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
//! [String]-like wrappers around [Bytes] and [BytesMut].
//!
//! The [Bytes] and [BytesMut] provide a buffer of bytes with ability to create owned slices into
//! the same shared memory allocation. This allows cheap manipulation of data.
//!
//! Strings are mostly just byte buffers with extra APIs to manipulate them. The standard [String]
//! type is built as a wrapper around [Vec]. We build similar wrappers around the [Bytes] and
//! [BytesMut], gaining the ability to create owned shared slices for textual data as well.
//!
//! Users are expected to use the [Str] and [StrMut] types. Note that these are type aliases around
//! the [StrInner] type. The latter is means to implement both in one go and contains all the
//! documentation, but is not meant to be used directly.
//!
//! # Splitting
//!
//! The [prim@str] type from standard library (which the types here dereference to) allows for
//! slicing and splitting in many convenient ways. They, however, return borrowed string slices
//! (`&str`), which might pose some problems.
//!
//! The [Str], and to certain extent, the [StrMut] type additionally allow cheap splitting and
//! slicing that produce owned [Str] and [StrMut] respectively. They are slightly more expensive
//! than the slicing than the ones returning `&str`, but only by incrementing internal reference
//! counts. They do not clone the actual string data, like `.to_owned()` on the standard library
//! methods would. These methods are available in addition to the standard ones.
//!
//! There are three ways how this can be done:
//!
//! * By dedicated methods, like [lines_bytes][StrInner::lines_bytes] (in general, the name of the
//!   standard method suffixed with `_bytes`).
//! * By using the [BytesIter] iterator manually.
//! * By using the standard-library methods, producing `&str` and translating it back to [Str] with
//!   [slice][StrInner::slice] or [StrInner::slice_ref].
//!
//! # Examples
//!
//! ```rust
//! # use bytes::Bytes;
//! # use bytes_utils::{Str, StrMut};
//! let mut builder = StrMut::new();
//! builder += "Hello";
//! builder.push(' ');
//! builder.push_str("World");
//! assert_eq!("Hello World", builder);
//!
//! let s1 = builder.split_built().freeze();
//! // This is a cheap copy, in the form of incrementing a reference count.
//! let s2 = s1.clone();
//! assert_eq!("Hello World", s1);
//! assert_eq!("Hello World", s2);
//! // Slicing is cheap as well, even though the returned things are Str and therefore owned too.
//! assert_eq!("ello", s1.slice(1..5));
//! // We have taken the data out of the builder, but the rest of its capacity can be used for
//! // further things.
//! assert_eq!("", builder);
//!
//! // Creating from strings and similar works
//! let a = Str::from("Hello");
//! assert_eq!("Hello", a);
//!
//! let e = Str::new();
//! assert_eq!("", e);
//!
//! // And from static str in O(1)
//! let b = Str::from_static("World");
//! assert_eq!("World", b);
//!
//! // And from Bytes too.
//! let b = Str::try_from(Bytes::from_static(b"World")).expect("Must be utf8");
//! assert_eq!("World", b);
//! // Invalid utf8 is refused.
//! Str::try_from(Bytes::from_static(&[0, 0, 255])).unwrap_err();
//! ```

use alloc::borrow::Cow;
use alloc::boxed::Box;
use alloc::string::String;
use core::borrow::{Borrow, BorrowMut};
use core::cmp::Ordering;
use core::convert::Infallible;
use core::fmt::{Debug, Display, Formatter, Result as FmtResult, Write};
use core::hash::{Hash, Hasher};
use core::iter::{self, FromIterator};
use core::ops::{Add, AddAssign, Deref, DerefMut, Index, IndexMut};
use core::str::{self, FromStr};

#[cfg(feature = "std")]
use std::error::Error;

use bytes::{Bytes, BytesMut};
use either::Either;

#[cfg(feature = "serde")]
mod serde_impl;

/// Error when creating [Str] or [StrMut] from invalid UTF8 data.
#[derive(Copy, Clone, Debug)]
pub struct Utf8Error<S> {
    e: core::str::Utf8Error,
    inner: S,
}

impl<S> Utf8Error<S> {
    /// Returns the byte buffer back to the caller.
    pub fn into_inner(self) -> S {
        self.inner
    }

