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
//! Types for tracking pieces of source code within a crate.
//!
//! The [`SourceMap`] tracks all the source code used within a single crate, mapping
//! from integer byte positions to the original source code location. Each bit
//! of source parsed during crate parsing (typically files, in-memory strings,
//! or various bits of macro expansion) cover a continuous range of bytes in the
//! `SourceMap` and are represented by [`SourceFile`]s. Byte positions are stored in
//! [`Span`] and used pervasively in the compiler. They are absolute positions
//! within the `SourceMap`, which upon request can be converted to line and column
//! information, source code snippets, etc.

pub use crate::hygiene::{ExpnData, ExpnKind};
pub use crate::*;

use rustc_data_structures::fx::FxHashMap;
use rustc_data_structures::stable_hasher::StableHasher;
use rustc_data_structures::sync::{AtomicU32, Lrc, MappedReadGuard, ReadGuard, RwLock};
use std::hash::Hash;
use std::path::{Path, PathBuf};
use std::sync::atomic::Ordering;
use std::{clone::Clone, cmp};
use std::{convert::TryFrom, unreachable};

use std::fs;
use std::io;

#[cfg(test)]
mod tests;

/// Returns the span itself if it doesn't come from a macro expansion,
/// otherwise return the call site span up to the `enclosing_sp` by
/// following the `expn_data` chain.
pub fn original_sp(sp: Span, enclosing_sp: Span) -> Span {
    let expn_data1 = sp.ctxt().outer_expn_data();
    let expn_data2 = enclosing_sp.ctxt().outer_expn_data();
    if expn_data1.is_root() || !expn_data2.is_root() && expn_data1.call_site == expn_data2.call_site
    {
        sp
    } else {
        original_sp(expn_data1.call_site, enclosing_sp)
    }
}

pub mod monotonic {
    use std::ops::{Deref, DerefMut};

    /// A `MonotonicVec` is a `Vec` which can only be grown.
    /// Once inserted, an element can never be removed or swapped,
    /// guaranteeing that any indices into a `MonotonicVec` are stable
    // This is declared in its own module to ensure that the private
    // field is inaccessible
    pub struct MonotonicVec<T>(Vec<T>);
    impl<T> MonotonicVec<T> {
        pub fn new(val: Vec<T>) -> MonotonicVec<T> {
            MonotonicVec(val)
        }

        pub fn push(&mut self, val: T) {
            self.0.push(val);
        }
    }

    impl<T> Default for MonotonicVec<T> {
        fn default() -> Self {
            MonotonicVec::new(vec![])
        }
    }

    impl<T> Deref for MonotonicVec<T> {
        type Target = Vec<T>;
        fn deref(&self) -> &Self::Target {
            &self.0
        }
    }

    impl<T> !DerefMut for MonotonicVec<T> {}
}

#[derive(Clone, Encodable, Decodable, Debug, Copy, HashStable_Generic)]
pub struct Spanned<T> {
    pub node: T,
    pub span: Span,
}

pub fn respan<T>(sp: Span, t: T) -> Spanned<T> {
    Spanned { node: t, span: sp }
}

pub fn dummy_spanned<T>(t: T) -> Spanned<T> {
    respan(DUMMY_SP, t)
}

// _____________________________________________________________________________
// SourceFile, MultiByteChar, FileName, FileLines
//

/// An abstraction over the fs operations used by the Parser.
pub trait FileLoader {
    /// Query the existence of a file.
    fn file_exists(&self, path: &Path) -> bool;

    /// Read the contents of a UTF-8 file into memory.
    fn read_file(&self, path: &Path) -> io::Result<String>;
}

/// A FileLoader that uses std::fs to load real files.
pub struct RealFileLoader;

impl FileLoader for RealFileLoader {
    fn file_exists(&self, path: &Path) -> bool {
        path.exists()
    }

    fn read_file(&self, path: &Path) -> io::Result<String> {
        fs::read_to_string(path)
    }
}

/// This is a [SourceFile] identifier that is used to correlate source files between
/// subsequent compilation sessions (which is something we need to do during
/// incremental compilation).
///
/// The [StableSourceFileId] also contains the CrateNum of the crate the source
/// file was originally parsed for. This way we get two separate entries in
/// the [SourceMap] if the same file is part of both the local and an upstream
/// crate. Trying to only have one entry for both cases is problematic because
/// at the point where we discover that there's a local use of the file in
/// addition to the upstream one, we might already have made decisions based on
/// the assumption that it's an upstream file. Treating the two files as
/// different has no real downsides.
#[derive(Copy, Clone, PartialEq, Eq, Hash, Encodable, Decodable, Debug)]
pub struct StableSourceFileId {
    // A hash of the source file's FileName. This is hash so that it's size
    // is more predictable than if we included the actual FileName value.
    pub file_name_hash: u64,

    // The CrateNum of the crate this source file was originally parsed for.
    // We cannot include this information in the hash because at the time
    // of hashing we don't have the context to map from the CrateNum's numeric
    // value to a StableCrateId.
    pub cnum: CrateNum,
}

// FIXME: we need a more globally consistent approach to the problem solved by
// StableSourceFileId, perhaps built atop source_file.name_hash.
impl StableSourceFileId {
    pub fn new(source_file: &SourceFile) -> StableSourceFileId {
        StableSourceFileId::new_from_name(&source_file.name, source_file.cnum)
    }

    fn new_from_name(name: &FileName, cnum: CrateNum) -> StableSourceFileId {
        let mut hasher = StableHasher::new();
        name.hash(&mut hasher);
        StableSourceFileId { file_name_hash: hasher.finish(), cnum }
    }
}

