liblevenshtein 0.9.1

Levenshtein/Universal Automata for approximate string matching using various dictionary backends
Documentation
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
//! NFA compiler for phonetic regular expressions.
//!
//! This module compiles parsed regex ASTs into NFAs using Thompson's construction.
//!
//! # Examples
//!
//! ```ignore
//! use liblevenshtein::phonetic::nfa::compiler::compile;
//! use liblevenshtein::phonetic::regex::parse;
//!
//! // Compile a simple pattern
//! let nfa = compile(&parse("(ph|f)one").expect("doc: regex parse must succeed")).expect("doc: nfa compile must succeed");
//! assert!(nfa.accepts("phone"));
//! assert!(nfa.accepts("fone"));
//! assert!(!nfa.accepts("bone"));
//!
//! // Compile a rewrite rule
//! let rule = parse_rule("ph -> f").expect("doc: rewrite rule parse must succeed");
//! let rewrite = compile_rewrite(&rule).expect("doc: rewrite compile must succeed");
//! ```

use std::collections::HashMap;

use super::context::{BoundaryKind, ContextPattern, ContextPatternChar};
use super::optimizer::{NfaOptimizerChar, OptimizationConfig};
use super::thompson::{ThompsonBuilder, ThompsonBuilderChar};
use super::{NFAChar, NFA};
use crate::phonetic::regex::ast::{
    ContextExpr, ContextExprByte, Regex, RegexByte, RegexFlags, UnicodeNormalization,
};
use crate::phonetic::regex::error::{ParseError, ParseErrorKind, ParseResult, Position};
use crate::phonetic::regex::transform::apply_flags;

/// Result of compiling a regex with flag support.
///
/// Contains both the compiled NFA and runtime settings extracted from flags.
#[derive(Debug, Clone)]
pub struct CompileResultChar {
    /// The compiled NFA.
    pub nfa: NFAChar,
    /// Unicode normalization to apply to input at runtime.
    pub unicode_normalization: Option<UnicodeNormalization>,
    /// Whether multiline mode is enabled (for `^` and `$`).
    pub multiline: bool,
    /// Whether dotall mode is enabled (`.` matches newlines).
    pub dotall: bool,
    /// Local Levenshtein distance override (from `(?;N)` syntax).
    ///
    /// When set, this distance limit should be used for fuzzy matching
    /// instead of the global distance parameter.
    pub local_distance: Option<u8>,
}

/// Compile a character-level regex to an NFA.
///
/// This applies regex flag transformations (like `(?i)` case-insensitive)
/// before compilation. For flags that affect runtime behavior (unicode
/// normalization, multiline, dotall), use [`compile_with_flags`] instead
/// to access those settings.
///
/// # Arguments
///
/// * `regex` - The regex AST to compile
///
/// # Returns
///
/// An NFA that accepts the language defined by the regex.
///
/// # Examples
///
/// ```ignore
/// use liblevenshtein::phonetic::regex::parse;
/// use liblevenshtein::phonetic::nfa::compiler::compile;
///
/// let regex = parse("a+b*c").expect("doc: regex parse must succeed");
/// let nfa = compile(&regex).expect("doc: nfa compile must succeed");
/// assert!(nfa.accepts("ac"));
/// assert!(nfa.accepts("aaac"));
/// assert!(nfa.accepts("abbc"));
///
/// // Case-insensitive matching via (?i) flag
/// let regex = parse("(?i:hello)").expect("doc: regex parse must succeed");
/// let nfa = compile(&regex).expect("doc: nfa compile must succeed");
/// assert!(nfa.accepts("hello"));
/// assert!(nfa.accepts("HELLO"));
/// assert!(nfa.accepts("HeLLo"));
/// ```
pub fn compile(regex: &Regex) -> ParseResult<NFAChar> {
    let mut compiler = NFACompilerChar::new();
    compiler.compile(regex)
}

/// Compile a character-level regex to an NFA with full flag support.
///
/// This applies regex flag transformations and returns both the NFA and
/// runtime settings extracted from flags like `(?u:NFC)`, `(?m)`, and `(?s)`.
///
/// # Arguments
///
/// * `regex` - The regex AST to compile
///
/// # Returns
///
/// A [`CompileResultChar`] containing the NFA and runtime settings.
///
/// # Examples
///
/// ```ignore
/// use liblevenshtein::phonetic::regex::parse;
/// use liblevenshtein::phonetic::nfa::compiler::compile_with_flags;
///
/// let regex = parse("(?iu:café)").expect("doc: regex parse must succeed");
/// let result = compile_with_flags(&regex).expect("doc: compile_with_flags must succeed");
///
/// // NFA matches case-insensitively (transformed)
/// assert!(result.nfa.accepts("café"));
/// assert!(result.nfa.accepts("CAFÉ"));
///
/// // Unicode normalization available for runtime use
/// if let Some(norm) = result.unicode_normalization {
///     // Apply normalization to input before matching
/// }
/// ```
pub fn compile_with_flags(regex: &Regex) -> ParseResult<CompileResultChar> {
    let mut compiler = NFACompilerChar::new();
    compiler.compile_with_flags(regex)
}

/// Compile a byte-level regex to an NFA.
pub fn compile_bytes(regex: &RegexByte) -> ParseResult<NFA> {
    let mut compiler = NFACompilerByte::new();
    compiler.compile(regex)
}

/// A compiled rewrite rule.
///
/// Contains the source pattern NFA, replacement string, and optional context.
#[derive(Debug, Clone)]
pub struct CompiledRewriteChar {
    /// NFA matching the source pattern
    pub source: NFAChar,
    /// Characters to replace with
    pub replacement: Vec<char>,
    /// Optional left context (lookbehind)
    pub left_context: Option<ContextPatternChar>,
    /// Optional right context (lookahead)
    pub right_context: Option<ContextPatternChar>,
    /// Weight/cost for this rule
    pub weight: f64,
}

/// A compiled byte-level rewrite rule.
#[derive(Debug, Clone)]
pub struct CompiledRewrite {
    /// NFA matching the source pattern
    pub source: NFA,
    /// Bytes to replace with
    pub replacement: Vec<u8>,
    /// Optional left context (lookbehind)
    pub left_context: Option<ContextPattern>,
    /// Optional right context (lookahead)
    pub right_context: Option<ContextPattern>,
    /// Weight/cost for this rule
    pub weight: f64,
}