    /// The inner description of why the data is invalid UTF8.
    pub fn utf8_error(&self) -> str::Utf8Error {
        self.e
    }
}

impl<S> Display for Utf8Error<S> {
    fn fmt(&self, fmt: &mut Formatter) -> FmtResult {
        Display::fmt(&self.e, fmt)
    }
}

#[cfg(feature = "std")]
impl<S: Debug> Error for Utf8Error<S> {}

/// Direction of iteration.
///
/// See [BytesIter].
#[derive(Copy, Clone, Debug, Eq, PartialEq)]
pub enum Direction {
    /// Move forward (in the normal direction) in the string.
    Forward,

    /// Move backwards in the string.
    Backward,
}

/// Manual splitting iterator.
///
/// The methods on [Str] and [StrMut] that iterate use this internally. But it can also be used
/// manually to generate other iterators that split the original into parts.
#[derive(Clone, Debug)]
pub struct BytesIter<S, F> {
    bytes: Option<S>,
    extract: F,
    direction: Direction,
}

impl<S, F> BytesIter<S, F>
where
    S: Storage,
    F: FnMut(&str) -> Option<(usize, usize)>,
{
    /// A constructor of the iterator.
    ///
    /// The `direction` specifies in what order chunks should be yielded.
    ///
    /// The `ext` closure is always called with the rest of not yet split string. It shall return
    /// the byte indices of the chunk and separator border. In case of forward iteration, it is the
    /// end of them and the separator needs to end further to the string (or at the same position).
    /// In the backwards direction, it is in reverse ‒ they specify their starts and the separator
    /// is before the chunk.
    ///
    /// # Panics
    ///
    /// If the indices don't point at a character boundary, the iteration will panic. It'll also
    /// panic if the returned indices are reversed or if they are out of bounds.
    pub fn new(s: StrInner<S>, direction: Direction, ext: F) -> Self {
        Self {
            bytes: Some(s.0),
            extract: ext,
            direction,
        }
    }
}

impl<S, F> Iterator for BytesIter<S, F>
where
    S: Storage,
    F: FnMut(&str) -> Option<(usize, usize)>,
{
    type Item = StrInner<S>;

    fn next(&mut self) -> Option<StrInner<S>> {
        let storage = self.bytes.take()?;
        // Safety: we keep sure it is valid UTF8 on the API boundary.
        let whole_str = unsafe { str::from_utf8_unchecked(storage.as_ref()) };
        fn split<S: Storage>(storage: S, left: usize, right: usize) -> (S, S) {
            let whole_str = unsafe { str::from_utf8_unchecked(storage.as_ref()) };
            // Sanity-check we are not slicing in the middle of utf8 code point. This would
            // panic if we do. It would also panic if we are out of range, which is also good.
            assert!(whole_str.is_char_boundary(left));
            assert!(whole_str.is_char_boundary(right));

            // Now that we are sure this is legal, we are going to slice the byte data for real.
            let (with_sep, end) = storage.split_at(right);
            let (start, _sep) = with_sep.split_at(left);
            (start, end)
        }
        match ((self.extract)(whole_str), self.direction) {
            (Some((chunk_end, sep_end)), Direction::Forward) => {
                assert!(chunk_end <= sep_end);
                let (start, end) = split(storage, chunk_end, sep_end);

                self.bytes = Some(end);
                Some(StrInner(start))
            }
            (Some((chunk_start, sep_start)), Direction::Backward) => {
                assert!(sep_start <= chunk_start);
                let (start, end) = split(storage, sep_start, chunk_start);

                self.bytes = Some(start);
                Some(StrInner(end))
            }
            (None, _) => {
                // No separator found -> return the whole rest (and keep None in ourselves)
                Some(StrInner(storage))
            }
        }
    }
}

/// Find a separator position, for use with the [BytesIter].
fn sep_find<F: Fn(char) -> bool>(s: &str, is_sep: F) -> Option<(usize, usize)> {
    let sep_start = s.find(&is_sep)?;
    let sep_end = s[sep_start..]
        .find(|c| !is_sep(c))
        .map(|e| e + sep_start)
        .unwrap_or_else(|| s.len());
    Some((sep_start, sep_end))
}