// _____________________________________________________________________________
// SourceMap
//

#[derive(Default)]
pub(super) struct SourceMapFiles {
    source_files: monotonic::MonotonicVec<Lrc<SourceFile>>,
    stable_id_to_source_file: FxHashMap<StableSourceFileId, Lrc<SourceFile>>,
}

pub struct SourceMap {
    /// The address space below this value is currently used by the files in the source map.
    used_address_space: AtomicU32,

    files: RwLock<SourceMapFiles>,
    file_loader: Box<dyn FileLoader + Sync + Send>,
    // This is used to apply the file path remapping as specified via
    // `--remap-path-prefix` to all `SourceFile`s allocated within this `SourceMap`.
    path_mapping: FilePathMapping,

    /// The algorithm used for hashing the contents of each source file.
    hash_kind: SourceFileHashAlgorithm,
}

impl SourceMap {
    pub fn new(path_mapping: FilePathMapping) -> SourceMap {
        Self::with_file_loader_and_hash_kind(
            Box::new(RealFileLoader),
            path_mapping,
            SourceFileHashAlgorithm::Md5,
        )
    }

    pub fn with_file_loader_and_hash_kind(
        file_loader: Box<dyn FileLoader + Sync + Send>,
        path_mapping: FilePathMapping,
        hash_kind: SourceFileHashAlgorithm,
    ) -> SourceMap {
        SourceMap {
            used_address_space: AtomicU32::new(0),
            files: Default::default(),
            file_loader,
            path_mapping,
            hash_kind,
        }
    }

    pub fn path_mapping(&self) -> &FilePathMapping {
        &self.path_mapping
    }

    pub fn file_exists(&self, path: &Path) -> bool {
        self.file_loader.file_exists(path)
    }

    pub fn load_file(&self, path: &Path) -> io::Result<Lrc<SourceFile>> {
        let src = self.file_loader.read_file(path)?;
        let filename = path.to_owned().into();
        Ok(self.new_source_file(filename, src))
    }

    /// Loads source file as a binary blob.
    ///
    /// Unlike `load_file`, guarantees that no normalization like BOM-removal
    /// takes place.
    pub fn load_binary_file(&self, path: &Path) -> io::Result<Vec<u8>> {
        // Ideally, this should use `self.file_loader`, but it can't
        // deal with binary files yet.
        let bytes = fs::read(path)?;

        // We need to add file to the `SourceMap`, so that it is present
        // in dep-info. There's also an edge case that file might be both
        // loaded as a binary via `include_bytes!` and as proper `SourceFile`
        // via `mod`, so we try to use real file contents and not just an
        // empty string.
        let text = std::str::from_utf8(&bytes).unwrap_or("").to_string();
        self.new_source_file(path.to_owned().into(), text);
        Ok(bytes)
    }

    // By returning a `MonotonicVec`, we ensure that consumers cannot invalidate
    // any existing indices pointing into `files`.
    pub fn files(&self) -> MappedReadGuard<'_, monotonic::MonotonicVec<Lrc<SourceFile>>> {
        ReadGuard::map(self.files.borrow(), |files| &files.source_files)
    }

    pub fn source_file_by_stable_id(
        &self,
        stable_id: StableSourceFileId,
    ) -> Option<Lrc<SourceFile>> {
        self.files.borrow().stable_id_to_source_file.get(&stable_id).cloned()
    }

    fn allocate_address_space(&self, size: usize) -> Result<usize, OffsetOverflowError> {
        let size = u32::try_from(size).map_err(|_| OffsetOverflowError)?;

        loop {
            let current = self.used_address_space.load(Ordering::Relaxed);
            let next = current
                .checked_add(size)
                // Add one so there is some space between files. This lets us distinguish
                // positions in the `SourceMap`, even in the presence of zero-length files.
                .and_then(|next| next.checked_add(1))
                .ok_or(OffsetOverflowError)?;

            if self
                .used_address_space
                .compare_exchange(current, next, Ordering::Relaxed, Ordering::Relaxed)
                .is_ok()
            {
                return Ok(usize::try_from(current).unwrap());
            }
        }
    }

    /// Creates a new `SourceFile`.
    /// If a file already exists in the `SourceMap` with the same ID, that file is returned
    /// unmodified.
    pub fn new_source_file(&self, filename: FileName, src: String) -> Lrc<SourceFile> {
        self.try_new_source_file(filename, src).unwrap_or_else(|OffsetOverflowError| {
            eprintln!("fatal error: rustc does not support files larger than 4GB");
            crate::fatal_error::FatalError.raise()
        })
    }

    fn try_new_source_file(
        &self,
        filename: FileName,
        src: String,
    ) -> Result<Lrc<SourceFile>, OffsetOverflowError> {
        // Note that filename may not be a valid path, eg it may be `<anon>` etc,
        // but this is okay because the directory determined by `path.pop()` will
        // be empty, so the working directory will be used.
        let (filename, _) = self.path_mapping.map_filename_prefix(&filename);

        let file_id = StableSourceFileId::new_from_name(&filename, LOCAL_CRATE);

        let lrc_sf = match self.source_file_by_stable_id(file_id) {
            Some(lrc_sf) => lrc_sf,
            None => {
                let start_pos = self.allocate_address_space(src.len())?;

                let source_file = Lrc::new(SourceFile::new(
                    filename,
                    src,
                    Pos::from_usize(start_pos),
                    self.hash_kind,
                ));

