PIdx(prods.len().as_()));
let start_prod = match implicit_start_rule {
None => {
// Add ^: S;
vec![Symbol::Rule(rule_map[start_name])]
}
Some(ref s) => {
// An implicit rule has been specified, so the special start rule
// needs to reference the intermediate start rule required. Therefore add:
// ^: ^~;
vec![Symbol::Rule(rule_map[s])]
}
};
prods.push(Some(start_prod));
prod_precs.push(Some(None));
prods_rules.push(Some(ridx));
actions.push(None);
continue;
} else if implicit_start_rule
.as_ref()
.map_or(false, |s| s == astrulename)
{
// Add the intermediate start rule (handling implicit tokens at the beginning of
// the file):
// ^~: ~ S;
rules_prods[usize::from(rule_map[astrulename])].push(PIdx(prods.len().as_()));
prods.push(Some(vec![
Symbol::Rule(rule_map[implicit_rule.as_ref().unwrap()]),
Symbol::Rule(rule_map[start_name]),
]));
prod_precs.push(Some(None));
prods_rules.push(Some(ridx));
continue;
} else if implicit_rule.as_ref().map_or(false, |s| s == astrulename) {
// Add the implicit rule: ~: "IMPLICIT_TOKEN_1" ~ | ... | "IMPLICIT_TOKEN_N" ~ | ;
let implicit_prods = &mut rules_prods[usize::from(rule_map[astrulename])];
// Add a production for each implicit token
for (t, _) in ast.implicit_tokens.as_ref().unwrap().iter() {
implicit_prods.push(PIdx(prods.len().as_()));
prods.push(Some(vec![Symbol::Token(token_map[t]), Symbol::Rule(ridx)]));
prod_precs.push(Some(None));
prods_rules.push(Some(ridx));
}
// Add an empty production
implicit_prods.push(PIdx(prods.len().as_()));
prods.push(Some(vec![]));
prod_precs.push(Some(None));
prods_rules.push(Some(ridx));
continue;
} else {
actiontypes[usize::from(ridx)] = ast.rules[astrulename].actiont.clone();
}
let rule = &mut rules_prods[usize::from(ridx)];
for &pidx in &ast.rules[astrulename].pidxs {
let astprod = &ast.prods[pidx];
let mut prod = Vec::with_capacity(astprod.symbols.len());
for astsym in &astprod.symbols {
match *astsym {
ast::Symbol::Rule(ref n, _) => {
prod.push(Symbol::Rule(rule_map[n]));
}
ast::Symbol::Token(ref n, _) => {
prod.push(Symbol::Token(token_map[n]));
if let Some(implicit_rule) = &implicit_rule {
prod.push(Symbol::Rule(rule_map[implicit_rule]));
}
}
};
}
let mut prec = None;
if let Some(ref n) = astprod.precedence {
prec = Some(ast.precs[n]);
} else {
for astsym in astprod.symbols.iter().rev() {
if let ast::Symbol::Token(ref n, _) = *astsym {
if let Some(p) = ast.precs.get(n) {
prec = Some(*p);
}
break;
}
}
}
(*rule).push(PIdx(pidx.as_()));
prods[pidx] = Some(prod);
prod_precs[pidx] = Some(prec.map(|(prec, _)| prec));
prods_rules[pidx] = Some(ridx);
if let Some(ref s) = astprod.action {
actions[pidx] = Some(s.clone());
}
}
}
let avoid_insert = if let Some(ai) = &ast.avoid_insert {
let mut aiv = Vob::from_elem(false, token_names.len());
for n in ai.keys() {
aiv.set(usize::from(token_map[n]), true);
}
Some(aiv)
} else {
None
};
assert!(!token_names.is_empty());
assert!(!rule_names.is_empty());
Ok(YaccGrammar {
rules_len: RIdx(rule_names.len().as_()),
rule_names: rule_names.into_boxed_slice(),
tokens_len: TIdx(token_names.len().as_()),
eof_token_idx,
token_names: token_names.into_boxed_slice(),
token_precs: token_precs.into_boxed_slice(),
token_epp: token_epp.into_boxed_slice(),
prods_len: PIdx(prods.len().as_()),
start_prod: rules_prods[usize::from(rule_map[&start_rule])][0],
rules_prods: rules_prods
.iter()
.map(|x| x.iter().copied().collect())
.collect(),
prods_rules: prods_rules.into_iter().map(Option::unwrap).collect(),
prods: prods
.into_iter()
.map(|x| x.unwrap().into_boxed_slice())
.collect(),
prod_precs: prod_precs.into_iter().map(Option::unwrap).collect(),
implicit_rule: implicit_rule.map(|x| rule_map[&x]),
actions: actions.into_boxed_slice(),
parse_param: ast.parse_param.clone(),
programs: ast.programs.clone(),
avoid_insert,
actiontypes: actiontypes.into_boxed_slice(),
expect: ast.expect.map(|(n, _)| n),
expectrr: ast.expectrr.map(|(n, _)| n),
})
}
/// How many productions does this grammar have?
pub fn prods_len(&self) -> PIdx<StorageT> {
self.prods_len
}
/// Return an iterator which produces (in order from `0..self.prods_len()`) all this
/// grammar's valid `PIdx`s.
pub fn iter_pidxs(&self) -> impl Iterator<Item = PIdx<StorageT>> {
// We can use as_ safely, because we know that we're only generating integers from
// 0..self.rules_len() and, since rules_len() returns an RIdx<StorageT>, then by
// definition the integers we're creating fit within StorageT.
Box::new((0..usize::from(self.prods_len())).map(|x| PIdx(x.as_())))
}
/// Get the sequence of symbols for production `pidx`. Panics if `pidx` doesn't exist.
pub fn prod(&self, pidx: PIdx<StorageT>) -> &[Symbol<StorageT>] {
&self.prods[usize::from(pidx)]
}
/// How many symbols does production `pidx` have? Panics if `pidx` doesn't exist.
pub fn prod_len(&self, pidx: PIdx<StorageT>) -> SIdx<StorageT> {
// Since we've already checked that StorageT can store all the symbols for every production
// in the grammar, the call to as_ is safe.