/// Compile a rewrite rule regex to a compiled rewrite structure.
pub fn compile_rewrite(regex: &Regex) -> ParseResult<CompiledRewriteChar> {
    let mut compiler = NFACompilerChar::new();
    compiler.compile_rewrite(regex)
}

/// Compile a byte-level rewrite rule.
pub fn compile_rewrite_bytes(regex: &RegexByte) -> ParseResult<CompiledRewrite> {
    let mut compiler = NFACompilerByte::new();
    compiler.compile_rewrite(regex)
}

// ============================================================================
// Trampolining support for stack-safe compilation
// ============================================================================

/// Work item for trampolined regex compilation.
///
/// This enum represents the operations that need to be performed during
/// iterative (trampolined) compilation. Using an explicit work stack instead
/// of the call stack allows compilation of deeply nested patterns without
/// risk of stack overflow.
#[derive(Debug)]
enum CompileWork<'a> {
    /// Compile this regex node and push result to value stack.
    Compile(&'a Regex),
    /// Pop two values, concatenate them, push result.
    DoConcatenate,
    /// Pop two values, create alternation, push result.
    DoAlternation,
    /// Pop one value, apply Kleene star, push result.
    DoStar,
    /// Pop one value, apply Kleene plus, push result.
    DoPlus,
    /// Pop one value, apply optional, push result.
    DoOptional,
    /// Pop one value, apply exact repetition, push result.
    DoRepeatExact(usize),
    /// Pop one value, apply range repetition, push result.
    DoRepeatRange(usize, Option<usize>),
}

/// Character-level NFA compiler.
pub struct NFACompilerChar {
    builder: ThompsonBuilderChar,
    /// Symbol table for pattern expansion
    symbols: HashMap<String, Vec<char>>,
    /// Regex flags (multiline, dotall, etc.)
    flags: RegexFlags,
    /// Use trampolining for stack-safe compilation of deep patterns.
    /// When false (default), uses recursive compilation which is faster
    /// but may stack overflow on deeply nested patterns.
    use_trampolining: bool,
    /// Optimization configuration. When Some, the NFA is optimized after compilation.
    /// Default: full optimization enabled.
    optimization: Option<OptimizationConfig>,
}

impl NFACompilerChar {
    /// Create a new compiler with recursive compilation and optimization enabled (default).
    pub fn new() -> Self {
        Self {
            builder: ThompsonBuilderChar::new(),
            symbols: HashMap::new(),
            flags: RegexFlags::default(),
            use_trampolining: false,
            optimization: Some(OptimizationConfig::full()),
        }
    }

    /// Enable trampolining for stack-safe compilation.
    ///
    /// Use this when compiling potentially deep patterns from untrusted input.
    /// Trampolining uses an explicit heap-allocated stack instead of the call stack,
    /// preventing stack overflow on deeply nested patterns like `((((a))))`.
    ///
    /// # Example
    /// ```
    /// use liblevenshtein::phonetic::nfa::NFACompilerChar;
    ///
    /// // For untrusted input, use trampolining
    /// let compiler = NFACompilerChar::new().with_trampolining();
    /// ```
    pub fn with_trampolining(mut self) -> Self {
        self.use_trampolining = true;
        self
    }

    /// Set trampolining mode.
    ///
    /// When `enabled` is true, compilation uses an iterative trampoline-based
    /// approach that is safe for deeply nested patterns but slightly slower.
    /// When false (default), uses recursive compilation which is faster.
    pub fn set_trampolining(&mut self, enabled: bool) {
        self.use_trampolining = enabled;
    }

    /// Check if trampolining is enabled.
    pub fn is_trampolining(&self) -> bool {
        self.use_trampolining
    }

    /// Enable optimization with the given configuration.
    ///
    /// By default, full optimization is enabled. Use this to customize
    /// which optimization passes are applied.
    ///
    /// # Example
    /// ```
    /// use liblevenshtein::phonetic::nfa::{NFACompilerChar, OptimizationConfig};
    ///
    /// // Use quick optimization (no epsilon elimination)
    /// let compiler = NFACompilerChar::new()
    ///     .with_optimization(OptimizationConfig::quick());
    /// ```
    pub fn with_optimization(mut self, config: OptimizationConfig) -> Self {
        self.optimization = Some(config);
        self
    }

    /// Disable optimization entirely.
    ///
    /// This is useful for debugging or when the NFA will be used only once.
    /// The resulting NFA may have epsilon transitions and unreachable states.
    ///
    /// # Example
    /// ```
    /// use liblevenshtein::phonetic::nfa::NFACompilerChar;
    ///
    /// let compiler = NFACompilerChar::new().without_optimization();
    /// ```
    pub fn without_optimization(mut self) -> Self {
        self.optimization = None;
        self
    }

    /// Set the optimization configuration.
    ///
    /// Pass `None` to disable optimization, or `Some(config)` to enable it.
    pub fn set_optimization(&mut self, config: Option<OptimizationConfig>) {
        self.optimization = config;
    }

    /// Check if optimization is enabled.
    pub fn is_optimization_enabled(&self) -> bool {
        self.optimization.is_some()
    }

    /// Get the current optimization configuration.
    pub fn optimization_config(&self) -> Option<&OptimizationConfig> {
        self.optimization.as_ref()
    }

    /// Add a symbol to the symbol table.
    ///
    /// Symbols can be referenced in patterns and will be expanded during compilation.
    pub fn add_symbol(&mut self, name: impl Into<String>, chars: Vec<char>) {
        self.symbols.insert(name.into(), chars);
    }

    /// Set the regex flags for compilation.
    pub fn set_flags(&mut self, flags: RegexFlags) {
        self.flags = flags;
    }

    /// Get the current flags.
    pub fn flags(&self) -> &RegexFlags {
        &self.flags
    }

    /// Get a symbol by name.
    pub fn get_symbol(&self, name: &str) -> Option<&Vec<char>> {
        self.symbols.get(name)
    }