                // Let's make sure the file_id we generated above actually matches
                // the ID we generate for the SourceFile we just created.
                debug_assert_eq!(StableSourceFileId::new(&source_file), file_id);

                let mut files = self.files.borrow_mut();

                files.source_files.push(source_file.clone());
                files.stable_id_to_source_file.insert(file_id, source_file.clone());

                source_file
            }
        };
        Ok(lrc_sf)
    }

    /// Allocates a new `SourceFile` representing a source file from an external
    /// crate. The source code of such an "imported `SourceFile`" is not available,
    /// but we still know enough to generate accurate debuginfo location
    /// information for things inlined from other crates.
    pub fn new_imported_source_file(
        &self,
        filename: FileName,
        src_hash: SourceFileHash,
        name_hash: u128,
        source_len: usize,
        cnum: CrateNum,
        file_local_lines: Lock<SourceFileLines>,
        mut file_local_multibyte_chars: Vec<MultiByteChar>,
        mut file_local_non_narrow_chars: Vec<NonNarrowChar>,
        mut file_local_normalized_pos: Vec<NormalizedPos>,
        original_start_pos: BytePos,
        metadata_index: u32,
    ) -> Lrc<SourceFile> {
        let start_pos = self
            .allocate_address_space(source_len)
            .expect("not enough address space for imported source file");

        let end_pos = Pos::from_usize(start_pos + source_len);
        let start_pos = Pos::from_usize(start_pos);

        // Translate these positions into the new global frame of reference,
        // now that the offset of the SourceFile is known.
        //
        // These are all unsigned values. `original_start_pos` may be larger or
        // smaller than `start_pos`, but `pos` is always larger than both.
        // Therefore, `(pos - original_start_pos) + start_pos` won't overflow
        // but `start_pos - original_start_pos` might. So we use the former
        // form rather than pre-computing the offset into a local variable. The
        // compiler backend can optimize away the repeated computations in a
        // way that won't trigger overflow checks.
        match &mut *file_local_lines.borrow_mut() {
            SourceFileLines::Lines(lines) => {
                for pos in lines {
                    *pos = (*pos - original_start_pos) + start_pos;
                }
            }
            SourceFileLines::Diffs(SourceFileDiffs { line_start, .. }) => {
                *line_start = (*line_start - original_start_pos) + start_pos;
            }
        }
        for mbc in &mut file_local_multibyte_chars {
            mbc.pos = (mbc.pos - original_start_pos) + start_pos;
        }
        for swc in &mut file_local_non_narrow_chars {
            *swc = (*swc - original_start_pos) + start_pos;
        }
        for nc in &mut file_local_normalized_pos {
            nc.pos = (nc.pos - original_start_pos) + start_pos;
        }

        let source_file = Lrc::new(SourceFile {
            name: filename,
            src: None,
            src_hash,
            external_src: Lock::new(ExternalSource::Foreign {
                kind: ExternalSourceKind::AbsentOk,
                metadata_index,
            }),
            start_pos,
            end_pos,
            lines: file_local_lines,
            multibyte_chars: file_local_multibyte_chars,
            non_narrow_chars: file_local_non_narrow_chars,
            normalized_pos: file_local_normalized_pos,
            name_hash,
            cnum,
        });

        let mut files = self.files.borrow_mut();

        files.source_files.push(source_file.clone());
        files
            .stable_id_to_source_file
            .insert(StableSourceFileId::new(&source_file), source_file.clone());

        source_file
    }

    // If there is a doctest offset, applies it to the line.
    pub fn doctest_offset_line(&self, file: &FileName, orig: usize) -> usize {
        match file {
            FileName::DocTest(_, offset) => {
                if *offset < 0 {
                    orig - (-(*offset)) as usize
                } else {
                    orig + *offset as usize
                }
            }
            _ => orig,
        }
    }

    /// Return the SourceFile that contains the given `BytePos`
    pub fn lookup_source_file(&self, pos: BytePos) -> Lrc<SourceFile> {
        let idx = self.lookup_source_file_idx(pos);
        (*self.files.borrow().source_files)[idx].clone()
    }

    /// Looks up source information about a `BytePos`.
    pub fn lookup_char_pos(&self, pos: BytePos) -> Loc {
        let sf = self.lookup_source_file(pos);
        let (line, col, col_display) = sf.lookup_file_pos_with_col_display(pos);
        Loc { file: sf, line, col, col_display }
    }

    // If the corresponding `SourceFile` is empty, does not return a line number.
    pub fn lookup_line(&self, pos: BytePos) -> Result<SourceFileAndLine, Lrc<SourceFile>> {
        let f = self.lookup_source_file(pos);

        match f.lookup_line(pos) {
            Some(line) => Ok(SourceFileAndLine { sf: f, line }),
            None => Err(f),
        }
    }

    fn span_to_string(&self, sp: Span, filename_display_pref: FileNameDisplayPreference) -> String {
        if self.files.borrow().source_files.is_empty() || sp.is_dummy() {
            return "no-location".to_string();
        }

        let lo = self.lookup_char_pos(sp.lo());
        let hi = self.lookup_char_pos(sp.hi());
        format!(
            "{}:{}:{}: {}:{}",
            lo.file.name.display(filename_display_pref),
            lo.line,
            lo.col.to_usize() + 1,
            hi.line,
            hi.col.to_usize() + 1,
        )
    }

    /// Format the span location suitable for embedding in build artifacts
    pub fn span_to_embeddable_string(&self, sp: Span) -> String {
        self.span_to_string(sp, FileNameDisplayPreference::Remapped)
    }