SIdx(self.prods[usize::from(pidx)].len().as_())
}
/// Return the rule index of the production `pidx`. Panics if `pidx` doesn't exist.
pub fn prod_to_rule(&self, pidx: PIdx<StorageT>) -> RIdx<StorageT> {
self.prods_rules[usize::from(pidx)]
}
/// Return the precedence of production `pidx` (where `None` indicates "no precedence specified").
/// Panics if `pidx` doesn't exist.
pub fn prod_precedence(&self, pidx: PIdx<StorageT>) -> Option<Precedence> {
self.prod_precs[usize::from(pidx)]
}
/// Return the production index of the start rule's sole production (for Yacc grammars the
/// start rule is defined to have precisely one production).
pub fn start_prod(&self) -> PIdx<StorageT> {
self.start_prod
}
/// How many rules does this grammar have?
pub fn rules_len(&self) -> RIdx<StorageT> {
self.rules_len
}
/// Return an iterator which produces (in order from `0..self.rules_len()`) all this
/// grammar's valid `RIdx`s.
pub fn iter_rules(&self) -> impl Iterator<Item = RIdx<StorageT>> {
// We can use as_ safely, because we know that we're only generating integers from
// 0..self.rules_len() and, since rules_len() returns an RIdx<StorageT>, then by
// definition the integers we're creating fit within StorageT.
Box::new((0..usize::from(self.rules_len())).map(|x| RIdx(x.as_())))
}
/// Return the productions for rule `ridx`. Panics if `ridx` doesn't exist.
pub fn rule_to_prods(&self, ridx: RIdx<StorageT>) -> &[PIdx<StorageT>] {
&self.rules_prods[usize::from(ridx)]
}
/// Return the name of rule `ridx`. Panics if `ridx` doesn't exist.
#[deprecated(since = "0.13.0", note = "Please use rule_name_str instead")]
pub fn rule_name(&self, ridx: RIdx<StorageT>) -> &str {
self.rule_name_str(ridx)
}
/// Return the name of rule `ridx`. Panics if `ridx` doesn't exist.
pub fn rule_name_str(&self, ridx: RIdx<StorageT>) -> &str {
let (name, _) = &self.rule_names[usize::from(ridx)];
name.as_str()
}
/// Return the span of rule `ridx`. Panics if `ridx` doesn't exist.
pub fn rule_name_span(&self, ridx: RIdx<StorageT>) -> Span {
let (_, span) = self.rule_names[usize::from(ridx)];
span
}
/// Return the `RIdx` of the implict rule if it exists, or `None` otherwise.
pub fn implicit_rule(&self) -> Option<RIdx<StorageT>> {
self.implicit_rule
}
/// Return the index of the rule named `n` or `None` if it doesn't exist.
pub fn rule_idx(&self, n: &str) -> Option<RIdx<StorageT>> {
self.rule_names
.iter()
.position(|(x, _)| x == n)
// The call to as_() is safe because rule_names is guaranteed to be
// small enough to fit into StorageT
.map(|x| RIdx(x.as_()))
}
/// What is the index of the start rule? Note that cfgrammar will have inserted at least one
/// rule "above" the user's start rule.
pub fn start_rule_idx(&self) -> RIdx<StorageT> {
self.prod_to_rule(self.start_prod)
}
/// How many tokens does this grammar have?
pub fn tokens_len(&self) -> TIdx<StorageT> {
self.tokens_len
}
/// Return an iterator which produces (in order from `0..self.tokens_len()`) all this
/// grammar's valid `TIdx`s.
pub fn iter_tidxs(&self) -> impl Iterator<Item = TIdx<StorageT>> {
// We can use as_ safely, because we know that we're only generating integers from
// 0..self.rules_len() and, since rules_len() returns an TIdx<StorageT>, then by
// definition the integers we're creating fit within StorageT.
Box::new((0..usize::from(self.tokens_len())).map(|x| TIdx(x.as_())))
}
/// Return the index of the end token.
pub fn eof_token_idx(&self) -> TIdx<StorageT> {
self.eof_token_idx
}
/// Return the name of token `tidx` (where `None` indicates "the rule has no name"). Panics if
/// `tidx` doesn't exist.
pub fn token_name(&self, tidx: TIdx<StorageT>) -> Option<&str> {
self.token_names[usize::from(tidx)]
.as_ref()
.map(|(_, x)| x.as_str())
}
/// Return the precedence of token `tidx` (where `None` indicates "no precedence specified").
/// Panics if `tidx` doesn't exist.
pub fn token_precedence(&self, tidx: TIdx<StorageT>) -> Option<Precedence> {
self.token_precs[usize::from(tidx)]
}
/// Return the %epp entry for token `tidx` (where `None` indicates "the token has no
/// pretty-printed value"). Panics if `tidx` doesn't exist.
pub fn token_epp(&self, tidx: TIdx<StorageT>) -> Option<&str> {
self.token_epp[usize::from(tidx)].as_deref()
}
/// Return the span for token given by `tidx` if one exists.
/// If `None`, the token is either implicit and not derived from a token
/// in the source, otherwise the `YaccGrammar` itself may not derived from a
/// textual source in which case the token may be explicit but still lack spans
/// from its construction.
pub fn token_span(&self, tidx: TIdx<StorageT>) -> Option<Span> {
self.token_names[usize::from(tidx)]
.as_ref()
.map(|(span, _)| *span)
}
/// Get the action for production `pidx`. Panics if `pidx` doesn't exist.
pub fn action(&self, pidx: PIdx<StorageT>) -> &Option<String> {
&self.actions[usize::from(pidx)]
}
pub fn actiontype(&self, ridx: RIdx<StorageT>) -> &Option<String> {
&self.actiontypes[usize::from(ridx)]
}
pub fn parse_param(&self) -> &Option<(String, String)> {
&self.parse_param
}
/// Get the programs part of the grammar
pub fn programs(&self) -> &Option<String> {
&self.programs
}
/// Returns a map from names to `TIdx`s of all tokens that a lexer will need to generate valid
/// inputs from this grammar.