    /// Compile a regex AST to an NFA.
    ///
    /// This applies flag transformations (like `(?i)` case-insensitive) before
    /// compilation. If optimization is enabled (default), the NFA is automatically
    /// optimized after Thompson construction.
    ///
    /// For access to runtime flags (unicode normalization, multiline, dotall),
    /// use [`compile_with_flags`] instead.
    pub fn compile(&mut self, regex: &Regex) -> ParseResult<NFAChar> {
        // Apply flag transformations before compilation
        let transform_result = apply_flags(regex);
        let nfa = self.compile_regex(&transform_result.regex)?;

        // Apply optimization if configured
        let mut nfa = if let Some(ref config) = self.optimization {
            let optimizer = NfaOptimizerChar::new(config.clone());
            let (optimized, _stats) = optimizer.optimize(nfa);
            optimized
        } else {
            nfa
        };

        // H9: Finalize CSR transition table
        nfa.finalize();
        Ok(nfa)
    }

    /// Compile a regex AST to an NFA with full flag support.
    ///
    /// This applies flag transformations and returns both the NFA and runtime
    /// settings extracted from flags like `(?u:NFC)`, `(?m)`, and `(?s)`.
    pub fn compile_with_flags(&mut self, regex: &Regex) -> ParseResult<CompileResultChar> {
        // Apply flag transformations before compilation
        let transform_result = apply_flags(regex);
        let nfa = self.compile_regex(&transform_result.regex)?;

        // Apply optimization if configured
        let nfa = if let Some(ref config) = self.optimization {
            let optimizer = NfaOptimizerChar::new(config.clone());
            let (optimized, _stats) = optimizer.optimize(nfa);
            optimized
        } else {
            nfa
        };

        // H9: Finalize CSR transition table
        let mut nfa = nfa;
        nfa.finalize();

        Ok(CompileResultChar {
            nfa,
            unicode_normalization: transform_result.unicode_normalization,
            multiline: transform_result.multiline,
            dotall: transform_result.dotall,
            local_distance: transform_result.local_distance,
        })
    }

    /// Compile a rewrite rule.
    ///
    /// The source pattern NFA is automatically optimized if optimization is enabled.
    pub fn compile_rewrite(&mut self, regex: &Regex) -> ParseResult<CompiledRewriteChar> {
        match regex {
            Regex::RewriteRule {
                pattern,
                replacement,
                context,
                weight,
            } => {
                // Compile and optimize the source pattern
                let source = self.compile_regex(pattern)?;
                let mut source = if let Some(ref config) = self.optimization {
                    let optimizer = NfaOptimizerChar::new(config.clone());
                    let (optimized, _stats) = optimizer.optimize(source);
                    optimized
                } else {
                    source
                };
                // H9: Finalize CSR transition table
                source.finalize();

                let replacement_chars = self.regex_to_literal(replacement)?;

                let (left_context, right_context) = if let Some(ctx) = context {
                    (
                        ctx.left
                            .as_ref()
                            .map(|l| self.compile_context_expr(l))
                            .transpose()?,
                        ctx.right
                            .as_ref()
                            .map(|r| self.compile_context_expr(r))
                            .transpose()?,
                    )
                } else {
                    (None, None)
                };

                Ok(CompiledRewriteChar {
                    source,
                    replacement: replacement_chars,
                    left_context,
                    right_context,
                    weight: *weight,
                })
            }
            _ => Err(ParseError::new(
                ParseErrorKind::InvalidRewriteRule("expected rewrite rule".to_string()),
                Position::start(),
            )),
        }
    }

    /// Compile a context expression to a context pattern.
    fn compile_context_expr(&mut self, expr: &ContextExpr) -> ParseResult<ContextPatternChar> {
        match expr {
            ContextExpr::Pattern(regex) => {
                let nfa = self.compile_regex(regex)?;
                // Apply optimization if configured
                let mut nfa = if let Some(ref config) = self.optimization {
                    let optimizer = NfaOptimizerChar::new(config.clone());
                    let (optimized, _stats) = optimizer.optimize(nfa);
                    optimized
                } else {
                    nfa
                };
                // H9: Finalize CSR transition table
                nfa.finalize();
                Ok(ContextPatternChar::Nfa(nfa))
            }
            ContextExpr::WordBoundary => {
                // Word boundary is represented as a boundary kind
                // The actual direction (start/end) is determined by position in the context
                Ok(ContextPatternChar::Boundary(BoundaryKind::WordStart))
            }
            ContextExpr::And(a, b) => {
                let left = self.compile_context_expr(a)?;
                let right = self.compile_context_expr(b)?;
                Ok(ContextPatternChar::And(Box::new(left), Box::new(right)))
            }
            ContextExpr::Or(a, b) => {
                let left = self.compile_context_expr(a)?;
                let right = self.compile_context_expr(b)?;
                Ok(ContextPatternChar::Or(Box::new(left), Box::new(right)))
            }
            ContextExpr::Not(inner) => {
                let pattern = self.compile_context_expr(inner)?;
                Ok(ContextPatternChar::Not(Box::new(pattern)))
            }
        }
    }

    /// Compile a regex to NFA.
    ///
    /// This method dispatches to either the recursive or trampolined
    /// implementation based on the `use_trampolining` flag.
    fn compile_regex(&mut self, regex: &Regex) -> ParseResult<NFAChar> {
        if self.use_trampolining {
            self.compile_regex_trampolined(regex)
        } else {
            self.compile_regex_recursive(regex)
        }
    }