/// Separator for an empty pattern.
fn empty_sep(s: &str, limit: usize) -> Option<(usize, usize)> {
    let char_end = s
        .char_indices()
        .skip(1)
        .map(|(i, _)| i)
        .chain(iter::once(s.len()).take((!s.is_empty()) as usize))
        .take(limit)
        .next()?;
    Some((char_end, char_end))
}

fn rempty_sep(s: &str, limit: usize) -> Option<(usize, usize)> {
    let char_start = s.char_indices().rev().map(|(i, _)| i).take(limit).next()?;
    Some((char_start, char_start))
}

/// The backing storage for [StrInner]
///
/// This is currently a technical detail of the crate, users are not expected to implement this
/// trait. Use [Str] or [StrMut] type aliases.
///
/// # Safety
///
/// The storage must act "sane". But what exactly it means is not yet analyzed and may change in
/// future versions. Don't implement the trait (at least not yet).
pub unsafe trait Storage: AsRef<[u8]> + Default + Sized {
    /// A type that can be used to build the storage incrementally.
    ///
    /// For mutable storages, it may be itself. For immutable one, there needs to be a mutable
    /// counterpart that can be converted to immutable later on.
    type Creator: Default + StorageMut;

    /// Converts the creator (mutable storage) to self.
    ///
    /// In case of mutable storages, this should be identity.
    fn from_creator(creator: Self::Creator) -> Self;

    /// Splits the storage at the given byte index and creates two non-overlapping instances.
    fn split_at(self, at: usize) -> (Self, Self);
}

unsafe impl Storage for Bytes {
    type Creator = BytesMut;
    fn from_creator(creator: Self::Creator) -> Self {
        creator.freeze()
    }
    fn split_at(mut self, at: usize) -> (Self, Self) {
        let right = self.split_off(at);
        (self, right)
    }
}

unsafe impl Storage for BytesMut {
    type Creator = BytesMut;
    fn from_creator(creator: Self::Creator) -> Self {
        creator
    }
    fn split_at(mut self, at: usize) -> (Self, Self) {
        let right = self.split_off(at);
        (self, right)
    }
}

/// Trait for extra functionality of a mutable storage.
///
/// This is in addition to what an immutable storage must satisfy.
///
/// # Safety
///
/// The storage must act "sane". But what exactly it means is not yet analyzed and may change in
/// future versions. Don't implement the trait (at least not yet).
pub unsafe trait StorageMut: Storage + AsMut<[u8]> {
    /// An immutable counter-part storage.
    type Immutable: Storage<Creator = Self>;

    /// Adds some more bytes to the end of the storage.
    fn push_slice(&mut self, s: &[u8]);
}

unsafe impl StorageMut for BytesMut {
    type Immutable = Bytes;
    fn push_slice(&mut self, s: &[u8]) {
        self.extend_from_slice(s)
    }
}

/// Implementation of the [Str] and [StrMut] types.
///
/// For technical reasons, both are implemented in one go as this type. For the same reason, most
/// of the documentation can be found here. Users are expected to use the [Str] and [StrMut]
/// instead.
#[derive(Clone, Default)]
pub struct StrInner<S>(S);

impl<S: Storage> StrInner<S> {
    /// Creates an empty instance.
    pub fn new() -> Self {
        Self::default()
    }

    /// Extracts the inner byte storage.
    pub fn into_inner(self) -> S {
        self.0
    }

    /// Access to the inner storage.
    pub fn inner(&self) -> &S {
        &self.0
    }

    /// Creates an instance from an existing byte storage.
    ///
    /// It may fail if the content is not valid UTF8.
    ///
    /// A [try_from][TryFrom::try_from] may be used instead.
    pub fn from_inner(s: S) -> Result<Self, Utf8Error<S>> {
        match str::from_utf8(s.as_ref()) {
            Ok(_) => Ok(Self(s)),
            Err(e) => Err(Utf8Error { e, inner: s }),
        }
    }

    /// Same as [from_inner][StrInner::from_inner], but without the checks.
    ///
    /// # Safety
    ///
    /// The caller must ensure content is valid UTF8.
    pub const unsafe fn from_inner_unchecked(s: S) -> Self {
        Self(s)
    }

    /// Splits the string into two at the given index.
    ///
    /// # Panics
    ///
    /// If the index is not at char boundary.
    pub fn split_at_bytes(self, at: usize) -> (Self, Self) {
        assert!(self.deref().is_char_boundary(at));
        let (l, r) = self.0.split_at(at);
        (Self(l), Self(r))
    }

