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
// Copyright © 2019 The Rust Fuzz Project Developers.
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.

//! Wrappers around raw, unstructured bytes.

use crate::{Arbitrary, Error, Result};
use std::marker::PhantomData;
use std::ops::ControlFlow;
use std::{mem, ops};

/// A source of unstructured data.
///
/// An `Unstructured` helps `Arbitrary` implementations interpret raw data
/// (typically provided by a fuzzer) as a "DNA string" that describes how to
/// construct the `Arbitrary` type. The goal is that a small change to the "DNA
/// string" (the raw data wrapped by an `Unstructured`) results in a small
/// change to the generated `Arbitrary` instance. This helps a fuzzer
/// efficiently explore the `Arbitrary`'s input space.
///
/// `Unstructured` is deterministic: given the same raw data, the same series of
/// API calls will return the same results (modulo system resource constraints,
/// like running out of memory). However, `Unstructured` does not guarantee
/// anything beyond that: it makes not guarantee that it will yield bytes from
/// the underlying data in any particular order.
///
/// You shouldn't generally need to use an `Unstructured` unless you are writing
/// a custom `Arbitrary` implementation by hand, instead of deriving it. Mostly,
/// you should just be passing it through to nested `Arbitrary::arbitrary`
/// calls.
///
/// # Example
///
/// Imagine you were writing a color conversion crate. You might want to write
/// fuzz tests that take a random RGB color and assert various properties, run
/// functions and make sure nothing panics, etc.
///
/// Below is what translating the fuzzer's raw input into an `Unstructured` and
/// using that to generate an arbitrary RGB color might look like:
///
/// ```
/// # #[cfg(feature = "derive")] fn foo() {
/// use arbitrary::{Arbitrary, Unstructured};
///
/// /// An RGB color.
/// #[derive(Arbitrary)]
/// pub struct Rgb {
///     r: u8,
///     g: u8,
///     b: u8,
/// }
///
/// // Get the raw bytes from the fuzzer.
/// #   let get_input_from_fuzzer = || &[];
/// let raw_data: &[u8] = get_input_from_fuzzer();
///
/// // Wrap it in an `Unstructured`.
/// let mut unstructured = Unstructured::new(raw_data);
///
/// // Generate an `Rgb` color and run our checks.
/// if let Ok(rgb) = Rgb::arbitrary(&mut unstructured) {
/// #   let run_my_color_conversion_checks = |_| {};
///     run_my_color_conversion_checks(rgb);
/// }
/// # }
/// ```
pub struct Unstructured<'a> {
    data: &'a [u8],
}

impl<'a> Unstructured<'a> {
    /// Create a new `Unstructured` from the given raw data.
    ///
    /// # Example
    ///
    /// ```
    /// use arbitrary::Unstructured;
    ///
    /// let u = Unstructured::new(&[1, 2, 3, 4]);
    /// ```
    pub fn new(data: &'a [u8]) -> Self {
        Unstructured { data }
    }

    /// Get the number of remaining bytes of underlying data that are still
    /// available.
    ///
    /// # Example
    ///
    /// ```
    /// use arbitrary::{Arbitrary, Unstructured};
    ///
    /// let mut u = Unstructured::new(&[1, 2, 3]);
    ///
    /// // Initially have three bytes of data.
    /// assert_eq!(u.len(), 3);
    ///
    /// // Generating a `bool` consumes one byte from the underlying data, so
    /// // we are left with two bytes afterwards.
    /// let _ = bool::arbitrary(&mut u);
    /// assert_eq!(u.len(), 2);
    /// ```
    #[inline]
    pub fn len(&self) -> usize {
        self.data.len()
    }

    /// Is the underlying unstructured data exhausted?
    ///
    /// `unstructured.is_empty()` is the same as `unstructured.len() == 0`.
    ///
    /// # Example
    ///
    /// ```
    /// use arbitrary::{Arbitrary, Unstructured};
    ///
    /// let mut u = Unstructured::new(&[1, 2, 3, 4]);
    ///
    /// // Initially, we are not empty.
    /// assert!(!u.is_empty());
    ///
    /// // Generating a `u32` consumes all four bytes of the underlying data, so
    /// // we become empty afterwards.
    /// let _ = u32::arbitrary(&mut u);
    /// assert!(u.is_empty());
    /// ```
    #[inline]
    pub fn is_empty(&self) -> bool {
        self.len() == 0
    }