    /// Format the span location suitable for pretty printing anotations with relative line numbers
    pub fn span_to_relative_line_string(&self, sp: Span, relative_to: Span) -> String {
        if self.files.borrow().source_files.is_empty() || sp.is_dummy() || relative_to.is_dummy() {
            return "no-location".to_string();
        }

        let lo = self.lookup_char_pos(sp.lo());
        let hi = self.lookup_char_pos(sp.hi());
        let offset = self.lookup_char_pos(relative_to.lo());

        if lo.file.name != offset.file.name || !relative_to.contains(sp) {
            return self.span_to_embeddable_string(sp);
        }

        let lo_line = lo.line.saturating_sub(offset.line);
        let hi_line = hi.line.saturating_sub(offset.line);

        format!(
            "{}:+{}:{}: +{}:{}",
            lo.file.name.display(FileNameDisplayPreference::Remapped),
            lo_line,
            lo.col.to_usize() + 1,
            hi_line,
            hi.col.to_usize() + 1,
        )
    }

    /// Format the span location to be printed in diagnostics. Must not be emitted
    /// to build artifacts as this may leak local file paths. Use span_to_embeddable_string
    /// for string suitable for embedding.
    pub fn span_to_diagnostic_string(&self, sp: Span) -> String {
        self.span_to_string(sp, self.path_mapping.filename_display_for_diagnostics)
    }

    pub fn span_to_filename(&self, sp: Span) -> FileName {
        self.lookup_char_pos(sp.lo()).file.name.clone()
    }

    pub fn filename_for_diagnostics<'a>(&self, filename: &'a FileName) -> FileNameDisplay<'a> {
        filename.display(self.path_mapping.filename_display_for_diagnostics)
    }

    pub fn is_multiline(&self, sp: Span) -> bool {
        let lo = self.lookup_source_file_idx(sp.lo());
        let hi = self.lookup_source_file_idx(sp.hi());
        if lo != hi {
            return true;
        }
        let f = (*self.files.borrow().source_files)[lo].clone();
        f.lookup_line(sp.lo()) != f.lookup_line(sp.hi())
    }

    #[instrument(skip(self), level = "trace")]
    pub fn is_valid_span(&self, sp: Span) -> Result<(Loc, Loc), SpanLinesError> {
        let lo = self.lookup_char_pos(sp.lo());
        trace!(?lo);
        let hi = self.lookup_char_pos(sp.hi());
        trace!(?hi);
        if lo.file.start_pos != hi.file.start_pos {
            return Err(SpanLinesError::DistinctSources(DistinctSources {
                begin: (lo.file.name.clone(), lo.file.start_pos),
                end: (hi.file.name.clone(), hi.file.start_pos),
            }));
        }
        Ok((lo, hi))
    }

    pub fn is_line_before_span_empty(&self, sp: Span) -> bool {
        match self.span_to_prev_source(sp) {
            Ok(s) => s.rsplit_once('\n').unwrap_or(("", &s)).1.trim_start().is_empty(),
            Err(_) => false,
        }
    }

    pub fn span_to_lines(&self, sp: Span) -> FileLinesResult {
        debug!("span_to_lines(sp={:?})", sp);
        let (lo, hi) = self.is_valid_span(sp)?;
        assert!(hi.line >= lo.line);

        if sp.is_dummy() {
            return Ok(FileLines { file: lo.file, lines: Vec::new() });
        }

        let mut lines = Vec::with_capacity(hi.line - lo.line + 1);

        // The span starts partway through the first line,
        // but after that it starts from offset 0.
        let mut start_col = lo.col;

        // For every line but the last, it extends from `start_col`
        // and to the end of the line. Be careful because the line
        // numbers in Loc are 1-based, so we subtract 1 to get 0-based
        // lines.
        //
        // FIXME: now that we handle DUMMY_SP up above, we should consider
        // asserting that the line numbers here are all indeed 1-based.
        let hi_line = hi.line.saturating_sub(1);
        for line_index in lo.line.saturating_sub(1)..hi_line {
            let line_len = lo.file.get_line(line_index).map_or(0, |s| s.chars().count());
            lines.push(LineInfo { line_index, start_col, end_col: CharPos::from_usize(line_len) });
            start_col = CharPos::from_usize(0);
        }

        // For the last line, it extends from `start_col` to `hi.col`:
        lines.push(LineInfo { line_index: hi_line, start_col, end_col: hi.col });

        Ok(FileLines { file: lo.file, lines })
    }

    /// Extracts the source surrounding the given `Span` using the `extract_source` function. The
    /// extract function takes three arguments: a string slice containing the source, an index in
    /// the slice for the beginning of the span and an index in the slice for the end of the span.
    fn span_to_source<F, T>(&self, sp: Span, extract_source: F) -> Result<T, SpanSnippetError>
    where
        F: Fn(&str, usize, usize) -> Result<T, SpanSnippetError>,
    {
        let local_begin = self.lookup_byte_offset(sp.lo());
        let local_end = self.lookup_byte_offset(sp.hi());

        if local_begin.sf.start_pos != local_end.sf.start_pos {
            Err(SpanSnippetError::DistinctSources(DistinctSources {
                begin: (local_begin.sf.name.clone(), local_begin.sf.start_pos),
                end: (local_end.sf.name.clone(), local_end.sf.start_pos),
            }))
        } else {
            self.ensure_source_file_source_present(local_begin.sf.clone());

            let start_index = local_begin.pos.to_usize();
            let end_index = local_end.pos.to_usize();
            let source_len = (local_begin.sf.end_pos - local_begin.sf.start_pos).to_usize();

            if start_index > end_index || end_index > source_len {
                return Err(SpanSnippetError::MalformedForSourcemap(MalformedSourceMapPositions {
                    name: local_begin.sf.name.clone(),
                    source_len,
                    begin_pos: local_begin.pos,
                    end_pos: local_end.pos,
                }));
            }

            if let Some(ref src) = local_begin.sf.src {
                extract_source(src, start_index, end_index)
            } else if let Some(src) = local_begin.sf.external_src.borrow().get_source() {
                extract_source(src, start_index, end_index)
            } else {
                Err(SpanSnippetError::SourceNotAvailable { filename: local_begin.sf.name.clone() })
            }
        }
    }

    pub fn is_span_accessible(&self, sp: Span) -> bool {
        self.span_to_source(sp, |src, start_index, end_index| {
            Ok(src.get(start_index..end_index).is_some())
        })
        .map_or(false, |is_accessible| is_accessible)
    }