pub fn tokens_map(&self) -> HashMap<&str, TIdx<StorageT>> {
let mut m = HashMap::with_capacity(usize::from(self.tokens_len) - 1);
for tidx in self.iter_tidxs() {
if let Some((_, n)) = self.token_names[usize::from(tidx)].as_ref() {
m.insert(&**n, tidx);
}
}
m
}
/// Return the index of the token named `n` or `None` if it doesn't exist.
pub fn token_idx(&self, n: &str) -> Option<TIdx<StorageT>> {
self.token_names
.iter()
.position(|x| x.as_ref().map_or(false, |(_, x)| x == n))
// The call to as_() is safe because token_names is guaranteed to be small
// enough to fit into StorageT
.map(|x| TIdx(x.as_()))
}
/// Is the token `tidx` marked as `%avoid_insert`?
pub fn avoid_insert(&self, tidx: TIdx<StorageT>) -> bool {
if let Some(ai) = &self.avoid_insert {
ai.get(usize::from(tidx)).unwrap()
} else {
false
}
}
// How many shift/reduce conflicts were expected?
pub fn expect(&self) -> Option<usize> {
self.expect
}
// How many reduce/reduce conflicts were expected?
pub fn expectrr(&self) -> Option<usize> {
self.expectrr
}
/// Is there a path from the `from` rule to the `to` rule? Note that recursive rules
/// return `true` for a path from themselves to themselves.
pub fn has_path(&self, from: RIdx<StorageT>, to: RIdx<StorageT>) -> bool {
let mut seen = vec![];
seen.resize(usize::from(self.rules_len()), false);
let mut todo = vec![];
todo.resize(usize::from(self.rules_len()), false);
todo[usize::from(from)] = true;
loop {
let mut empty = true;
for ridx in self.iter_rules() {
if !todo[usize::from(ridx)] {
continue;
}
seen[usize::from(ridx)] = true;
todo[usize::from(ridx)] = false;
empty = false;
for pidx in self.rule_to_prods(ridx).iter() {
for sym in self.prod(*pidx) {
if let Symbol::Rule(p_ridx) = *sym {
if p_ridx == to {
return true;
}
if !seen[usize::from(p_ridx)] {
todo[usize::from(p_ridx)] = true;
}
}
}
}
}
if empty {
return false;
}
}
}
/// Returns the string representation of a given production `pidx`.
pub fn pp_prod(&self, pidx: PIdx<StorageT>) -> String {
let mut sprod = String::new();
let ridx = self.prod_to_rule(pidx);
sprod.push_str(self.rule_name_str(ridx));
sprod.push(':');
for sym in self.prod(pidx) {
let s = match sym {
Symbol::Token(tidx) => self.token_name(*tidx).unwrap(),
Symbol::Rule(ridx) => self.rule_name_str(*ridx),
};
write!(sprod, " \"{}\"", s).ok();
}
sprod
}
/// Return a `SentenceGenerator` which can then generate minimal sentences for any rule
/// based on the user-defined `token_cost` function which gives the associated cost for
/// generating each token (where the cost must be greater than 0). Note that multiple
/// tokens can have the same score. The simplest cost function is thus `|_| 1`.
pub fn sentence_generator<F>(&self, token_cost: F) -> SentenceGenerator<StorageT>
where
F: Fn(TIdx<StorageT>) -> u8,
{
SentenceGenerator::new(self, token_cost)
}
/// Return a `YaccFirsts` struct for this grammar.
pub fn firsts(&self) -> YaccFirsts<StorageT> {
YaccFirsts::new(self)
}
/// Return a `YaccFirsts` struct for this grammar.
pub fn follows(&self) -> YaccFollows<StorageT> {
YaccFollows::new(self)
}
}
/// A `SentenceGenerator` can generate minimal sentences for any given rule. e.g. for the
/// grammar:
///
/// ```text
/// %start A
/// %%
/// A: A B | ;
/// B: C | D;
/// C: 'x' B | 'x';
/// D: 'y' B | 'y' 'z';
/// ```
///
/// the following are valid minimal sentences:
///
/// ```text
/// A: []
/// B: [x]
/// C: [x]
/// D: [y, x] or [y, z]
/// ```
pub struct SentenceGenerator<'a, StorageT> {
grm: &'a YaccGrammar<StorageT>,
rule_min_costs: RefCell<Option<Vec<u16>>>,
rule_max_costs: RefCell<Option<Vec<u16>>>,
token_costs: Vec<u8>,
}
impl<'a, StorageT: 'static + PrimInt + Unsigned> SentenceGenerator<'a, StorageT>
where
usize: AsPrimitive<StorageT>,
{
fn new<F>(grm: &'a YaccGrammar<StorageT>, token_cost: F) -> Self
where
F: Fn(TIdx<StorageT>) -> u8,
{
let mut token_costs = Vec::with_capacity(usize::from(grm.tokens_len()));
for tidx in grm.iter_tidxs() {
token_costs.push(token_cost(tidx));
}
SentenceGenerator {
grm,
token_costs,
rule_min_costs: RefCell::new(None),
rule_max_costs: RefCell::new(None),
}
}
/// What is the cost of a minimal sentence for the rule `ridx`? Note that, unlike
/// `min_sentence`, this function does not actually *build* a sentence and it is thus much
/// faster.
pub fn min_sentence_cost(&self, ridx: RIdx<StorageT>) -> u16 {
self.rule_min_costs
.borrow_mut()
.get_or_insert_with(|| rule_min_costs(self.grm, &self.token_costs))[usize::from(ridx)]
}
/// What is the cost of a maximal sentence for the rule `ridx`? Rules which can generate
/// sentences of unbounded length return None; rules which can only generate maximal strings of
/// a finite length return a `Some(u16)`.