    /// Generate an arbitrary instance of `A`.
    ///
    /// This is simply a helper method that is equivalent to `<A as
    /// Arbitrary>::arbitrary(self)`. This helper is a little bit more concise,
    /// and can be used in situations where Rust's type inference will figure
    /// out what `A` should be.
    ///
    /// # Example
    ///
    /// ```
    /// # #[cfg(feature="derive")] fn foo() -> arbitrary::Result<()> {
    /// use arbitrary::{Arbitrary, Unstructured};
    ///
    /// #[derive(Arbitrary)]
    /// struct MyType {
    ///     // ...
    /// }
    ///
    /// fn do_stuff(value: MyType) {
    /// #   let _ = value;
    ///     // ...
    /// }
    ///
    /// let mut u = Unstructured::new(&[1, 2, 3, 4]);
    ///
    /// // Rust's type inference can figure out that `value` should be of type
    /// // `MyType` here:
    /// let value = u.arbitrary()?;
    /// do_stuff(value);
    /// # Ok(()) }
    /// ```
    pub fn arbitrary<A>(&mut self) -> Result<A>
    where
        A: Arbitrary<'a>,
    {
        <A as Arbitrary<'a>>::arbitrary(self)
    }

    /// Get the number of elements to insert when building up a collection of
    /// arbitrary `ElementType`s.
    ///
    /// This uses the [`<ElementType as
    /// Arbitrary>::size_hint`][crate::Arbitrary::size_hint] method to smartly
    /// choose a length such that we most likely have enough underlying bytes to
    /// construct that many arbitrary `ElementType`s.
    ///
    /// This should only be called within an `Arbitrary` implementation.
    ///
    /// # Example
    ///
    /// ```
    /// use arbitrary::{Arbitrary, Result, Unstructured};
    /// # pub struct MyCollection<T> { _t: std::marker::PhantomData<T> }
    /// # impl<T> MyCollection<T> {
    /// #     pub fn with_capacity(capacity: usize) -> Self { MyCollection { _t: std::marker::PhantomData } }
    /// #     pub fn insert(&mut self, element: T) {}
    /// # }
    ///
    /// impl<'a, T> Arbitrary<'a> for MyCollection<T>
    /// where
    ///     T: Arbitrary<'a>,
    /// {
    ///     fn arbitrary(u: &mut Unstructured<'a>) -> Result<Self> {
    ///         // Get the number of `T`s we should insert into our collection.
    ///         let len = u.arbitrary_len::<T>()?;
    ///
    ///         // And then create a collection of that length!
    ///         let mut my_collection = MyCollection::with_capacity(len);
    ///         for _ in 0..len {
    ///             let element = T::arbitrary(u)?;
    ///             my_collection.insert(element);
    ///         }
    ///
    ///         Ok(my_collection)
    ///     }
    /// }
    /// ```
    pub fn arbitrary_len<ElementType>(&mut self) -> Result<usize>
    where
        ElementType: Arbitrary<'a>,
    {
        let byte_size = self.arbitrary_byte_size()?;
        let (lower, upper) = <ElementType as Arbitrary>::size_hint(0);
        let elem_size = upper.unwrap_or_else(|| lower * 2);
        let elem_size = std::cmp::max(1, elem_size);
        Ok(byte_size / elem_size)
    }

    fn arbitrary_byte_size(&mut self) -> Result<usize> {
        if self.data.is_empty() {
            Ok(0)
        } else if self.data.len() == 1 {
            self.data = &[];
            Ok(0)
        } else {
            // Take lengths from the end of the data, since the `libFuzzer` folks
            // found that this lets fuzzers more efficiently explore the input
            // space.
            //
            // https://github.com/rust-fuzz/libfuzzer-sys/blob/0c450753/libfuzzer/utils/FuzzedDataProvider.h#L92-L97

            // We only consume as many bytes as necessary to cover the entire
            // range of the byte string.
            let len = if self.data.len() <= std::u8::MAX as usize + 1 {
                let bytes = 1;
                let max_size = self.data.len() - bytes;
                let (rest, for_size) = self.data.split_at(max_size);
                self.data = rest;
                Self::int_in_range_impl(0..=max_size as u8, for_size.iter().copied())?.0 as usize
            } else if self.data.len() <= std::u16::MAX as usize + 1 {
                let bytes = 2;
                let max_size = self.data.len() - bytes;
                let (rest, for_size) = self.data.split_at(max_size);
                self.data = rest;
                Self::int_in_range_impl(0..=max_size as u16, for_size.iter().copied())?.0 as usize
            } else if self.data.len() <= std::u32::MAX as usize + 1 {
                let bytes = 4;
                let max_size = self.data.len() - bytes;
                let (rest, for_size) = self.data.split_at(max_size);
                self.data = rest;
                Self::int_in_range_impl(0..=max_size as u32, for_size.iter().copied())?.0 as usize
            } else {
                let bytes = 8;
                let max_size = self.data.len() - bytes;
                let (rest, for_size) = self.data.split_at(max_size);
                self.data = rest;
                Self::int_in_range_impl(0..=max_size as u64, for_size.iter().copied())?.0 as usize
            };