    /// Returns the source snippet as `String` corresponding to the given `Span`.
    pub fn span_to_snippet(&self, sp: Span) -> Result<String, SpanSnippetError> {
        self.span_to_source(sp, |src, start_index, end_index| {
            src.get(start_index..end_index)
                .map(|s| s.to_string())
                .ok_or(SpanSnippetError::IllFormedSpan(sp))
        })
    }

    pub fn span_to_margin(&self, sp: Span) -> Option<usize> {
        Some(self.indentation_before(sp)?.len())
    }

    pub fn indentation_before(&self, sp: Span) -> Option<String> {
        self.span_to_source(sp, |src, start_index, _| {
            let before = &src[..start_index];
            let last_line = before.rsplit_once('\n').map_or(before, |(_, last)| last);
            Ok(last_line
                .split_once(|c: char| !c.is_whitespace())
                .map_or(last_line, |(indent, _)| indent)
                .to_string())
        })
        .ok()
    }

    /// Returns the source snippet as `String` before the given `Span`.
    pub fn span_to_prev_source(&self, sp: Span) -> Result<String, SpanSnippetError> {
        self.span_to_source(sp, |src, start_index, _| {
            src.get(..start_index).map(|s| s.to_string()).ok_or(SpanSnippetError::IllFormedSpan(sp))
        })
    }

    /// Extends the given `Span` to just after the previous occurrence of `c`. Return the same span
    /// if no character could be found or if an error occurred while retrieving the code snippet.
    pub fn span_extend_to_prev_char(&self, sp: Span, c: char, accept_newlines: bool) -> Span {
        if let Ok(prev_source) = self.span_to_prev_source(sp) {
            let prev_source = prev_source.rsplit(c).next().unwrap_or("");
            if !prev_source.is_empty() && (accept_newlines || !prev_source.contains('\n')) {
                return sp.with_lo(BytePos(sp.lo().0 - prev_source.len() as u32));
            }
        }

        sp
    }

    /// Extends the given `Span` to just after the previous occurrence of `pat` when surrounded by
    /// whitespace. Returns None if the pattern could not be found or if an error occurred while
    /// retrieving the code snippet.
    pub fn span_extend_to_prev_str(
        &self,
        sp: Span,
        pat: &str,
        accept_newlines: bool,
        include_whitespace: bool,
    ) -> Option<Span> {
        // assure that the pattern is delimited, to avoid the following
        //     fn my_fn()
        //           ^^^^ returned span without the check
        //     ---------- correct span
        let prev_source = self.span_to_prev_source(sp).ok()?;
        for ws in &[" ", "\t", "\n"] {
            let pat = pat.to_owned() + ws;
            if let Some(pat_pos) = prev_source.rfind(&pat) {
                let just_after_pat_pos = pat_pos + pat.len() - 1;
                let just_after_pat_plus_ws = if include_whitespace {
                    just_after_pat_pos
                        + prev_source[just_after_pat_pos..]
                            .find(|c: char| !c.is_whitespace())
                            .unwrap_or(0)
                } else {
                    just_after_pat_pos
                };
                let len = prev_source.len() - just_after_pat_plus_ws;
                let prev_source = &prev_source[just_after_pat_plus_ws..];
                if accept_newlines || !prev_source.trim_start().contains('\n') {
                    return Some(sp.with_lo(BytePos(sp.lo().0 - len as u32)));
                }
            }
        }

        None
    }

    /// Returns the source snippet as `String` after the given `Span`.
    pub fn span_to_next_source(&self, sp: Span) -> Result<String, SpanSnippetError> {
        self.span_to_source(sp, |src, _, end_index| {
            src.get(end_index..).map(|s| s.to_string()).ok_or(SpanSnippetError::IllFormedSpan(sp))
        })
    }

    /// Extends the given `Span` while the next character matches the predicate
    pub fn span_extend_while(
        &self,
        span: Span,
        f: impl Fn(char) -> bool,
    ) -> Result<Span, SpanSnippetError> {
        self.span_to_source(span, |s, _start, end| {
            let n = s[end..].char_indices().find(|&(_, c)| !f(c)).map_or(s.len() - end, |(i, _)| i);
            Ok(span.with_hi(span.hi() + BytePos(n as u32)))
        })
    }

    /// Extends the given `Span` to just before the next occurrence of `c`.
    pub fn span_extend_to_next_char(&self, sp: Span, c: char, accept_newlines: bool) -> Span {
        if let Ok(next_source) = self.span_to_next_source(sp) {
            let next_source = next_source.split(c).next().unwrap_or("");
            if !next_source.is_empty() && (accept_newlines || !next_source.contains('\n')) {
                return sp.with_hi(BytePos(sp.hi().0 + next_source.len() as u32));
            }
        }

        sp
    }

    /// Extends the given `Span` to contain the entire line it is on.
    pub fn span_extend_to_line(&self, sp: Span) -> Span {
        self.span_extend_to_prev_char(self.span_extend_to_next_char(sp, '\n', true), '\n', true)
    }