    /// Splits into whitespace separated "words".
    ///
    /// This acts like [split_whitespace][str::split_whitespace], but yields owned instances. It
    /// doesn't clone the content, it just increments some reference counts.
    pub fn split_whitespace_bytes(self) -> impl Iterator<Item = Self> {
        BytesIter::new(self, Direction::Forward, |s| {
            sep_find(s, char::is_whitespace)
        })
        .filter(|s| !s.is_empty())
    }

    /// Splits into whitespace separated "words".
    ///
    /// This acts like [split_ascii_whitespace][str::split_ascii_whitespace], but yields owned
    /// instances. This doesn't clone the content, it just increments some reference counts.
    pub fn split_ascii_whitespace_bytes(self) -> impl Iterator<Item = Self> {
        BytesIter::new(self, Direction::Forward, |s| {
            sep_find(s, |c| c.is_ascii() && (c as u8).is_ascii_whitespace())
        })
        .filter(|s| !s.is_empty())
    }

    /// Splits into lines.
    ///
    /// This acts like [lines][str::lines], but yields owned instances. The content is not cloned,
    /// this just increments some reference counts.
    pub fn lines_bytes(self) -> impl Iterator<Item = Self> {
        if self.is_empty() {
            Either::Left(iter::empty())
        } else {
            let iter = BytesIter::new(self, Direction::Forward, |s| sep_find(s, |c| c == '\n'))
                .map(|s| match s.chars().next() {
                    Some('\r') => s.split_at_bytes(1).1,
                    _ => s,
                });
            Either::Right(iter)
        }
    }

    /// Splits with the provided separator.
    ///
    /// This acts somewhat like [split][str::split], but yields owned instances. Also, it accepts
    /// only string patters (since the `Pattern` is not stable ☹). The content is not cloned, this
    /// just increments some reference counts.
    pub fn split_bytes<'s>(self, sep: &'s str) -> impl Iterator<Item = Self> + 's
    where
        S: 's,
    {
        if sep.is_empty() {
            let bulk = BytesIter::new(self, Direction::Forward, |s| empty_sep(s, usize::MAX));
            Either::Left(iter::once(Self::default()).chain(bulk))
        } else {
            let sep_find = move |s: &str| s.find(sep).map(|pos| (pos, pos + sep.len()));
            Either::Right(BytesIter::new(self, Direction::Forward, sep_find))
        }
    }

    /// Splits max. `n` times according to the given pattern.
    ///
    /// This acts somewhat like [splitn][str::splitn], but yields owned instances. Also, it accepts
    /// only string patters (since the `Pattern` is not stable ☹). The content is not cloned, this
    /// just increments some reference counts.
    pub fn splitn_bytes<'s>(self, mut n: usize, sep: &'s str) -> impl Iterator<Item = Self> + 's
    where
        S: 's,
    {
        // TODO: This seems to work, but is ugly. Any idea how to simplify?
        if sep.is_empty() {
            if n <= 1 {
                Either::Left(Either::Left(iter::once(self).take(n)))
            } else {
                n -= 1;
                let bulk = BytesIter::new(self, Direction::Forward, move |s| {
                    n -= 1;
                    empty_sep(s, n)
                });
                Either::Left(Either::Right(iter::once(Self::default()).chain(bulk)))
            }
        } else {
            let sep_find = move |s: &str| {
                n -= 1;
                if n == 0 {
                    None
                } else {
                    s.find(sep).map(|pos| (pos, pos + sep.len()))
                }
            };
            Either::Right(BytesIter::new(self, Direction::Forward, sep_find).take(n))
        }
    }

    /// A reverse version of [split_bytes][Self::split_bytes].
    pub fn rsplit_bytes<'s>(self, sep: &'s str) -> impl Iterator<Item = Self> + 's
    where
        S: 's,
    {
        if sep.is_empty() {
            let bulk = BytesIter::new(self, Direction::Backward, |s| rempty_sep(s, usize::MAX));
            Either::Left(iter::once(Self::default()).chain(bulk))
        } else {
            let sep_find = move |s: &str| s.rfind(sep).map(|pos| (pos + sep.len(), pos));
            Either::Right(BytesIter::new(self, Direction::Backward, sep_find))
        }
    }