    /// Recursive implementation of regex compilation.
    ///
    /// This is the default, faster implementation that uses the call stack.
    /// For deeply nested patterns, use `with_trampolining()` instead.
    #[allow(deprecated)]
    fn compile_regex_recursive(&mut self, regex: &Regex) -> ParseResult<NFAChar> {
        match regex {
            Regex::Empty => Ok(self.builder.epsilon()),
            Regex::Char(c) => Ok(self.builder.single_char(*c)),
            Regex::CharClass(class) => Ok(self.builder.char_class(class.clone())),
            Regex::Any => Ok(self.builder.any_char()),
            Regex::Concat(a, b) => {
                let nfa_a = self.compile_regex_recursive(a)?;
                let nfa_b = self.compile_regex_recursive(b)?;
                Ok(self.builder.concatenate(nfa_a, nfa_b))
            }
            Regex::Alt(a, b) => {
                let nfa_a = self.compile_regex_recursive(a)?;
                let nfa_b = self.compile_regex_recursive(b)?;
                Ok(self.builder.alternation(nfa_a, nfa_b))
            }
            Regex::Star(inner) => {
                let nfa = self.compile_regex_recursive(inner)?;
                Ok(self.builder.kleene_star(nfa))
            }
            Regex::Plus(inner) => {
                let nfa = self.compile_regex_recursive(inner)?;
                Ok(self.builder.kleene_plus(nfa))
            }
            Regex::Optional(inner) => {
                let nfa = self.compile_regex_recursive(inner)?;
                Ok(self.builder.optional(nfa))
            }
            Regex::RepeatExact(inner, n) => {
                let nfa = self.compile_regex_recursive(inner)?;
                Ok(self.builder.repeat_exact(nfa, *n))
            }
            Regex::RepeatRange(inner, min, max) => {
                let nfa = self.compile_regex_recursive(inner)?;
                Ok(self.builder.repeat_range(nfa, *min, *max))
            }
            // Capturing group: (...)
            Regex::CapturingGroup(_, inner) => {
                // Capturing groups compile like regular groups for now
                // (capture semantics would be added in a future phase)
                self.compile_regex_recursive(inner)
            }
            // Non-capturing group: (?:...)
            Regex::NonCapturingGroup(inner) => self.compile_regex_recursive(inner),
            // Named group: (?<name>...)
            Regex::NamedGroup(_, inner) => {
                // Named groups compile like regular groups for now
                self.compile_regex_recursive(inner)
            }
            // Group reference: (?&name)
            Regex::GroupRef(name) => {
                // Group references require the named group's pattern to be available.
                // For now, return an error - in a future phase, we would expand
                // the group reference inline or implement proper subroutine semantics.
                Err(ParseError::new(
                    ParseErrorKind::InvalidGroupReference(format!(
                        "group reference (?&{}) cannot be compiled - subroutine expansion not yet implemented",
                        name
                    )),
                    Position::start(),
                ))
            }
            // Flags group: (?flags:...) or (?flags)
            Regex::FlagsGroup { inner, .. } => {
                // For now, ignore flags and compile the inner pattern if present
                // Flag-based matching (case-insensitive, etc.) would be implemented
                // by transforming the AST before compilation in a future phase
                match inner {
                    Some(inner_regex) => self.compile_regex_recursive(inner_regex),
                    None => Ok(self.builder.epsilon()),
                }
            }
            Regex::WordBoundary => {
                // Word boundary is handled specially in context matching
                // For now, return epsilon (will be handled by context logic)
                Ok(self.builder.epsilon())
            }
            // Anchor assertions (zero-width)
            Regex::StartOfLine => Ok(self.builder.start_of_line()),
            Regex::EndOfLine => Ok(self.builder.end_of_line()),
            Regex::StartOfInput => Ok(self.builder.start_of_input()),
            Regex::EndOfInput => Ok(self.builder.end_of_input()),
            Regex::EndOfInputStrict => Ok(self.builder.end_of_input_strict()),
            Regex::RewriteRule { pattern, .. } => {
                // When compiling a rewrite rule as a pattern, just compile the pattern part
                self.compile_regex_recursive(pattern)
            }
        }
    }

    /// Trampolined (iterative) implementation of regex compilation.
    ///
    /// This implementation uses an explicit heap-allocated work stack instead
    /// of the call stack, making it safe for deeply nested patterns that would
    /// otherwise cause stack overflow.
    #[allow(deprecated)]
    fn compile_regex_trampolined(&mut self, regex: &Regex) -> ParseResult<NFAChar> {
        let mut work_stack: Vec<CompileWork<'_>> = vec![CompileWork::Compile(regex)];
        let mut value_stack: Vec<NFAChar> = Vec::new();

        while let Some(work) = work_stack.pop() {
            match work {
                CompileWork::Compile(node) => {
                    self.process_compile_node(node, &mut work_stack, &mut value_stack)?;
                }
                CompileWork::DoConcatenate => {
                    let b = value_stack.pop().expect("second operand for concat");
                    let a = value_stack.pop().expect("first operand for concat");
                    value_stack.push(self.builder.concatenate(a, b));
                }
                CompileWork::DoAlternation => {
                    let b = value_stack.pop().expect("second operand for alt");
                    let a = value_stack.pop().expect("first operand for alt");
                    value_stack.push(self.builder.alternation(a, b));
                }
                CompileWork::DoStar => {
                    let inner = value_stack.pop().expect("operand for star");
                    value_stack.push(self.builder.kleene_star(inner));
                }
                CompileWork::DoPlus => {
                    let inner = value_stack.pop().expect("operand for plus");
                    value_stack.push(self.builder.kleene_plus(inner));
                }
                CompileWork::DoOptional => {
                    let inner = value_stack.pop().expect("operand for optional");
                    value_stack.push(self.builder.optional(inner));
                }
                CompileWork::DoRepeatExact(n) => {
                    let inner = value_stack.pop().expect("operand for repeat");
                    value_stack.push(self.builder.repeat_exact(inner, n));
                }
                CompileWork::DoRepeatRange(min, max) => {
                    let inner = value_stack.pop().expect("operand for repeat range");
                    value_stack.push(self.builder.repeat_range(inner, min, max));
                }
            }
        }

        value_stack.pop().ok_or_else(|| {
            ParseError::new(
                ParseErrorKind::InternalError("empty value stack after compilation".into()),
                Position::start(),
            )
        })
    }

    /// Process a single regex node for trampolined compilation.
    ///
    /// This method handles each regex variant by either:
    /// - Pushing the result directly to the value stack (base cases)
    /// - Pushing work items for children to the work stack (recursive cases)
    #[allow(deprecated)]
    fn process_compile_node<'a>(
        &mut self,
        node: &'a Regex,
        work_stack: &mut Vec<CompileWork<'a>>,
        value_stack: &mut Vec<NFAChar>,
    ) -> ParseResult<()> {
        match node {
            // === Base cases: push result directly ===
            Regex::Empty => value_stack.push(self.builder.epsilon()),
            Regex::Char(c) => value_stack.push(self.builder.single_char(*c)),
            Regex::CharClass(class) => value_stack.push(self.builder.char_class(class.clone())),
            Regex::Any => value_stack.push(self.builder.any_char()),
            Regex::WordBoundary => value_stack.push(self.builder.epsilon()),
            Regex::StartOfLine => value_stack.push(self.builder.start_of_line()),
            Regex::EndOfLine => value_stack.push(self.builder.end_of_line()),
            Regex::StartOfInput => value_stack.push(self.builder.start_of_input()),
            Regex::EndOfInput => value_stack.push(self.builder.end_of_input()),
            Regex::EndOfInputStrict => value_stack.push(self.builder.end_of_input_strict()),