            Ok(len)
        }
    }

    /// Generate an integer within the given range.
    ///
    /// Do not use this to generate the size of a collection. Use
    /// `arbitrary_len` instead.
    ///
    /// # Panics
    ///
    /// Panics if `range.start >= range.end`. That is, the given range must be
    /// non-empty.
    ///
    /// # Example
    ///
    /// ```
    /// use arbitrary::{Arbitrary, Unstructured};
    ///
    /// let mut u = Unstructured::new(&[1, 2, 3, 4]);
    ///
    /// let x: i32 = u.int_in_range(-5_000..=-1_000)
    ///     .expect("constructed `u` with enough bytes to generate an `i32`");
    ///
    /// assert!(-5_000 <= x);
    /// assert!(x <= -1_000);
    /// ```
    pub fn int_in_range<T>(&mut self, range: ops::RangeInclusive<T>) -> Result<T>
    where
        T: Int,
    {
        let (result, bytes_consumed) = Self::int_in_range_impl(range, self.data.iter().cloned())?;
        self.data = &self.data[bytes_consumed..];
        Ok(result)
    }

    fn int_in_range_impl<T>(
        range: ops::RangeInclusive<T>,
        mut bytes: impl Iterator<Item = u8>,
    ) -> Result<(T, usize)>
    where
        T: Int,
    {
        let start = range.start();
        let end = range.end();
        assert!(
            start <= end,
            "`arbitrary::Unstructured::int_in_range` requires a non-empty range"
        );

        // When there is only one possible choice, don't waste any entropy from
        // the underlying data.
        if start == end {
            return Ok((*start, 0));
        }

        let range: T::Widest = end.as_widest() - start.as_widest();
        let mut result = T::Widest::ZERO;
        let mut offset: usize = 0;

        while offset < mem::size_of::<T>()
            && (range >> T::Widest::from_usize(offset * 8)) > T::Widest::ZERO
        {
            let byte = bytes.next().ok_or(Error::NotEnoughData)?;
            result = (result << 8) | T::Widest::from_u8(byte);
            offset += 1;
        }

        // Avoid division by zero.
        if let Some(range) = range.checked_add(T::Widest::ONE) {
            result = result % range;
        }

        Ok((
            T::from_widest(start.as_widest().wrapping_add(result)),
            offset,
        ))
    }

    /// Choose one of the given choices.
    ///
    /// This should only be used inside of `Arbitrary` implementations.
    ///
    /// Returns an error if there is not enough underlying data to make a
    /// choice or if no choices are provided.
    ///
    /// # Examples
    ///
    /// Selecting from an array of choices:
    ///
    /// ```
    /// use arbitrary::Unstructured;
    ///
    /// let mut u = Unstructured::new(&[1, 2, 3, 4, 5, 6, 7, 8, 9, 0]);
    /// let choices = ['a', 'b', 'c', 'd', 'e', 'f', 'g'];
    ///
    /// let choice = u.choose(&choices).unwrap();
    ///
    /// println!("chose {}", choice);
    /// ```
    ///
    /// An error is returned if no choices are provided:
    ///
    /// ```
    /// use arbitrary::Unstructured;
    ///
    /// let mut u = Unstructured::new(&[1, 2, 3, 4, 5, 6, 7, 8, 9, 0]);
    /// let choices: [char; 0] = [];
    ///
    /// let result = u.choose(&choices);
    ///
    /// assert!(result.is_err());
    /// ```
    pub fn choose<'b, T>(&mut self, choices: &'b [T]) -> Result<&'b T> {
        if choices.is_empty() {
            return Err(Error::EmptyChoose);
        }
        let idx = self.int_in_range(0..=choices.len() - 1)?;
        Ok(&choices[idx])
    }