    /// Given a `Span`, tries to get a shorter span ending before the first occurrence of `char`
    /// `c`.
    pub fn span_until_char(&self, sp: Span, c: char) -> Span {
        match self.span_to_snippet(sp) {
            Ok(snippet) => {
                let snippet = snippet.split(c).next().unwrap_or("").trim_end();
                if !snippet.is_empty() && !snippet.contains('\n') {
                    sp.with_hi(BytePos(sp.lo().0 + snippet.len() as u32))
                } else {
                    sp
                }
            }
            _ => sp,
        }
    }

    /// Given a `Span`, tries to get a shorter span ending just after the first occurrence of `char`
    /// `c`.
    pub fn span_through_char(&self, sp: Span, c: char) -> Span {
        if let Ok(snippet) = self.span_to_snippet(sp) {
            if let Some(offset) = snippet.find(c) {
                return sp.with_hi(BytePos(sp.lo().0 + (offset + c.len_utf8()) as u32));
            }
        }
        sp
    }

    /// Given a `Span`, gets a new `Span` covering the first token and all its trailing whitespace
    /// or the original `Span`.
    ///
    /// If `sp` points to `"let mut x"`, then a span pointing at `"let "` will be returned.
    pub fn span_until_non_whitespace(&self, sp: Span) -> Span {
        let mut whitespace_found = false;

        self.span_take_while(sp, |c| {
            if !whitespace_found && c.is_whitespace() {
                whitespace_found = true;
            }

            !whitespace_found || c.is_whitespace()
        })
    }

    /// Given a `Span`, gets a new `Span` covering the first token without its trailing whitespace
    /// or the original `Span` in case of error.
    ///
    /// If `sp` points to `"let mut x"`, then a span pointing at `"let"` will be returned.
    pub fn span_until_whitespace(&self, sp: Span) -> Span {
        self.span_take_while(sp, |c| !c.is_whitespace())
    }

    /// Given a `Span`, gets a shorter one until `predicate` yields `false`.
    pub fn span_take_while<P>(&self, sp: Span, predicate: P) -> Span
    where
        P: for<'r> FnMut(&'r char) -> bool,
    {
        if let Ok(snippet) = self.span_to_snippet(sp) {
            let offset = snippet.chars().take_while(predicate).map(|c| c.len_utf8()).sum::<usize>();

            sp.with_hi(BytePos(sp.lo().0 + (offset as u32)))
        } else {
            sp
        }
    }

    /// Given a `Span`, return a span ending in the closest `{`. This is useful when you have a
    /// `Span` enclosing a whole item but we need to point at only the head (usually the first
    /// line) of that item.
    ///
    /// *Only suitable for diagnostics.*
    pub fn guess_head_span(&self, sp: Span) -> Span {
        // FIXME: extend the AST items to have a head span, or replace callers with pointing at
        // the item's ident when appropriate.
        self.span_until_char(sp, '{')
    }

    /// Returns a new span representing just the first character of the given span.
    pub fn start_point(&self, sp: Span) -> Span {
        let width = {
            let sp = sp.data();
            let local_begin = self.lookup_byte_offset(sp.lo);
            let start_index = local_begin.pos.to_usize();
            let src = local_begin.sf.external_src.borrow();

            let snippet = if let Some(ref src) = local_begin.sf.src {
                Some(&src[start_index..])
            } else if let Some(src) = src.get_source() {
                Some(&src[start_index..])
            } else {
                None
            };

            match snippet {
                None => 1,
                Some(snippet) => match snippet.chars().next() {
                    None => 1,
                    Some(c) => c.len_utf8(),
                },
            }
        };

        sp.with_hi(BytePos(sp.lo().0 + width as u32))
    }

    /// Returns a new span representing just the last character of this span.
    pub fn end_point(&self, sp: Span) -> Span {
        let pos = sp.hi().0;

        let width = self.find_width_of_character_at_span(sp, false);
        let corrected_end_position = pos.checked_sub(width).unwrap_or(pos);

        let end_point = BytePos(cmp::max(corrected_end_position, sp.lo().0));
        sp.with_lo(end_point)
    }

    /// Returns a new span representing the next character after the end-point of this span.
    pub fn next_point(&self, sp: Span) -> Span {
        if sp.is_dummy() {
            return sp;
        }
        let start_of_next_point = sp.hi().0;

        let width = self.find_width_of_character_at_span(sp.shrink_to_hi(), true);
        // If the width is 1, then the next span should point to the same `lo` and `hi`. However,
        // in the case of a multibyte character, where the width != 1, the next span should
        // span multiple bytes to include the whole character.
        let end_of_next_point =
            start_of_next_point.checked_add(width - 1).unwrap_or(start_of_next_point);

        let end_of_next_point = BytePos(cmp::max(sp.lo().0 + 1, end_of_next_point));
        Span::new(BytePos(start_of_next_point), end_of_next_point, sp.ctxt(), None)
    }