            // === Binary operations: push operation, then children in reverse ===
            Regex::Concat(a, b) => {
                work_stack.push(CompileWork::DoConcatenate);
                work_stack.push(CompileWork::Compile(b));
                work_stack.push(CompileWork::Compile(a));
            }
            Regex::Alt(a, b) => {
                work_stack.push(CompileWork::DoAlternation);
                work_stack.push(CompileWork::Compile(b));
                work_stack.push(CompileWork::Compile(a));
            }

            // === Unary operations: push operation, then child ===
            Regex::Star(inner) => {
                work_stack.push(CompileWork::DoStar);
                work_stack.push(CompileWork::Compile(inner));
            }
            Regex::Plus(inner) => {
                work_stack.push(CompileWork::DoPlus);
                work_stack.push(CompileWork::Compile(inner));
            }
            Regex::Optional(inner) => {
                work_stack.push(CompileWork::DoOptional);
                work_stack.push(CompileWork::Compile(inner));
            }
            Regex::RepeatExact(inner, n) => {
                work_stack.push(CompileWork::DoRepeatExact(*n));
                work_stack.push(CompileWork::Compile(inner));
            }
            Regex::RepeatRange(inner, min, max) => {
                work_stack.push(CompileWork::DoRepeatRange(*min, *max));
                work_stack.push(CompileWork::Compile(inner));
            }

            // === Group wrappers: pass-through to inner ===
            Regex::CapturingGroup(_, inner)
            | Regex::NonCapturingGroup(inner)
            | Regex::NamedGroup(_, inner) => {
                work_stack.push(CompileWork::Compile(inner));
            }
            Regex::FlagsGroup {
                inner: Some(inner_regex),
                ..
            } => {
                work_stack.push(CompileWork::Compile(inner_regex));
            }
            Regex::FlagsGroup { inner: None, .. } => {
                value_stack.push(self.builder.epsilon());
            }
            Regex::RewriteRule { pattern, .. } => {
                work_stack.push(CompileWork::Compile(pattern));
            }

            // === Error case ===
            Regex::GroupRef(name) => {
                return Err(ParseError::new(
                    ParseErrorKind::InvalidGroupReference(format!(
                        "group reference (?&{}) cannot be compiled - subroutine expansion not yet implemented",
                        name
                    )),
                    Position::start(),
                ));
            }
        }
        Ok(())
    }

    /// Convert a regex to a literal string (for replacement).
    #[allow(deprecated)]
    fn regex_to_literal(&self, regex: &Regex) -> ParseResult<Vec<char>> {
        match regex {
            Regex::Empty => Ok(Vec::new()),
            Regex::Char(c) => Ok(vec![*c]),
            Regex::Concat(a, b) => {
                let mut chars = self.regex_to_literal(a)?;
                chars.extend(self.regex_to_literal(b)?);
                Ok(chars)
            }
            // All group types extract the inner literal
            Regex::CapturingGroup(_, inner)
            | Regex::NonCapturingGroup(inner)
            | Regex::NamedGroup(_, inner) => self.regex_to_literal(inner),
            // Flags group with inner pattern
            Regex::FlagsGroup {
                inner: Some(inner_regex),
                ..
            } => self.regex_to_literal(inner_regex),
            // Flags group without inner (inline flags)
            Regex::FlagsGroup { inner: None, .. } => Ok(Vec::new()),
            _ => Err(ParseError::new(
                ParseErrorKind::InvalidRewriteRule(
                    "replacement must be a literal string".to_string(),
                ),
                Position::start(),
            )),
        }
    }
}

impl Default for NFACompilerChar {
    fn default() -> Self {
        Self::new()
    }
}

/// Byte-level NFA compiler.
pub struct NFACompilerByte {
    builder: ThompsonBuilder,
}

impl NFACompilerByte {
    /// Create a new compiler.
    pub fn new() -> Self {
        Self {
            builder: ThompsonBuilder::new(),
        }
    }

    /// Compile a regex AST to an NFA.
    pub fn compile(&mut self, regex: &RegexByte) -> ParseResult<NFA> {
        let mut nfa = self.compile_regex(regex)?;
        // H9: Finalize CSR transition table
        nfa.finalize();
        Ok(nfa)
    }

    /// Compile a rewrite rule.
    pub fn compile_rewrite(&mut self, regex: &RegexByte) -> ParseResult<CompiledRewrite> {
        match regex {
            RegexByte::RewriteRule {
                pattern,
                replacement,
                context,
                weight,
            } => {
                let mut source = self.compile_regex(pattern)?;
                // H9: Finalize CSR transition table
                source.finalize();
                let replacement_bytes = self.regex_to_literal(replacement)?;

                let (left_context, right_context) = if let Some(ctx) = context {
                    (
                        ctx.left
                            .as_ref()
                            .map(|l| self.compile_context_expr(l))
                            .transpose()?,
                        ctx.right
                            .as_ref()
                            .map(|r| self.compile_context_expr(r))
                            .transpose()?,
                    )
                } else {
                    (None, None)
                };

                Ok(CompiledRewrite {
                    source,
                    replacement: replacement_bytes,
                    left_context,
                    right_context,
                    weight: *weight,
                })
            }
            _ => Err(ParseError::new(
                ParseErrorKind::InvalidRewriteRule("expected rewrite rule".to_string()),
                Position::start(),
            )),
        }
    }