    /// Generate a boolean according to the given ratio.
    ///
    /// # Panics
    ///
    /// Panics when the numerator and denominator do not meet these constraints:
    ///
    /// * `0 < numerator <= denominator`
    ///
    /// # Example
    ///
    /// Generate a boolean that is `true` five sevenths of the time:
    ///
    /// ```
    /// # fn foo() -> arbitrary::Result<()> {
    /// use arbitrary::Unstructured;
    ///
    /// # let my_data = [1, 2, 3, 4, 5, 6, 7, 8, 9, 0];
    /// let mut u = Unstructured::new(&my_data);
    ///
    /// if u.ratio(5, 7)? {
    ///     // Take this branch 5/7 of the time.
    /// }
    /// # Ok(())
    /// # }
    /// ```
    pub fn ratio<T>(&mut self, numerator: T, denominator: T) -> Result<bool>
    where
        T: Int,
    {
        assert!(T::ZERO < numerator);
        assert!(numerator <= denominator);
        let x = self.int_in_range(T::ONE..=denominator)?;
        Ok(x <= numerator)
    }

    /// Fill a `buffer` with bytes from the underlying raw data.
    ///
    /// This should only be called within an `Arbitrary` implementation. This is
    /// a very low-level operation. You should generally prefer calling nested
    /// `Arbitrary` implementations like `<Vec<u8>>::arbitrary` and
    /// `String::arbitrary` over using this method directly.
    ///
    /// If this `Unstructured` does not have enough underlying data to fill the
    /// whole `buffer`, it pads the buffer out with zeros.
    ///
    /// # Example
    ///
    /// ```
    /// use arbitrary::Unstructured;
    ///
    /// let mut u = Unstructured::new(&[1, 2, 3, 4]);
    ///
    /// let mut buf = [0; 2];
    ///
    /// assert!(u.fill_buffer(&mut buf).is_ok());
    /// assert_eq!(buf, [1, 2]);
    ///
    /// assert!(u.fill_buffer(&mut buf).is_ok());
    /// assert_eq!(buf, [3, 4]);
    ///
    /// assert!(u.fill_buffer(&mut buf).is_ok());
    /// assert_eq!(buf, [0, 0]);
    /// ```
    pub fn fill_buffer(&mut self, buffer: &mut [u8]) -> Result<()> {
        let n = std::cmp::min(buffer.len(), self.data.len());
        buffer[..n].copy_from_slice(&self.data[..n]);
        for byte in buffer[n..].iter_mut() {
            *byte = 0;
        }
        self.data = &self.data[n..];
        Ok(())
    }

    /// Provide `size` bytes from the underlying raw data.
    ///
    /// This should only be called within an `Arbitrary` implementation. This is
    /// a very low-level operation. You should generally prefer calling nested
    /// `Arbitrary` implementations like `<Vec<u8>>::arbitrary` and
    /// `String::arbitrary` over using this method directly.
    ///
    /// # Example
    ///
    /// ```
    /// use arbitrary::Unstructured;
    ///
    /// let mut u = Unstructured::new(&[1, 2, 3, 4]);
    ///
    /// assert!(u.bytes(2).unwrap() == &[1, 2]);
    /// assert!(u.bytes(2).unwrap() == &[3, 4]);
    /// ```
    pub fn bytes(&mut self, size: usize) -> Result<&'a [u8]> {
        if self.data.len() < size {
            return Err(Error::NotEnoughData);
        }

        let (for_buf, rest) = self.data.split_at(size);
        self.data = rest;
        Ok(for_buf)
    }

    /// Peek at `size` number of bytes of the underlying raw input.
    ///
    /// Does not consume the bytes, only peeks at them.
    ///
    /// Returns `None` if there are not `size` bytes left in the underlying raw
    /// input.
    ///
    /// # Example
    ///
    /// ```
    /// use arbitrary::Unstructured;
    ///
    /// let u = Unstructured::new(&[1, 2, 3]);
    ///
    /// assert_eq!(u.peek_bytes(0).unwrap(), []);
    /// assert_eq!(u.peek_bytes(1).unwrap(), [1]);
    /// assert_eq!(u.peek_bytes(2).unwrap(), [1, 2]);
    /// assert_eq!(u.peek_bytes(3).unwrap(), [1, 2, 3]);
    ///
    /// assert!(u.peek_bytes(4).is_none());
    /// ```
    pub fn peek_bytes(&self, size: usize) -> Option<&'a [u8]> {
        self.data.get(..size)
    }

    /// Consume all of the rest of the remaining underlying bytes.
    ///
    /// Returns a slice of all the remaining, unconsumed bytes.
    ///
    /// # Example
    ///
    /// ```
    /// use arbitrary::Unstructured;
    ///
    /// let mut u = Unstructured::new(&[1, 2, 3]);
    ///
    /// let mut remaining = u.take_rest();
    ///
    /// assert_eq!(remaining, [1, 2, 3]);
    /// ```
    pub fn take_rest(mut self) -> &'a [u8] {
        mem::replace(&mut self.data, &[])
    }