    /// A reverse version of [splitn_bytes][Self::splitn_bytes].
    pub fn rsplitn_bytes<'s>(self, mut n: usize, sep: &'s str) -> impl Iterator<Item = Self> + 's
    where
        S: 's,
    {
        // TODO: This seems to work, but is ugly. Any idea how to simplify?
        if sep.is_empty() {
            if n <= 1 {
                Either::Left(Either::Left(iter::once(self).take(n)))
            } else {
                n -= 1;
                let bulk = BytesIter::new(self, Direction::Backward, move |s| {
                    n -= 1;
                    rempty_sep(s, n)
                });
                Either::Left(Either::Right(iter::once(Self::default()).chain(bulk)))
            }
        } else {
            let sep_find = move |s: &str| {
                n -= 1;
                if n == 0 {
                    None
                } else {
                    s.rfind(sep).map(|pos| (pos + sep.len(), pos))
                }
            };
            Either::Right(BytesIter::new(self, Direction::Backward, sep_find).take(n))
        }
    }
}

impl<S: StorageMut> StrInner<S> {
    /// Appends a string.
    pub fn push_str(&mut self, s: &str) {
        self.0.push_slice(s.as_bytes());
    }

    /// Appends one character.
    pub fn push(&mut self, c: char) {
        self.push_str(c.encode_utf8(&mut [0; 4]));
    }

    /// Provides mutable access to the inner buffer.
    ///
    /// # Safety
    ///
    /// The caller must ensure that the content stays valid UTF8.
    pub unsafe fn inner_mut(&mut self) -> &mut S {
        &mut self.0
    }

    /// Turns the mutable variant into an immutable one.
    ///
    /// The advantage is that it can then be shared (also by small parts).
    pub fn freeze(self) -> StrInner<S::Immutable> {
        StrInner(S::Immutable::from_creator(self.0))
    }
}

impl<S: Storage> Deref for StrInner<S> {
    type Target = str;

    fn deref(&self) -> &str {
        unsafe { str::from_utf8_unchecked(self.0.as_ref()) }
    }
}

impl<S: StorageMut> DerefMut for StrInner<S> {
    fn deref_mut(&mut self) -> &mut str {
        unsafe { str::from_utf8_unchecked_mut(self.0.as_mut()) }
    }
}

impl<S, T> AsRef<T> for StrInner<S>
where
    S: Storage,
    str: AsRef<T>,
{
    fn as_ref(&self) -> &T {
        self.deref().as_ref()
    }
}

impl<S: StorageMut> AsMut<str> for StrInner<S> {
    fn as_mut(&mut self) -> &mut str {
        self.deref_mut()
    }
}

impl<S: Storage> Borrow<str> for StrInner<S> {
    fn borrow(&self) -> &str {
        self.deref()
    }
}

impl<S: StorageMut> BorrowMut<str> for StrInner<S> {
    fn borrow_mut(&mut self) -> &mut str {
        self.deref_mut()
    }
}

impl<S: Storage> Debug for StrInner<S> {
    fn fmt(&self, fmt: &mut Formatter) -> FmtResult {
        Debug::fmt(self.deref(), fmt)
    }
}

impl<S: Storage> Display for StrInner<S> {
    fn fmt(&self, fmt: &mut Formatter) -> FmtResult {
        Display::fmt(self.deref(), fmt)
    }
}

impl<S: Storage> Hash for StrInner<S> {
    fn hash<H: Hasher>(&self, state: &mut H) {
        self.deref().hash(state)
    }
}

impl<S, I> Index<I> for StrInner<S>
where
    S: Storage,
    str: Index<I>,
{
    type Output = <str as Index<I>>::Output;

    fn index(&self, index: I) -> &Self::Output {
        self.deref().index(index)
    }
}

impl<S, I> IndexMut<I> for StrInner<S>
where
    S: StorageMut,
    str: IndexMut<I>,
{
    fn index_mut(&mut self, index: I) -> &mut Self::Output {
        self.deref_mut().index_mut(index)
    }
}

impl<S: StorageMut> Add<&str> for StrInner<S> {
    type Output = Self;

    fn add(mut self, rhs: &str) -> Self::Output {
        self.push_str(rhs);
        self
    }
}

impl<S: StorageMut> AddAssign<&str> for StrInner<S> {
    fn add_assign(&mut self, rhs: &str) {
        self.push_str(rhs);
    }
}