    /// Compile a context expression to a context pattern.
    fn compile_context_expr(&mut self, expr: &ContextExprByte) -> ParseResult<ContextPattern> {
        match expr {
            ContextExprByte::Pattern(regex) => {
                let mut nfa = self.compile_regex(regex)?;
                // H9: Finalize CSR transition table
                nfa.finalize();
                Ok(ContextPattern::Nfa(nfa))
            }
            ContextExprByte::WordBoundary => Ok(ContextPattern::Boundary(BoundaryKind::WordStart)),
            ContextExprByte::And(a, b) => {
                let left = self.compile_context_expr(a)?;
                let right = self.compile_context_expr(b)?;
                Ok(ContextPattern::And(Box::new(left), Box::new(right)))
            }
            ContextExprByte::Or(a, b) => {
                let left = self.compile_context_expr(a)?;
                let right = self.compile_context_expr(b)?;
                Ok(ContextPattern::Or(Box::new(left), Box::new(right)))
            }
            ContextExprByte::Not(inner) => {
                let pattern = self.compile_context_expr(inner)?;
                Ok(ContextPattern::Not(Box::new(pattern)))
            }
        }
    }

    /// Compile a regex to NFA.
    #[allow(deprecated)]
    fn compile_regex(&mut self, regex: &RegexByte) -> ParseResult<NFA> {
        match regex {
            RegexByte::Empty => Ok(self.builder.epsilon()),
            RegexByte::Byte(b) => Ok(self.builder.single_byte(*b)),
            RegexByte::ByteClass(class) => Ok(self.builder.byte_class(class.clone())),
            RegexByte::Any => Ok(self.builder.any_byte()),
            RegexByte::Concat(a, b) => {
                let nfa_a = self.compile_regex(a)?;
                let nfa_b = self.compile_regex(b)?;
                Ok(self.builder.concatenate(nfa_a, nfa_b))
            }
            RegexByte::Alt(a, b) => {
                let nfa_a = self.compile_regex(a)?;
                let nfa_b = self.compile_regex(b)?;
                Ok(self.builder.alternation(nfa_a, nfa_b))
            }
            RegexByte::Star(inner) => {
                let nfa = self.compile_regex(inner)?;
                Ok(self.builder.kleene_star(nfa))
            }
            RegexByte::Plus(inner) => {
                let nfa = self.compile_regex(inner)?;
                Ok(self.builder.kleene_plus(nfa))
            }
            RegexByte::Optional(inner) => {
                let nfa = self.compile_regex(inner)?;
                Ok(self.builder.optional(nfa))
            }
            RegexByte::RepeatExact(inner, n) => {
                let nfa = self.compile_regex(inner)?;
                Ok(self.builder.repeat_exact(nfa, *n))
            }
            RegexByte::RepeatRange(inner, min, max) => {
                let nfa = self.compile_regex(inner)?;
                Ok(self.builder.repeat_range(nfa, *min, *max))
            }
            // Capturing group: (...)
            RegexByte::CapturingGroup(_, inner) => {
                self.compile_regex(inner)
            }
            // Non-capturing group: (?:...)
            RegexByte::NonCapturingGroup(inner) => {
                self.compile_regex(inner)
            }
            // Named group: (?<name>...)
            RegexByte::NamedGroup(_, inner) => {
                self.compile_regex(inner)
            }
            // Group reference: (?&name)
            RegexByte::GroupRef(name) => {
                Err(ParseError::new(
                    ParseErrorKind::InvalidGroupReference(format!(
                        "group reference (?&{}) cannot be compiled - subroutine expansion not yet implemented",
                        name
                    )),
                    Position::start(),
                ))
            }
            // Flags group: (?flags:...) or (?flags)
            RegexByte::FlagsGroup { inner, .. } => {
                match inner {
                    Some(inner_regex) => self.compile_regex(inner_regex),
                    None => Ok(self.builder.epsilon()),
                }
            }
            RegexByte::WordBoundary => {
                Ok(self.builder.epsilon())
            }
            // Anchor assertions (zero-width)
            RegexByte::StartOfLine => Ok(self.builder.start_of_line()),
            RegexByte::EndOfLine => Ok(self.builder.end_of_line()),
            RegexByte::StartOfInput => Ok(self.builder.start_of_input()),
            RegexByte::EndOfInput => Ok(self.builder.end_of_input()),
            RegexByte::EndOfInputStrict => Ok(self.builder.end_of_input_strict()),
            RegexByte::RewriteRule { pattern, .. } => {
                self.compile_regex(pattern)
            }
        }
    }

    /// Convert a regex to a literal byte string (for replacement).
    #[allow(deprecated)]
    fn regex_to_literal(&self, regex: &RegexByte) -> ParseResult<Vec<u8>> {
        match regex {
            RegexByte::Empty => Ok(Vec::new()),
            RegexByte::Byte(b) => Ok(vec![*b]),
            RegexByte::Concat(a, b) => {
                let mut bytes = self.regex_to_literal(a)?;
                bytes.extend(self.regex_to_literal(b)?);
                Ok(bytes)
            }
            // All group types extract the inner literal
            RegexByte::CapturingGroup(_, inner)
            | RegexByte::NonCapturingGroup(inner)
            | RegexByte::NamedGroup(_, inner) => self.regex_to_literal(inner),
            // Flags group with inner pattern
            RegexByte::FlagsGroup {
                inner: Some(inner_regex),
                ..
            } => self.regex_to_literal(inner_regex),
            // Flags group without inner (inline flags)
            RegexByte::FlagsGroup { inner: None, .. } => Ok(Vec::new()),
            _ => Err(ParseError::new(
                ParseErrorKind::InvalidRewriteRule(
                    "replacement must be a literal string".to_string(),
                ),
                Position::start(),
            )),
        }
    }
}

impl Default for NFACompilerByte {
    fn default() -> Self {
        Self::new()
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::phonetic::regex::{parse, parse_rule};

    #[test]
    fn test_compile_literal() {
        let regex = parse("phone").expect("test: parse phone");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(nfa.accepts("phone"));
        assert!(!nfa.accepts("fone"));
    }

    #[test]
    fn test_compile_alternation() {
        let regex = parse("ph|f").expect("test: parse ph|f");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(nfa.accepts("ph"));
        assert!(nfa.accepts("f"));
        assert!(!nfa.accepts("g"));
    }

    #[test]
    fn test_compile_group() {
        let regex = parse("(ph|f)one").expect("test: parse (ph|f)one");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(nfa.accepts("phone"));
        assert!(nfa.accepts("fone"));
        assert!(!nfa.accepts("bone"));
    }

    #[test]
    fn test_compile_star() {
        let regex = parse("a*").expect("test: parse a*");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(nfa.accepts(""));
        assert!(nfa.accepts("a"));
        assert!(nfa.accepts("aaa"));
        assert!(!nfa.accepts("b"));
    }