pub fn max_sentence_cost(&self, ridx: RIdx<StorageT>) -> Option<u16> {
let v = self
.rule_max_costs
.borrow_mut()
.get_or_insert_with(|| rule_max_costs(self.grm, &self.token_costs))[usize::from(ridx)];
if v == u16::MAX {
None
} else {
Some(v)
}
}
/// Non-deterministically return a minimal sentence from the set of minimal sentences for the
/// rule `ridx`.
pub fn min_sentence(&self, ridx: RIdx<StorageT>) -> Vec<TIdx<StorageT>> {
let cheapest_prod = |p_ridx: RIdx<StorageT>| -> PIdx<StorageT> {
let mut low_sc = None;
let mut low_idx = None;
for &pidx in self.grm.rule_to_prods(p_ridx).iter() {
let mut sc = 0;
for sym in self.grm.prod(pidx).iter() {
sc += match *sym {
Symbol::Rule(i) => self.min_sentence_cost(i),
Symbol::Token(i) => u16::from(self.token_costs[usize::from(i)]),
};
}
if low_sc.is_none() || Some(sc) < low_sc {
low_sc = Some(sc);
low_idx = Some(pidx);
}
}
low_idx.unwrap()
};
let mut s = vec![];
let mut st = vec![(cheapest_prod(ridx), 0)];
while let Some((pidx, sym_idx)) = st.pop() {
let prod = self.grm.prod(pidx);
for (sidx, sym) in prod.iter().enumerate().skip(sym_idx) {
match sym {
Symbol::Rule(s_ridx) => {
st.push((pidx, sidx + 1));
st.push((cheapest_prod(*s_ridx), 0));
}
Symbol::Token(s_tidx) => {
s.push(*s_tidx);
}
}
}
}
s
}
/// Return (in arbitrary order) all the minimal sentences for the rule `ridx`.
pub fn min_sentences(&self, ridx: RIdx<StorageT>) -> Vec<Vec<TIdx<StorageT>>> {
let cheapest_prods = |p_ridx: RIdx<StorageT>| -> Vec<PIdx<StorageT>> {
let mut low_sc = None;
let mut low_idxs = vec![];
for &pidx in self.grm.rule_to_prods(p_ridx).iter() {
let mut sc = 0;
for sym in self.grm.prod(pidx).iter() {
sc += match *sym {
Symbol::Rule(s_ridx) => self.min_sentence_cost(s_ridx),
Symbol::Token(s_tidx) => u16::from(self.token_costs[usize::from(s_tidx)]),
};
}
if low_sc.is_none() || Some(sc) <= low_sc {
if Some(sc) < low_sc {
low_idxs.clear();
}
low_sc = Some(sc);
low_idxs.push(pidx);
}
}
low_idxs
};
let mut sts = Vec::new(); // Output sentences
for pidx in cheapest_prods(ridx) {
let prod = self.grm.prod(pidx);
if prod.is_empty() {
sts.push(vec![]);
continue;
}
// We construct the minimal sentences in two phases.
//
// First, for each symbol in the production, we gather all the possible minimal
// sentences for it. If, for the grammar:
// X: 'a' Y
// Y: 'b' | 'c'
// we ask for the minimal sentences of X's only production we'll end up with a vec of
// vecs as follows:
// [[['a']], [['b'], ['c']]]
let mut ms = Vec::with_capacity(prod.len());
for sym in prod {
match *sym {
Symbol::Rule(s_ridx) => ms.push(self.min_sentences(s_ridx)),
Symbol::Token(s_tidx) => ms.push(vec![vec![s_tidx]]),
}
}
// Second, we need to generate all combinations of the gathered sentences. We do this
// by writing our own simple numeric incrementing scheme. If we rewrite the list from
// above as follows:
//
// 0 1 <- call this axis "i"
// 0: a b
// 1: c
// ^
// |
// call this axis "todo"
//
// this hopefully becomes easier to see. Let's call the list "ms": the combinations we
// need to generate are thus:
//
// ms[0][0] + ms[1][0] (i.e. 'ab')
// ms[0][0] + ms[1][1] (i.e. 'ac')
//
// The easiest way to model this is to have a list (todo) with each entry starting at
// 0. After each iteration around the loop (i) we add 1 to the last todo column's
// entry: if that spills over the length of the corresponding ms entry, then we reset
// that column to zero, and try adding 1 to the previous column (as many times as
// needed). If the first column spills, then we're done. This is basically normal
// arithmetic but with each digit having an arbitrary base.
let mut todo = vec![0; prod.len()];
let mut cur = Vec::new();
'b: loop {
for i in 0..todo.len() {
cur.extend(&ms[i][todo[i]]);
}
sts.push(std::mem::take(&mut cur));
let mut j = todo.len() - 1;
loop {
if todo[j] + 1 == ms[j].len() {
if j == 0 {
break 'b;
}
todo[j] = 0;
j -= 1;
} else {
todo[j] += 1;
break;
}
}
}
}
sts
}
}
/// Return the cost of a minimal string for each rule in this grammar. The cost of a
/// token is specified by the user-defined `token_cost` function.
#[allow(clippy::unnecessary_unwrap)]
fn rule_min_costs<StorageT: 'static + PrimInt + Unsigned>(
grm: &YaccGrammar<StorageT>,
token_costs: &[u8],
) -> Vec<u16>
where
usize: AsPrimitive<StorageT>,
{
// We use a simple(ish) fixed-point algorithm to determine costs. We maintain two lists
// "costs" and "done". An integer costs[i] starts at 0 and monotonically increments
// until done[i] is true, at which point costs[i] value is fixed. We also use the done
// list as a simple "todo" list: whilst there is at least one false value in done, there is
// still work to do.
//
// On each iteration of the loop, we examine each rule in the todo list to see if
// we can get a better idea of its true cost. Some are trivial:
// * A rule with an empty production immediately has a cost of 0.
// * Rules whose productions don't reference any rules (i.e. only contain tokens) can be
// immediately given a cost by calculating the lowest-cost production.