impl<S: StorageMut> Extend<char> for StrInner<S> {
    fn extend<T: IntoIterator<Item = char>>(&mut self, iter: T) {
        for c in iter {
            self.push(c);
        }
    }
}

impl<'a, S: StorageMut> Extend<&'a char> for StrInner<S> {
    fn extend<T: IntoIterator<Item = &'a char>>(&mut self, iter: T) {
        for c in iter {
            self.push(*c);
        }
    }
}

macro_rules! impl_extend {
    ($ty:ty $(, $lifetimes:lifetime )* ) => {
        impl<$($lifetimes, )* S: StorageMut> Extend<$ty> for StrInner<S> {
            fn extend<T: IntoIterator<Item = $ty>>(&mut self, iter: T) {
                for i in iter {
                    self.push_str(i.as_ref());
                }
            }
        }

        impl<$($lifetimes, )* S> FromIterator<$ty> for StrInner<S>
        where
            S: Storage,
        {
            fn from_iter<T: IntoIterator<Item = $ty>>(iter: T) -> Self {
                let mut creator = StrInner(S::Creator::default());
                creator.extend(iter);
                StrInner(S::from_creator(creator.0))
            }
        }
    };
}

impl_extend!(String);
impl_extend!(Box<str>);
impl_extend!(&'a String, 'a);
impl_extend!(&'a str, 'a);
impl_extend!(Cow<'a, str>, 'a);

macro_rules! impl_from {
    ($ty:ty $(, $lifetimes:lifetime )* ) => {
        impl<$($lifetimes, )* S> From<$ty> for StrInner<S>
        where
            S: Storage,
        {
            fn from(s: $ty) -> Self {
                iter::once(s).collect()
            }
        }
    };
}

impl_from!(&'a String, 'a);
impl_from!(&'a str, 'a);
impl_from!(Cow<'a, str>, 'a);

impl From<String> for Str {
    fn from(s: String) -> Self {
        let inner = Bytes::from(s.into_bytes());
        // Safety: inner is constructed from a str
        unsafe { Str::from_inner_unchecked(inner) }
    }
}

impl From<Box<str>> for Str {
    fn from(s: Box<str>) -> Self {
        let s: Box<[u8]> = s.into();
        let inner = Bytes::from(s);
        // Safety: inner is constructed from a str
        unsafe { Str::from_inner_unchecked(inner) }
    }
}

macro_rules! impl_try_from {
    ($ty: ty) => {
        impl TryFrom<$ty> for StrInner<$ty> {
            type Error = Utf8Error<$ty>;
            fn try_from(s: $ty) -> Result<Self, Utf8Error<$ty>> {
                Self::from_inner(s)
            }
        }

        impl From<StrInner<$ty>> for $ty {
            fn from(s: StrInner<$ty>) -> $ty {
                s.0
            }
        }
    };
}

impl_try_from!(Bytes);
impl_try_from!(BytesMut);

impl From<StrMut> for Str {
    fn from(s: StrMut) -> Self {
        s.freeze()
    }
}

impl<S: Storage> FromStr for StrInner<S> {
    type Err = Infallible;

    fn from_str(s: &str) -> Result<Self, Self::Err> {
        Ok(s.into())
    }
}

impl<S: Storage> PartialEq for StrInner<S> {
    fn eq(&self, other: &Self) -> bool {
        self.deref() == other.deref()
    }
}

impl<S: Storage> Eq for StrInner<S> {}

impl<S: Storage> PartialOrd for StrInner<S> {
    fn partial_cmp(&self, other: &Self) -> Option<Ordering> {
        Some(Ord::cmp(self, other))
    }
}

impl<S: Storage> Ord for StrInner<S> {
    fn cmp(&self, other: &Self) -> Ordering {
        self.deref().cmp(other.deref())
    }
}

macro_rules! impl_partrial_eq {
    ($ty: ty $(, $lifetimes:lifetime )* ) => {
        impl<$($lifetimes, )* S: Storage> PartialEq<$ty> for StrInner<S> {
            fn eq(&self, other: &$ty) -> bool {
                self.deref() == other.deref()
            }
        }

        impl<$($lifetimes, )* S: Storage> PartialEq<StrInner<S>> for $ty {
            fn eq(&self, other: &StrInner<S>) -> bool {
                self.deref() == other.deref()
            }
        }