    #[test]
    fn test_compile_plus() {
        let regex = parse("a+").expect("test: parse a+");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(!nfa.accepts(""));
        assert!(nfa.accepts("a"));
        assert!(nfa.accepts("aaa"));
    }

    #[test]
    fn test_compile_optional() {
        let regex = parse("a?b").expect("test: parse a?b");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(nfa.accepts("b"));
        assert!(nfa.accepts("ab"));
        assert!(!nfa.accepts("aab"));
    }

    #[test]
    fn test_compile_char_class() {
        let regex = parse("[aeiou]").expect("test: parse [aeiou]");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(nfa.accepts("a"));
        assert!(nfa.accepts("e"));
        assert!(!nfa.accepts("b"));
    }

    #[test]
    fn test_compile_any() {
        let regex = parse("a.c").expect("test: parse a.c");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(nfa.accepts("abc"));
        assert!(nfa.accepts("axc"));
        assert!(!nfa.accepts("ac"));
    }

    #[test]
    fn test_compile_repeat_exact() {
        let regex = parse("a{3}").expect("test: parse a{3}");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(!nfa.accepts("aa"));
        assert!(nfa.accepts("aaa"));
        assert!(!nfa.accepts("aaaa"));
    }

    #[test]
    fn test_compile_repeat_range() {
        let regex = parse("a{2,4}").expect("test: parse a{2,4}");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(!nfa.accepts("a"));
        assert!(nfa.accepts("aa"));
        assert!(nfa.accepts("aaa"));
        assert!(nfa.accepts("aaaa"));
        assert!(!nfa.accepts("aaaaa"));
    }

    #[test]
    fn test_compile_rewrite_rule_simple() {
        let regex = parse_rule("ph -> f").expect("test: parse rule ph -> f");
        let rewrite = compile_rewrite(&regex).expect("test: compile_rewrite");
        assert!(rewrite.source.accepts("ph"));
        assert_eq!(rewrite.replacement, vec!['f']);
        assert!(rewrite.left_context.is_none());
        assert!(rewrite.right_context.is_none());
    }

    #[test]
    fn test_compile_rewrite_rule_with_context() {
        let regex = parse_rule("c -> s / _[ei]").expect("test: parse rule c -> s / _[ei]");
        let rewrite = compile_rewrite(&regex).expect("test: compile_rewrite");
        assert!(rewrite.source.accepts("c"));
        assert_eq!(rewrite.replacement, vec!['s']);
        assert!(rewrite.left_context.is_none());
        assert!(rewrite.right_context.is_some());
        let right = rewrite
            .right_context
            .expect("test: right_context is_some checked above");
        assert!(right.accepts("e"));
        assert!(right.accepts("i"));
        assert!(!right.accepts("a"));
    }

    #[test]
    fn test_compile_rewrite_rule_empty_replacement() {
        let regex = parse_rule("e -> / _#").expect("test: parse rule e -> / _#");
        let rewrite = compile_rewrite(&regex).expect("test: compile_rewrite");
        assert!(rewrite.source.accepts("e"));
        assert!(rewrite.replacement.is_empty());
    }

    #[test]
    fn test_compile_complex_pattern() {
        let regex = parse("(ph|f)one[s]?").expect("test: parse (ph|f)one[s]?");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(nfa.accepts("phone"));
        assert!(nfa.accepts("phones"));
        assert!(nfa.accepts("fone"));
        assert!(nfa.accepts("fones"));
        assert!(!nfa.accepts("bone"));
    }

    // Byte-level tests

    #[test]
    fn test_compile_bytes_literal() {
        let regex = crate::phonetic::regex::parse_bytes(b"phone").expect("test: parse_bytes phone");
        let nfa = compile_bytes(&regex).expect("test: compile_bytes nfa");
        assert!(nfa.accepts(b"phone"));
        assert!(!nfa.accepts(b"fone"));
    }

    #[test]
    fn test_compile_bytes_alternation() {
        let regex = crate::phonetic::regex::parse_bytes(b"ph|f").expect("test: parse_bytes ph|f");
        let nfa = compile_bytes(&regex).expect("test: compile_bytes nfa");
        assert!(nfa.accepts(b"ph"));
        assert!(nfa.accepts(b"f"));
    }

    #[test]
    fn test_compile_bytes_rewrite() {
        let regex = crate::phonetic::regex::parse_rule_bytes(b"ph -> f")
            .expect("test: parse_rule_bytes ph -> f");
        let rewrite = compile_rewrite_bytes(&regex).expect("test: compile_rewrite_bytes");
        assert!(rewrite.source.accepts(b"ph"));
        assert_eq!(rewrite.replacement, vec![b'f']);
    }

    // --- Anchor tests ---

    #[test]
    fn test_compile_start_of_line() {
        let regex = parse("^hello").expect("test: parse ^hello");
        let nfa = compile(&regex).expect("test: compile nfa");
        // NFA should have anchor transition followed by literal
        assert!(nfa.state_count() >= 3);
    }

    #[test]
    fn test_compile_end_of_line() {
        let regex = parse("hello$").expect("test: parse hello$");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(nfa.state_count() >= 3);
    }

    #[test]
    fn test_compile_anchored_pattern() {
        let regex = parse("^hello$").expect("test: parse ^hello$");
        let nfa = compile(&regex).expect("test: compile nfa");
        // start_of_line + hello (5 chars) + end_of_line
        assert!(nfa.state_count() >= 4);
    }

    #[test]
    fn test_compile_start_of_input() {
        let regex = parse(r"\Ahello").expect("test: parse \\Ahello");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(nfa.state_count() >= 3);
    }

    #[test]
    fn test_compile_end_of_input() {
        let regex = parse(r"hello\Z").expect("test: parse hello\\Z");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(nfa.state_count() >= 3);
    }

    #[test]
    fn test_compile_strict_end_of_input() {
        let regex = parse(r"hello\z").expect("test: parse hello\\z");
        let nfa = compile(&regex).expect("test: compile nfa");
        assert!(nfa.state_count() >= 3);
    }