// However if a rule A references another rule B, we may need to wait until
// we've fully analysed B before we can cost A. This might seem to cause problems with
// recursive rules, so we introduce the concept of "incomplete costs" i.e. if a production
// references a rule we can work out its minimum possible cost simply by counting
// the production's token costs. Since rules can have a mix of complete and
// incomplete productions, this is sometimes enough to allow us to assign a final cost to
// a rule (if the lowest complete production's cost is lower than or equal to all
// the lowest incomplete production's cost). This allows us to make progress, since it
// means that we can iteratively improve our knowledge of a token's minimum cost:
// eventually we will reach a point where we can determine it definitively.
let mut costs = vec![0; usize::from(grm.rules_len())];
let mut done = vec![false; usize::from(grm.rules_len())];
loop {
let mut all_done = true;
for i in 0..done.len() {
if done[i] {
continue;
}
all_done = false;
let mut ls_cmplt = None; // lowest completed cost
let mut ls_noncmplt = None; // lowest non-completed cost
// The call to as_() is guaranteed safe because done.len() == grm.rules_len(), and
// we guarantee that grm.rules_len() can fit in StorageT.
for pidx in grm.rule_to_prods(RIdx(i.as_())).iter() {
let mut c: u16 = 0; // production cost
let mut cmplt = true;
for sym in grm.prod(*pidx) {
let sc = match *sym {
Symbol::Token(tidx) => u16::from(token_costs[usize::from(tidx)]),
Symbol::Rule(ridx) => {
if !done[usize::from(ridx)] {
cmplt = false;
}
costs[usize::from(ridx)]
}
};
c = c
.checked_add(sc)
.expect("Overflow occurred when calculating rule costs");
}
if cmplt && (ls_cmplt.is_none() || Some(c) < ls_cmplt) {
ls_cmplt = Some(c);
} else if !cmplt && (ls_noncmplt.is_none() || Some(c) < ls_noncmplt) {
ls_noncmplt = Some(c);
}
}
if ls_cmplt.is_some() && (ls_noncmplt.is_none() || ls_cmplt < ls_noncmplt) {
debug_assert!(ls_cmplt.unwrap() >= costs[i]);
costs[i] = ls_cmplt.unwrap();
done[i] = true;
} else if let Some(ls_noncmplt) = ls_noncmplt {
debug_assert!(ls_noncmplt >= costs[i]);
costs[i] = ls_noncmplt;
}
}
if all_done {
debug_assert!(done.iter().all(|x| *x));
break;
}
}
costs
}
/// Return the cost of the maximal string for each rule in this grammar (u32::max_val()
/// representing "this rule can generate strings of infinite length"). The cost of a
/// token is specified by the user-defined `token_cost` function.
#[allow(clippy::unnecessary_unwrap)]
fn rule_max_costs<StorageT: 'static + PrimInt + Unsigned>(
grm: &YaccGrammar<StorageT>,
token_costs: &[u8],
) -> Vec<u16>
where
usize: AsPrimitive<StorageT>,
{
let mut done = vec![false; usize::from(grm.rules_len())];
let mut costs = vec![0; usize::from(grm.rules_len())];
// First mark all recursive rules.
for ridx in grm.iter_rules() {
// Calling has_path so frequently is not exactly efficient...
if grm.has_path(ridx, ridx) {
costs[usize::from(ridx)] = u16::MAX;
done[usize::from(ridx)] = true;
}
}
loop {
let mut all_done = true;
for i in 0..done.len() {
if done[i] {
continue;
}
all_done = false;
let mut hs_cmplt = None; // highest completed cost
let mut hs_noncmplt = None; // highest non-completed cost
// The call to as_() is guaranteed safe because done.len() == grm.rules_len(), and
// we guarantee that grm.rules_len() can fit in StorageT.
'a: for pidx in grm.rule_to_prods(RIdx(i.as_())).iter() {
let mut c: u16 = 0; // production cost
let mut cmplt = true;
for sym in grm.prod(*pidx) {
let sc = match *sym {
Symbol::Token(s_tidx) => u16::from(token_costs[usize::from(s_tidx)]),
Symbol::Rule(s_ridx) => {
if costs[usize::from(s_ridx)] == u16::MAX {
// As soon as we find reference to an infinite rule, we
// can stop looking.
hs_cmplt = Some(u16::MAX);
break 'a;
}
if !done[usize::from(s_ridx)] {
cmplt = false;
}
costs[usize::from(s_ridx)]
}
};
c = c
.checked_add(sc)
.expect("Overflow occurred when calculating rule costs");
if c == u16::MAX {
panic!("Unable to represent cost in 64 bits.");
}
}
if cmplt && (hs_cmplt.is_none() || Some(c) > hs_cmplt) {
hs_cmplt = Some(c);
} else if !cmplt && (hs_noncmplt.is_none() || Some(c) > hs_noncmplt) {