    /// Finds the width of the character, either before or after the end of provided span,
    /// depending on the `forwards` parameter.
    fn find_width_of_character_at_span(&self, sp: Span, forwards: bool) -> u32 {
        let sp = sp.data();
        if sp.lo == sp.hi {
            debug!("find_width_of_character_at_span: early return empty span");
            return 1;
        }

        let local_begin = self.lookup_byte_offset(sp.lo);
        let local_end = self.lookup_byte_offset(sp.hi);
        debug!(
            "find_width_of_character_at_span: local_begin=`{:?}`, local_end=`{:?}`",
            local_begin, local_end
        );

        if local_begin.sf.start_pos != local_end.sf.start_pos {
            debug!("find_width_of_character_at_span: begin and end are in different files");
            return 1;
        }

        let start_index = local_begin.pos.to_usize();
        let end_index = local_end.pos.to_usize();
        debug!(
            "find_width_of_character_at_span: start_index=`{:?}`, end_index=`{:?}`",
            start_index, end_index
        );

        // Disregard indexes that are at the start or end of their spans, they can't fit bigger
        // characters.
        if (!forwards && end_index == usize::MIN) || (forwards && start_index == usize::MAX) {
            debug!("find_width_of_character_at_span: start or end of span, cannot be multibyte");
            return 1;
        }

        let source_len = (local_begin.sf.end_pos - local_begin.sf.start_pos).to_usize();
        debug!("find_width_of_character_at_span: source_len=`{:?}`", source_len);
        // Ensure indexes are also not malformed.
        if start_index > end_index || end_index > source_len {
            debug!("find_width_of_character_at_span: source indexes are malformed");
            return 1;
        }

        let src = local_begin.sf.external_src.borrow();

        // We need to extend the snippet to the end of the src rather than to end_index so when
        // searching forwards for boundaries we've got somewhere to search.
        let snippet = if let Some(ref src) = local_begin.sf.src {
            &src[start_index..]
        } else if let Some(src) = src.get_source() {
            &src[start_index..]
        } else {
            return 1;
        };
        debug!("find_width_of_character_at_span: snippet=`{:?}`", snippet);

        let mut target = if forwards { end_index + 1 } else { end_index - 1 };
        debug!("find_width_of_character_at_span: initial target=`{:?}`", target);

        while !snippet.is_char_boundary(target - start_index) && target < source_len {
            target = if forwards {
                target + 1
            } else {
                match target.checked_sub(1) {
                    Some(target) => target,
                    None => {
                        break;
                    }
                }
            };
            debug!("find_width_of_character_at_span: target=`{:?}`", target);
        }
        debug!("find_width_of_character_at_span: final target=`{:?}`", target);

        if forwards { (target - end_index) as u32 } else { (end_index - target) as u32 }
    }

    pub fn get_source_file(&self, filename: &FileName) -> Option<Lrc<SourceFile>> {
        // Remap filename before lookup
        let filename = self.path_mapping().map_filename_prefix(filename).0;
        for sf in self.files.borrow().source_files.iter() {
            if filename == sf.name {
                return Some(sf.clone());
            }
        }
        None
    }

    /// For a global `BytePos`, computes the local offset within the containing `SourceFile`.
    pub fn lookup_byte_offset(&self, bpos: BytePos) -> SourceFileAndBytePos {
        let idx = self.lookup_source_file_idx(bpos);
        let sf = (*self.files.borrow().source_files)[idx].clone();
        let offset = bpos - sf.start_pos;
        SourceFileAndBytePos { sf, pos: offset }
    }

    // Returns the index of the `SourceFile` (in `self.files`) that contains `pos`.
    // This index is guaranteed to be valid for the lifetime of this `SourceMap`,
    // since `source_files` is a `MonotonicVec`
    pub fn lookup_source_file_idx(&self, pos: BytePos) -> usize {
        self.files
            .borrow()
            .source_files
            .binary_search_by_key(&pos, |key| key.start_pos)
            .unwrap_or_else(|p| p - 1)
    }

    pub fn count_lines(&self) -> usize {
        self.files().iter().fold(0, |a, f| a + f.count_lines())
    }

    pub fn ensure_source_file_source_present(&self, source_file: Lrc<SourceFile>) -> bool {
        source_file.add_external_src(|| {
            match source_file.name {
                FileName::Real(ref name) if let Some(local_path) = name.local_path() => {
                    self.file_loader.read_file(local_path).ok()
                }
                _ => None,
            }
        })
    }

    pub fn is_imported(&self, sp: Span) -> bool {
        let source_file_index = self.lookup_source_file_idx(sp.lo());
        let source_file = &self.files()[source_file_index];
        source_file.is_imported()
    }

    /// Gets the span of a statement. If the statement is a macro expansion, the
    /// span in the context of the block span is found. The trailing semicolon is included
    /// on a best-effort basis.
    pub fn stmt_span(&self, stmt_span: Span, block_span: Span) -> Span {
        if !stmt_span.from_expansion() {
            return stmt_span;
        }
        let mac_call = original_sp(stmt_span, block_span);
        self.mac_call_stmt_semi_span(mac_call).map_or(mac_call, |s| mac_call.with_hi(s.hi()))
    }

    /// Tries to find the span of the semicolon of a macro call statement.
    /// The input must be the *call site* span of a statement from macro expansion.
    /// ```ignore (illustrative)
    /// //       v output
    ///    mac!();
    /// // ^^^^^^ input
    /// ```
    pub fn mac_call_stmt_semi_span(&self, mac_call: Span) -> Option<Span> {
        let span = self.span_extend_while(mac_call, char::is_whitespace).ok()?;
        let span = span.shrink_to_hi().with_hi(BytePos(span.hi().0.checked_add(1)?));
        if self.span_to_snippet(span).as_deref() != Ok(";") {
            return None;
        }
        Some(span)
    }
}