    #[test]
    fn test_compile_multiline_flags() {
        // Pattern with multiline flag
        let regex = parse("(?m)^line$").expect("test: parse (?m)^line$");
        let nfa = compile(&regex).expect("test: compile nfa");
        // Should compile with anchor transitions
        assert!(nfa.state_count() >= 3);
    }

    #[test]
    fn test_trampolining_api() {
        // Test that the trampolining API works correctly
        let mut compiler = NFACompilerChar::new();
        assert!(!compiler.is_trampolining(), "default should be recursive");

        compiler.set_trampolining(true);
        assert!(
            compiler.is_trampolining(),
            "should be trampolining after set"
        );

        compiler.set_trampolining(false);
        assert!(
            !compiler.is_trampolining(),
            "should be recursive after unset"
        );

        let compiler2 = NFACompilerChar::new().with_trampolining();
        assert!(
            compiler2.is_trampolining(),
            "builder should enable trampolining"
        );
    }

    #[test]
    fn test_trampolined_matches_recursive_simple() {
        // Simple patterns should produce equivalent NFAs
        let patterns = ["a", "ab", "a|b", "a*", "a+", "a?", "[aeiou]"];

        for pattern in patterns {
            let regex = parse(pattern).expect("parse");

            // Compile with recursive (default)
            let nfa_recursive = compile(&regex).expect("recursive compile");

            // Compile with trampolining
            let mut compiler = NFACompilerChar::new().with_trampolining();
            let nfa_trampolined = compiler.compile(&regex).expect("trampolined compile");

            // Both should have the same state count
            assert_eq!(
                nfa_recursive.state_count(),
                nfa_trampolined.state_count(),
                "state count mismatch for pattern '{}'",
                pattern
            );

            // Both should accept/reject the same inputs
            let test_inputs = ["", "a", "b", "ab", "aaa", "aeiou"];
            for input in test_inputs {
                assert_eq!(
                    nfa_recursive.accepts(input),
                    nfa_trampolined.accepts(input),
                    "behavior mismatch for pattern '{}' on input '{}'",
                    pattern,
                    input
                );
            }
        }
    }

    #[test]
    fn test_trampolined_matches_recursive_complex() {
        // More complex patterns
        let patterns = [
            "(ab)+",
            "a{2,5}",
            "(?:a|b)*c",
            "(ph|f)one",
            "[a-z]+",
            "a{3}",
        ];

        for pattern in patterns {
            let regex = parse(pattern).expect("parse");

            let nfa_recursive = compile(&regex).expect("recursive compile");

            let mut compiler = NFACompilerChar::new().with_trampolining();
            let nfa_trampolined = compiler.compile(&regex).expect("trampolined compile");

            assert_eq!(
                nfa_recursive.state_count(),
                nfa_trampolined.state_count(),
                "state count mismatch for pattern '{}'",
                pattern
            );
        }
    }

    #[test]
    fn test_deep_nesting_with_trampolining() {
        // Build deeply nested pattern: ((((((((((a))))))))))
        // Note: The depth is limited by the parser's recursion, not the compiler.
        // The trampolining helps with compilation, but parsing is still recursive.
        let depth = 50; // Parser can handle this depth
        let mut pattern = "a".to_string();
        for _ in 0..depth {
            pattern = format!("({})", pattern);
        }

        let regex = parse(&pattern).expect("parse deeply nested pattern");

        // Compile with trampolining
        let mut compiler = NFACompilerChar::new().with_trampolining();
        let nfa = compiler
            .compile(&regex)
            .expect("trampolined compile of deep pattern");

        // Verify it works correctly
        assert!(nfa.accepts("a"), "should accept 'a'");
        assert!(!nfa.accepts("b"), "should reject 'b'");
        assert!(!nfa.accepts("aa"), "should reject 'aa'");

        // Also verify recursive compilation works at this depth
        let nfa_recursive = compile(&regex).expect("recursive compile");
        assert_eq!(nfa.state_count(), nfa_recursive.state_count());
    }

    #[test]
    fn test_deep_alternation_with_trampolining() {
        // Build deeply nested alternation: (((a|b)|c)|d)...
        // Parser depth limited
        let depth = 40;
        let mut pattern = "a".to_string();
        for i in 0..depth {
            let c = char::from(b'a' + ((i + 1) % 26) as u8);
            pattern = format!("({}|{})", pattern, c);
        }

        let regex = parse(&pattern).expect("parse deeply nested alternation");

        let mut compiler = NFACompilerChar::new().with_trampolining();
        let nfa = compiler
            .compile(&regex)
            .expect("trampolined compile of deep alternation");

        // Should accept any single letter
        assert!(nfa.accepts("a"));
        assert!(nfa.accepts("b"));
        assert!(nfa.accepts("c"));

        // Verify equivalence with recursive
        let nfa_recursive = compile(&regex).expect("recursive compile");
        assert_eq!(nfa.state_count(), nfa_recursive.state_count());
    }

    #[test]
    fn test_deep_concat_with_trampolining() {
        // Build deeply concatenated pattern via groups
        // Parser depth limited
        let depth = 100;
        let mut pattern = String::new();
        for _ in 0..depth {
            pattern.push_str("(?:a)");
        }

        let regex = parse(&pattern).expect("parse deeply concatenated pattern");

        let mut compiler = NFACompilerChar::new().with_trampolining();
        let nfa = compiler
            .compile(&regex)
            .expect("trampolined compile of deep concat");

        // Should accept exactly 'depth' 'a' characters
        let expected: String = std::iter::repeat('a').take(depth).collect();
        assert!(nfa.accepts(&expected), "should accept {} 'a' chars", depth);

        let too_short: String = std::iter::repeat('a').take(depth - 1).collect();
        assert!(
            !nfa.accepts(&too_short),
            "should reject {} 'a' chars",
            depth - 1
        );

        // Verify equivalence with recursive
        let nfa_recursive = compile(&regex).expect("recursive compile");
        assert_eq!(nfa.state_count(), nfa_recursive.state_count());
    }

    #[test]
    fn test_recursive_mode_still_works() {
        // Verify recursive mode (default) still works for normal patterns
        let patterns = ["hello", "world|earth", "a+b*c?", "[0-9]+"];

        for pattern in patterns {
            let regex = parse(pattern).expect("parse");

            // Default is recursive
            let nfa = compile(&regex).expect("recursive compile");
            assert!(nfa.state_count() > 0);
        }
    }
}