hs_noncmplt = Some(c);
}
}
if hs_cmplt.is_some() && (hs_noncmplt.is_none() || hs_cmplt > hs_noncmplt) {
debug_assert!(hs_cmplt.unwrap() >= costs[i]);
costs[i] = hs_cmplt.unwrap();
done[i] = true;
} else if let Some(hs_noncmplt) = hs_noncmplt {
debug_assert!(hs_noncmplt >= costs[i]);
costs[i] = hs_noncmplt;
}
}
if all_done {
debug_assert!(done.iter().all(|x| *x));
break;
}
}
costs
}
#[cfg(test)]
mod test {
use super::{
super::{AssocKind, Precedence, YaccGrammar, YaccKind, YaccOriginalActionKind},
rule_max_costs, rule_min_costs, IMPLICIT_RULE, IMPLICIT_START_RULE,
};
use crate::{PIdx, RIdx, Span, Symbol, TIdx};
use std::collections::HashMap;
macro_rules! bslice {
() => (
::Vec::new().into_boxed_slice()
);
($elem:expr; $n:expr) => (
::vec::from_elem($elem, $n).into_boxed_slice()
);
($($x:expr),+ $(,)?) => (
<[_]>::into_vec(
Box::new([$($x),+])
).into_boxed_slice()
);
}
#[test]
fn test_minimal() {
let grm = YaccGrammar::new(
YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
"%start R %token T %% R: 'T';",
)
.unwrap();
assert_eq!(grm.start_prod, PIdx(1));
assert_eq!(grm.implicit_rule(), None);
grm.rule_idx("^").unwrap();
grm.rule_idx("R").unwrap();
grm.token_idx("T").unwrap();
assert_eq!(&*grm.rules_prods, &[bslice![PIdx(1)], bslice![PIdx(0)]]);
let start_prod = grm.prod(grm.rules_prods[usize::from(grm.rule_idx("^").unwrap())][0]);
assert_eq!(*start_prod, [Symbol::Rule(grm.rule_idx("R").unwrap())]);
let r_prod = grm.prod(grm.rules_prods[usize::from(grm.rule_idx("R").unwrap())][0]);
assert_eq!(*r_prod, [Symbol::Token(grm.token_idx("T").unwrap())]);
assert_eq!(&*grm.prods_rules, &[RIdx(1), RIdx(0)]);
assert_eq!(
grm.tokens_map(),
[("T", TIdx(0))]
.iter()
.cloned()
.collect::<HashMap<&str, TIdx<_>>>()
);
assert_eq!(grm.iter_rules().collect::<Vec<_>>(), vec![RIdx(0), RIdx(1)]);
}
#[test]
fn test_rule_ref() {
let grm = YaccGrammar::new(
YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
"%start R %token T %% R : S; S: 'T';",
)
.unwrap();
grm.rule_idx("^").unwrap();
grm.rule_idx("R").unwrap();
grm.rule_idx("S").unwrap();
grm.token_idx("T").unwrap();
assert!(grm.token_name(grm.eof_token_idx()).is_none());
assert_eq!(
&*grm.rules_prods,
&[bslice![PIdx(2)], bslice![PIdx(0)], bslice![PIdx(1)]]
);
let start_prod = grm.prod(grm.rules_prods[usize::from(grm.rule_idx("^").unwrap())][0]);
assert_eq!(*start_prod, [Symbol::Rule(grm.rule_idx("R").unwrap())]);
let r_prod = grm.prod(grm.rules_prods[usize::from(grm.rule_idx("R").unwrap())][0]);
assert_eq!(r_prod.len(), 1);
assert_eq!(r_prod[0], Symbol::Rule(grm.rule_idx("S").unwrap()));
let s_prod = grm.prod(grm.rules_prods[usize::from(grm.rule_idx("S").unwrap())][0]);
assert_eq!(s_prod.len(), 1);
assert_eq!(s_prod[0], Symbol::Token(grm.token_idx("T").unwrap()));
}
#[test]
#[rustfmt::skip]
fn test_long_prod() {
let grm = YaccGrammar::new(
YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
"%start R %token T1 T2 %% R : S 'T1' S; S: 'T2';"
).unwrap();
grm.rule_idx("^").unwrap();
grm.rule_idx("R").unwrap();
grm.rule_idx("S").unwrap();
grm.token_idx("T1").unwrap();
grm.token_idx("T2").unwrap();
assert_eq!(&*grm.rules_prods, &[bslice![PIdx(2)],
bslice![PIdx(0)],
bslice![PIdx(1)]]);
assert_eq!(&*grm.prods_rules, &[RIdx(1),
RIdx(2),
RIdx(0)]);
let start_prod = grm.prod(grm.rules_prods[usize::from(grm.rule_idx("^").unwrap())][0]);
assert_eq!(*start_prod, [Symbol::Rule(grm.rule_idx("R").unwrap())]);
let r_prod = grm.prod(grm.rules_prods[usize::from(grm.rule_idx("R").unwrap())][0]);
assert_eq!(r_prod.len(), 3);
assert_eq!(r_prod[0], Symbol::Rule(grm.rule_idx("S").unwrap()));
assert_eq!(r_prod[1], Symbol::Token(grm.token_idx("T1").unwrap()));
assert_eq!(r_prod[2], Symbol::Rule(grm.rule_idx("S").unwrap()));
let s_prod = grm.prod(grm.rules_prods[usize::from(grm.rule_idx("S").unwrap())][0]);
assert_eq!(s_prod.len(), 1);
assert_eq!(s_prod[0], Symbol::Token(grm.token_idx("T2").unwrap()));
}
#[test]
fn test_prods_rules() {
let grm = YaccGrammar::new(
YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
"
%start A
%%
A: B
| C;
B: 'x';
C: 'y'
| 'z';
",
)
.unwrap();
assert_eq!(
&*grm.prods_rules,