        impl<$($lifetimes, )* S: Storage> PartialOrd<$ty> for StrInner<S> {
            fn partial_cmp(&self, other: &$ty) -> Option<Ordering> {
                Some(self.deref().cmp(other.deref()))
            }
        }

        impl<$($lifetimes, )* S: Storage> PartialOrd<StrInner<S>> for $ty {
            fn partial_cmp(&self, other: &StrInner<S>) -> Option<Ordering> {
                Some(self.deref().cmp(other.deref()))
            }
        }
    };
}

impl_partrial_eq!(String);
impl_partrial_eq!(Box<str>);
impl_partrial_eq!(&'a str, 'a);
impl_partrial_eq!(&'a mut str, 'a);
impl_partrial_eq!(Cow<'a, str>, 'a);

impl<S: StorageMut> Write for StrInner<S> {
    fn write_str(&mut self, s: &str) -> FmtResult {
        self.push_str(s);
        Ok(())
    }
}
/// The [format] macro, but returning [Str].
///
/// # Examples
///
/// ```
/// use bytes_utils::{format_bytes, Str};
/// let s: Str = format_bytes!("Hello {}", "world");
/// assert_eq!("Hello world", s);
/// ```
#[macro_export]
macro_rules! format_bytes {
    ($($arg: tt)*) => {
        $crate::format_bytes_mut!($($arg)*).freeze()
    }
}

/// The [format] macro, but returning [StrMut].
///
/// # Examples
///
/// ```
/// use bytes_utils::{format_bytes_mut, StrMut};
/// let s: StrMut = format_bytes_mut!("Hello {}", "world");
/// assert_eq!("Hello world", s);
/// ```
#[macro_export]
macro_rules! format_bytes_mut {
    ($($arg: tt)*) => {{
        use std::fmt::Write;
        let mut buf = $crate::StrMut::default();
        write!(buf, $($arg)*).unwrap();
        buf
    }}
}

/// An immutable variant of [Bytes]-backed string.
///
/// The methods and their documentation are on [StrInner], but users are mostly expected to use
/// this and the [StrMut] aliases.
pub type Str = StrInner<Bytes>;

impl Str {
    /// Extracts a subslice of the string as an owned [Str].
    ///
    /// # Panics
    ///
    /// If the byte indices in the range are not on char boundaries.
    pub fn slice<R>(&self, range: R) -> Str
    where
        str: Index<R, Output = str>,
    {
        self.slice_ref(&self[range])
    }

    /// Extracts owned representation of the slice passed.
    ///
    /// This method accepts a string sub-slice of `self`. It then extracts the slice but as the
    /// [Str] type. This makes it easier to use "ordinary" string parsing/manipulation and then go
    /// back to holding the [Bytes]-based representation.
    ///
    /// This is zero-copy, the common part will be shared by reference counting.
    ///
    /// # Panics
    ///
    /// If the provided slice is not a sub-slice of `self`. This is checked based on address of the
    /// slice, not on the content.
    ///
    /// # Example
    ///
    /// ```rust
    /// # use bytes_utils::Str;
    /// let owned = Str::from("Hello World");
    /// let borrowed_mid: &str = &owned[2..5];
    ///
    /// let mid: Str = owned.slice_ref(borrowed_mid);
    /// assert_eq!("Hello World", owned);
    /// assert_eq!("llo", mid);
    /// ```
    pub fn slice_ref(&self, subslice: &str) -> Self {
        let sub = self.0.slice_ref(subslice.as_bytes());
        Self(sub)
    }

    /// Create [`Str`] from static string in O(1).
    pub const fn from_static(s: &'static str) -> Self {
        let bytes = Bytes::from_static(s.as_bytes());
        // Safety: bytes is constructed from str
        unsafe { Str::from_inner_unchecked(bytes) }
    }
}

/// A mutable variant of [BytesMut]-backed string.
///
/// Unlike [Str], this one allows modifications (mostly additions), but also doesn't allow
/// overlapping/shared chunks.
///
/// This is internally backed by the [StrInner] type, so the documentation of the methods are on
/// that.
pub type StrMut = StrInner<BytesMut>;

impl StrMut {
    /// Splits and returns the part of already built string, but keeps the extra capacity.
    pub fn split_built(&mut self) -> StrMut {
        StrInner(self.0.split())
    }
}