#[derive(Clone)]
pub struct FilePathMapping {
    mapping: Vec<(PathBuf, PathBuf)>,
    filename_display_for_diagnostics: FileNameDisplayPreference,
}

impl FilePathMapping {
    pub fn empty() -> FilePathMapping {
        FilePathMapping::new(Vec::new())
    }

    pub fn new(mapping: Vec<(PathBuf, PathBuf)>) -> FilePathMapping {
        let filename_display_for_diagnostics = if mapping.is_empty() {
            FileNameDisplayPreference::Local
        } else {
            FileNameDisplayPreference::Remapped
        };

        FilePathMapping { mapping, filename_display_for_diagnostics }
    }

    /// Applies any path prefix substitution as defined by the mapping.
    /// The return value is the remapped path and a boolean indicating whether
    /// the path was affected by the mapping.
    pub fn map_prefix(&self, path: PathBuf) -> (PathBuf, bool) {
        if path.as_os_str().is_empty() {
            // Exit early if the path is empty and therefore there's nothing to remap.
            // This is mostly to reduce spam for `RUSTC_LOG=[remap_path_prefix]`.
            return (path, false);
        }

        return remap_path_prefix(&self.mapping, path);

        #[instrument(level = "debug", skip(mapping), ret)]
        fn remap_path_prefix(mapping: &[(PathBuf, PathBuf)], path: PathBuf) -> (PathBuf, bool) {
            // NOTE: We are iterating over the mapping entries from last to first
            //       because entries specified later on the command line should
            //       take precedence.
            for &(ref from, ref to) in mapping.iter().rev() {
                debug!("Trying to apply {from:?} => {to:?}");

                if let Ok(rest) = path.strip_prefix(from) {
                    let remapped = if rest.as_os_str().is_empty() {
                        // This is subtle, joining an empty path onto e.g. `foo/bar` will
                        // result in `foo/bar/`, that is, there'll be an additional directory
                        // separator at the end. This can lead to duplicated directory separators
                        // in remapped paths down the line.
                        // So, if we have an exact match, we just return that without a call
                        // to `Path::join()`.
                        to.clone()
                    } else {
                        to.join(rest)
                    };
                    debug!("Match - remapped");

                    return (remapped, true);
                } else {
                    debug!("No match - prefix {from:?} does not match");
                }
            }

            debug!("not remapped");
            (path, false)
        }
    }

    fn map_filename_prefix(&self, file: &FileName) -> (FileName, bool) {
        match file {
            FileName::Real(realfile) if let RealFileName::LocalPath(local_path) = realfile => {
                let (mapped_path, mapped) = self.map_prefix(local_path.to_path_buf());
                let realfile = if mapped {
                    RealFileName::Remapped {
                        local_path: Some(local_path.clone()),
                        virtual_name: mapped_path,
                    }
                } else {
                    realfile.clone()
                };
                (FileName::Real(realfile), mapped)
            }
            FileName::Real(_) => unreachable!("attempted to remap an already remapped filename"),
            other => (other.clone(), false),
        }
    }

    /// Expand a relative path to an absolute path with remapping taken into account.
    /// Use this when absolute paths are required (e.g. debuginfo or crate metadata).
    ///
    /// The resulting `RealFileName` will have its `local_path` portion erased if
    /// possible (i.e. if there's also a remapped path).
    pub fn to_embeddable_absolute_path(
        &self,
        file_path: RealFileName,
        working_directory: &RealFileName,
    ) -> RealFileName {
        match file_path {
            // Anything that's already remapped we don't modify, except for erasing
            // the `local_path` portion.
            RealFileName::Remapped { local_path: _, virtual_name } => {
                RealFileName::Remapped {
                    // We do not want any local path to be exported into metadata
                    local_path: None,
                    // We use the remapped name verbatim, even if it looks like a relative
                    // path. The assumption is that the user doesn't want us to further
                    // process paths that have gone through remapping.
                    virtual_name,
                }
            }

            RealFileName::LocalPath(unmapped_file_path) => {
                // If no remapping has been applied yet, try to do so
                let (new_path, was_remapped) = self.map_prefix(unmapped_file_path);
                if was_remapped {
                    // It was remapped, so don't modify further
                    return RealFileName::Remapped { local_path: None, virtual_name: new_path };
                }

                if new_path.is_absolute() {
                    // No remapping has applied to this path and it is absolute,
                    // so the working directory cannot influence it either, so
                    // we are done.
                    return RealFileName::LocalPath(new_path);
                }

                debug_assert!(new_path.is_relative());
                let unmapped_file_path_rel = new_path;

                match working_directory {
                    RealFileName::LocalPath(unmapped_working_dir_abs) => {
                        let file_path_abs = unmapped_working_dir_abs.join(unmapped_file_path_rel);

                        // Although neither `working_directory` nor the file name were subject
                        // to path remapping, the concatenation between the two may be. Hence
                        // we need to do a remapping here.
                        let (file_path_abs, was_remapped) = self.map_prefix(file_path_abs);
                        if was_remapped {
                            RealFileName::Remapped {
                                // Erase the actual path
                                local_path: None,
                                virtual_name: file_path_abs,
                            }
                        } else {
                            // No kind of remapping applied to this path, so
                            // we leave it as it is.
                            RealFileName::LocalPath(file_path_abs)
                        }
                    }
                    RealFileName::Remapped {
                        local_path: _,
                        virtual_name: remapped_working_dir_abs,
                    } => {
                        // If working_directory has been remapped, then we emit
                        // Remapped variant as the expanded path won't be valid
                        RealFileName::Remapped {
                            local_path: None,
                            virtual_name: Path::new(remapped_working_dir_abs)
                                .join(unmapped_file_path_rel),
                        }
                    }
                }
            }
        }
    }
}