&[RIdx(1), RIdx(1), RIdx(2), RIdx(3), RIdx(3), RIdx(0)]
);
}
#[test]
#[rustfmt::skip]
fn test_left_right_nonassoc_precs() {
let grm = YaccGrammar::new(
YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
"
%start Expr
%right '='
%left '+' '-'
%left '/'
%left '*'
%nonassoc '~'
%%
Expr : Expr '=' Expr
| Expr '+' Expr
| Expr '-' Expr
| Expr '/' Expr
| Expr '*' Expr
| Expr '~' Expr
| 'id' ;
").unwrap();
assert_eq!(grm.prod_precs.len(), 8);
assert_eq!(grm.prod_precs[0].unwrap(), Precedence{level: 0, kind: AssocKind::Right});
assert_eq!(grm.prod_precs[1].unwrap(), Precedence{level: 1, kind: AssocKind::Left});
assert_eq!(grm.prod_precs[2].unwrap(), Precedence{level: 1, kind: AssocKind::Left});
assert_eq!(grm.prod_precs[3].unwrap(), Precedence{level: 2, kind: AssocKind::Left});
assert_eq!(grm.prod_precs[4].unwrap(), Precedence{level: 3, kind: AssocKind::Left});
assert_eq!(grm.prod_precs[5].unwrap(), Precedence{level: 4, kind: AssocKind::Nonassoc});
assert!(grm.prod_precs[6].is_none());
assert_eq!(grm.prod_precs[7], None);
}
#[test]
#[rustfmt::skip]
fn test_prec_override() {
let grm = YaccGrammar::new(
YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
"
%start expr
%left '+' '-'
%left '*' '/'
%%
expr : expr '+' expr
| expr '-' expr
| expr '*' expr
| expr '/' expr
| '-' expr %prec '*'
| 'id' ;
"
).unwrap();
assert_eq!(grm.prod_precs.len(), 7);
assert_eq!(grm.prod_precs[0].unwrap(), Precedence{level: 0, kind: AssocKind::Left});
assert_eq!(grm.prod_precs[1].unwrap(), Precedence{level: 0, kind: AssocKind::Left});
assert_eq!(grm.prod_precs[2].unwrap(), Precedence{level: 1, kind: AssocKind::Left});
assert_eq!(grm.prod_precs[3].unwrap(), Precedence{level: 1, kind: AssocKind::Left});
assert_eq!(grm.prod_precs[4].unwrap(), Precedence{level: 1, kind: AssocKind::Left});
assert!(grm.prod_precs[5].is_none());
assert_eq!(grm.prod_precs[6], None);
}
#[test]
#[rustfmt::skip]
fn test_implicit_tokens_rewrite() {
let grm = YaccGrammar::new(
YaccKind::Eco,
"
%implicit_tokens ws1 ws2
%start S
%%
S: 'a' | T;
T: 'c' |;
"
).unwrap();
// Check that the above grammar has been rewritten to:
// ^ : ^~;
// ^~: ~ S;
// ~ : ws1 | ws2 | ;
// S : 'a' ~ | T;
// T : 'c' ~ | ;
assert_eq!(grm.prod_precs.len(), 9);
let itfs_rule_idx = grm.rule_idx(IMPLICIT_START_RULE).unwrap();
assert_eq!(grm.rules_prods[usize::from(itfs_rule_idx)].len(), 1);
let itfs_prod1 = &grm.prods[usize::from(grm.rules_prods[usize::from(itfs_rule_idx)][0])];
assert_eq!(itfs_prod1.len(), 2);
assert_eq!(itfs_prod1[0], Symbol::Rule(grm.rule_idx(IMPLICIT_RULE).unwrap()));
assert_eq!(itfs_prod1[1], Symbol::Rule(grm.rule_idx("S").unwrap()));
let s_rule_idx = grm.rule_idx("S").unwrap();
assert_eq!(grm.rules_prods[usize::from(s_rule_idx)].len(), 2);
let s_prod1 = &grm.prods[usize::from(grm.rules_prods[usize::from(s_rule_idx)][0])];
assert_eq!(s_prod1.len(), 2);
assert_eq!(s_prod1[0], Symbol::Token(grm.token_idx("a").unwrap()));
assert_eq!(s_prod1[1], Symbol::Rule(grm.rule_idx(IMPLICIT_RULE).unwrap()));
let s_prod2 = &grm.prods[usize::from(grm.rules_prods[usize::from(s_rule_idx)][1])];
assert_eq!(s_prod2.len(), 1);
assert_eq!(s_prod2[0], Symbol::Rule(grm.rule_idx("T").unwrap()));
let t_rule_idx = grm.rule_idx("T").unwrap();
assert_eq!(grm.rules_prods[usize::from(s_rule_idx)].len(), 2);
let t_prod1 = &grm.prods[usize::from(grm.rules_prods[usize::from(t_rule_idx)][0])];
assert_eq!(t_prod1.len(), 2);
assert_eq!(t_prod1[0], Symbol::Token(grm.token_idx("c").unwrap()));
assert_eq!(t_prod1[1], Symbol::Rule(grm.rule_idx(IMPLICIT_RULE).unwrap()));
let t_prod2 = &grm.prods[usize::from(grm.rules_prods[usize::from(t_rule_idx)][1])];
assert_eq!(t_prod2.len(), 0);
assert_eq!(Some(grm.rule_idx(IMPLICIT_RULE).unwrap()), grm.implicit_rule());
let i_rule_idx = grm.rule_idx(IMPLICIT_RULE).unwrap();
assert_eq!(grm.rules_prods[usize::from(i_rule_idx)].len(), 3);
let i_prod1 = &grm.prods[usize::from(grm.rules_prods[usize::from(i_rule_idx)][0])];
let i_prod2 = &grm.prods[usize::from(grm.rules_prods[usize::from(i_rule_idx)][1])];
assert_eq!(i_prod1.len(), 2);
assert_eq!(i_prod2.len(), 2);
// We don't know what order the implicit rule will contain our tokens in,
// hence the awkward dance below.