#[cfg(test)]
mod tests {
    use itertools::Itertools;
    use std::panic;

    use super::*;

    #[test]
    fn split_w_byte_index() {
        let v = Str::from("😈 ").split_whitespace_bytes().collect_vec();
        assert_eq!(1, v.len());
        assert_eq!("😈", v[0]);
    }

    #[test]
    fn split_same() {
        let v = Str::from("a").split_bytes("a").collect_vec();
        assert_eq!(2, v.len());
        assert_eq!("", v[0]);
        assert_eq!("", v[1]);
    }

    #[test]
    fn split_empty_pat() {
        let v = Str::from("a").split_bytes("").collect_vec();
        assert_eq!(3, v.len());
        assert_eq!("", v[0]);
        assert_eq!("a", v[1]);
        assert_eq!("", v[2]);
    }

    #[test]
    fn slice_checks_char_boundaries() {
        let v = Str::from("😈");
        assert_eq!(4, v.len());
        panic::catch_unwind(|| v.slice(1..)).unwrap_err();
    }

    #[test]
    fn split_at_bytes_mid() {
        let v = Str::from("hello");
        let (l, r) = v.split_at_bytes(2);
        assert_eq!("he", l);
        assert_eq!("llo", r);
    }

    #[test]
    fn split_at_bytes_begin() {
        let v = Str::from("hello");
        let (l, r) = v.split_at_bytes(0);
        assert_eq!("", l);
        assert_eq!("hello", r);
    }

    #[test]
    fn split_at_bytes_end() {
        let v = Str::from("hello");
        let (l, r) = v.split_at_bytes(5);
        assert_eq!("hello", l);
        assert_eq!("", r);
    }

    #[test]
    fn split_at_bytes_panic() {
        let v = Str::from("😈");
        assert_eq!(4, v.len());
        panic::catch_unwind(|| v.split_at_bytes(2)).unwrap_err();
    }

    #[cfg(not(miri))]
    mod proptests {
        use proptest::prelude::*;

        use super::*;

        proptest! {
            #[test]
            fn split_whitespace(s: String) {
                let bstring = Str::from(&s);

                let bw = bstring.split_whitespace_bytes();
                let sw = s.split_whitespace();

                for (b, s) in bw.zip_eq(sw) {
                    prop_assert_eq!(b, s);
                }
            }

            #[test]
            fn split_ascii_whitespace(s: String) {
                let bstring = Str::from(&s);

                let bw = bstring.split_ascii_whitespace_bytes();
                let sw = s.split_ascii_whitespace();

                for (b, s) in bw.zip_eq(sw) {
                    prop_assert_eq!(b, s);
                }
            }

            #[test]
            fn lines(s: String) {
                let bstring = Str::from(&s);

                let bl = bstring.lines_bytes();
                let sl = s.lines();

                for (b, s) in bl.zip_eq(sl) {
                    prop_assert_eq!(b, s);
                }
            }

            #[test]
            fn split(s: String, pat: String) {
                let bstring = Str::from(&s);

                let bs = bstring.split_bytes(&pat);
                let ss = s.split(&pat);

                for (b, s) in bs.zip_eq(ss) {
                    prop_assert_eq!(b, s);
                }
            }

            #[test]
            fn split_n(s: String, pat: String, n in 0..5usize) {
                let bstring = Str::from(&s);

                let bs = bstring.splitn_bytes(n, &pat);
                let ss = s.splitn(n, &pat);

                for (b, s) in bs.zip_eq(ss) {
                    prop_assert_eq!(b, s);
                }
            }

            #[test]
            fn rsplit(s: String, pat: String) {
                let bstring = Str::from(&s);

                let bs = bstring.rsplit_bytes(&pat);
                let ss = s.rsplit(&pat);

                for (b, s) in bs.zip_eq(ss) {
                    prop_assert_eq!(b, s);
                }
            }

            #[test]
            fn rsplit_n(s: String, pat: String, n in 0..5usize) {
                let bstring = Str::from(&s);

                let bs = bstring.rsplitn_bytes(n, &pat);
                let ss = s.rsplitn(n, &pat);

                for (b, s) in bs.zip_eq(ss) {
                    prop_assert_eq!(b, s);
                }
            }
        }
    }
}