let cnd1 = bslice![
Symbol::Token(grm.token_idx("ws1").unwrap()),
Symbol::Rule(grm.implicit_rule().unwrap()),
];
let cnd2 = bslice![
Symbol::Token(grm.token_idx("ws2").unwrap()),
Symbol::Rule(grm.implicit_rule().unwrap()),
];
assert!((*i_prod1 == cnd1 && *i_prod2 == cnd2) || (*i_prod1 == cnd2 && *i_prod2 == cnd1));
let i_prod3 = &grm.prods[usize::from(grm.rules_prods[usize::from(i_rule_idx)][2])];
assert_eq!(i_prod3.len(), 0);
}
#[test]
#[rustfmt::skip]
fn test_has_path() {
let grm = YaccGrammar::new(
YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
"
%start A
%%
A: B;
B: B 'x' | C;
C: C 'y' | ;
"
).unwrap();
let a_ridx = grm.rule_idx("A").unwrap();
let b_ridx = grm.rule_idx("B").unwrap();
let c_ridx = grm.rule_idx("C").unwrap();
assert!(grm.has_path(a_ridx, b_ridx));
assert!(grm.has_path(a_ridx, c_ridx));
assert!(grm.has_path(b_ridx, b_ridx));
assert!(grm.has_path(b_ridx, c_ridx));
assert!(grm.has_path(c_ridx, c_ridx));
assert!(!grm.has_path(a_ridx, a_ridx));
assert!(!grm.has_path(b_ridx, a_ridx));
assert!(!grm.has_path(c_ridx, a_ridx));
}
#[test]
#[rustfmt::skip]
fn test_rule_min_costs() {
let grm = YaccGrammar::new(
YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
"
%start A
%%
A: A B | ;
B: C | D | E;
C: 'x' B | 'x';
D: 'y' B | 'y' 'z';
E: 'x' A | 'x' 'y';
"
).unwrap();
let scores = rule_min_costs(&grm, &[1, 1, 1]);
assert_eq!(scores[usize::from(grm.rule_idx("A").unwrap())], 0);
assert_eq!(scores[usize::from(grm.rule_idx("B").unwrap())], 1);
assert_eq!(scores[usize::from(grm.rule_idx("C").unwrap())], 1);
assert_eq!(scores[usize::from(grm.rule_idx("D").unwrap())], 2);
assert_eq!(scores[usize::from(grm.rule_idx("E").unwrap())], 1);
}
#[test]
fn test_min_sentences() {
let grm = YaccGrammar::new(
YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
"
%start A
%%
A: A B | ;
B: C | D;
C: 'x' B | 'x';
D: 'y' B | 'y' 'z';
",
)
.unwrap();
let sg = grm.sentence_generator(|_| 1);
let find = |nt_name: &str, str_cnds: Vec<Vec<&str>>| {
let cnds = str_cnds
.iter()
.map(|x| {
x.iter()
.map(|y| grm.token_idx(y).unwrap())
.collect::<Vec<_>>()
})
.collect::<Vec<_>>();
let ms = sg.min_sentence(grm.rule_idx(nt_name).unwrap());
if !cnds.iter().any(|x| x == &ms) {
panic!("{:?} doesn't have any matches in {:?}", ms, str_cnds);
}
let min_sts = sg.min_sentences(grm.rule_idx(nt_name).unwrap());
assert_eq!(cnds.len(), min_sts.len());
for ms in min_sts {
if !cnds.iter().any(|x| x == &ms) {
panic!("{:?} doesn't have any matches in {:?}", ms, str_cnds);
}
}
};
find("A", vec![vec![]]);
find("B", vec![vec!["x"]]);
find("C", vec![vec!["x"]]);
find("D", vec![vec!["y", "x"], vec!["y", "z"]]);
}
#[test]
#[rustfmt::skip]
fn test_rule_max_costs1() {
let grm = YaccGrammar::new(
YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
"
%start A
%%
A: A B | ;
B: C | D | E;
C: 'x' B | 'x';
D: 'y' B | 'y' 'z';
E: 'x' A | 'x' 'y';
"
).unwrap();
let scores = rule_max_costs(&grm, &[1, 1, 1]);
assert_eq!(scores[usize::from(grm.rule_idx("A").unwrap())], u16::MAX);
assert_eq!(scores[usize::from(grm.rule_idx("B").unwrap())], u16::MAX);
assert_eq!(scores[usize::from(grm.rule_idx("C").unwrap())], u16::MAX);
assert_eq!(scores[usize::from(grm.rule_idx("D").unwrap())], u16::MAX);
assert_eq!(scores[usize::from(grm.rule_idx("E").unwrap())], u16::MAX);
}
#[test]
#[rustfmt::skip]
fn test_rule_max_costs2() {
let grm = YaccGrammar::new(
YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
"
%start A
%%
A: A B | B;
B: C | D;
C: 'x' 'y' | 'x';
D: 'y' 'x' | 'y' 'x' 'z';
"
).unwrap();
let scores = rule_max_costs(&grm, &[1, 1, 1]);
assert_eq!(scores[usize::from(grm.rule_idx("A").unwrap())], u16::MAX);
assert_eq!(scores[usize::from(grm.rule_idx("B").unwrap())], 3);
assert_eq!(scores[usize::from(grm.rule_idx("C").unwrap())], 2);
assert_eq!(scores[usize::from(grm.rule_idx("D").unwrap())], 3);
}
#[test]
fn test_out_of_order_productions() {
// Example taken from p54 of Locally least-cost error repair in LR parsers, Carl Cerecke
let grm = YaccGrammar::new(
YaccKind::Original(YaccOriginalActionKind::GenericParseTree),
"
%start S
%%
S: A 'c' 'd'
| B 'c' 'e';
A: 'a';
B: 'a'
| 'b';
A: 'b';
",
)
.unwrap();
assert_eq!(
&*grm.prods_rules,
&[
RIdx(1),
RIdx(1),
RIdx(2),
RIdx(3),
RIdx(3),
RIdx(2),
RIdx(0)
]
);
}
#[test]
fn test_token_spans() {
let src = "%%\nAB: 'a' | 'foo';";
let grm =
YaccGrammar::new(YaccKind::Original(YaccOriginalActionKind::NoAction), src).unwrap();
let token_map = grm.tokens_map();
let a_tidx = token_map.get("a");
let foo_tidx = token_map.get("foo");
let a_span = grm.token_span(*a_tidx.unwrap()).unwrap();
let foo_span = grm.token_span(*foo_tidx.unwrap()).unwrap();
let ab_span = grm.rule_name_span(grm.rule_idx("AB").unwrap());
assert_eq!(a_span, Span::new(8, 9));
assert_eq!(foo_span, Span::new(14, 17));
assert_eq!(ab_span, Span::new(3, 5));
assert_eq!(&src[a_span.start()..a_span.end()], "a");
assert_eq!(&src[foo_span.start()..foo_span.end()], "foo");
assert_eq!(&src[ab_span.start()..ab_span.end()], "AB");
}
#[test]
fn token_span_issue296() {
let src = "%%
S: | AB;
A: 'a' 'b';
B: 'b' 'c';
AB: A AB | B ';' AB;
%%
";
let grm =
YaccGrammar::new(YaccKind::Original(YaccOriginalActionKind::NoAction), src).unwrap();
let token_map = grm.tokens_map();
let c_tidx = token_map.get("c").unwrap();
assert_eq!(grm.token_name(*c_tidx), Some("c"));
let c_span = grm.token_span(*c_tidx).unwrap();
assert_eq!(&src[c_span.start()..c_span.end()], "c");
}
}