    /// Provide an iterator over elements for constructing a collection
    ///
    /// This is useful for implementing [`Arbitrary::arbitrary`] on collections
    /// since the implementation is simply `u.arbitrary_iter()?.collect()`
    pub fn arbitrary_iter<'b, ElementType: Arbitrary<'a>>(
        &'b mut self,
    ) -> Result<ArbitraryIter<'a, 'b, ElementType>> {
        Ok(ArbitraryIter {
            u: &mut *self,
            _marker: PhantomData,
        })
    }

    /// Provide an iterator over elements for constructing a collection from
    /// all the remaining bytes.
    ///
    /// This is useful for implementing [`Arbitrary::arbitrary_take_rest`] on collections
    /// since the implementation is simply `u.arbitrary_take_rest_iter()?.collect()`
    pub fn arbitrary_take_rest_iter<ElementType: Arbitrary<'a>>(
        self,
    ) -> Result<ArbitraryTakeRestIter<'a, ElementType>> {
        let (lower, upper) = ElementType::size_hint(0);

        let elem_size = upper.unwrap_or(lower * 2);
        let elem_size = std::cmp::max(1, elem_size);
        let size = self.len() / elem_size;
        Ok(ArbitraryTakeRestIter {
            size,
            u: Some(self),
            _marker: PhantomData,
        })
    }

    /// Call the given function an arbitrary number of times.
    ///
    /// The function is given this `Unstructured` so that it can continue to
    /// generate arbitrary data and structures.
    ///
    /// You may optionaly specify minimum and maximum bounds on the number of
    /// times the function is called.
    ///
    /// You may break out of the loop early by returning
    /// `Ok(std::ops::ControlFlow::Break)`. To continue the loop, return
    /// `Ok(std::ops::ControlFlow::Continue)`.
    ///
    /// # Panics
    ///
    /// Panics if `min > max`.
    ///
    /// # Example
    ///
    /// Call a closure that generates an arbitrary type inside a context an
    /// arbitrary number of times:
    ///
    /// ```
    /// use arbitrary::{Result, Unstructured};
    /// use std::ops::ControlFlow;
    ///
    /// enum Type {
    ///     /// A boolean type.
    ///     Bool,
    ///
    ///     /// An integer type.
    ///     Int,
    ///
    ///     /// A list of the `i`th type in this type's context.
    ///     List(usize),
    /// }
    ///
    /// fn arbitrary_types_context(u: &mut Unstructured) -> Result<Vec<Type>> {
    ///     let mut context = vec![];
    ///
    ///     u.arbitrary_loop(Some(10), Some(20), |u| {
    ///         let num_choices = if context.is_empty() {
    ///             2
    ///         } else {
    ///             3
    ///         };
    ///         let ty = match u.int_in_range::<u8>(1..=num_choices)? {
    ///             1 => Type::Bool,
    ///             2 => Type::Int,
    ///             3 => Type::List(u.int_in_range(0..=context.len() - 1)?),
    ///             _ => unreachable!(),
    ///         };
    ///         context.push(ty);
    ///         Ok(ControlFlow::Continue(()))
    ///     })?;
    ///
    ///     // The number of loop iterations are constrained by the min/max
    ///     // bounds that we provided.
    ///     assert!(context.len() >= 10);
    ///     assert!(context.len() <= 20);
    ///
    ///     Ok(context)
    /// }
    /// ```
    pub fn arbitrary_loop(
        &mut self,
        min: Option<u32>,
        max: Option<u32>,
        mut f: impl FnMut(&mut Self) -> Result<ControlFlow<(), ()>>,
    ) -> Result<()> {
        let min = min.unwrap_or(0);
        let max = max.unwrap_or(u32::MAX);
        assert!(min <= max);

        for _ in 0..min {
            match f(self)? {
                ControlFlow::Continue(_) => continue,
                ControlFlow::Break(_) => return Ok(()),
            }
        }

        for _ in 0..(max - min) {
            let keep_going = self.arbitrary().unwrap_or(false);
            if !keep_going {
                break;
            }
            match f(self)? {
                ControlFlow::Continue(_) => continue,
                ControlFlow::Break(_) => break,
            }
        }

        Ok(())
    }
}

/// Utility iterator produced by [`Unstructured::arbitrary_iter`]
pub struct ArbitraryIter<'a, 'b, ElementType> {
    u: &'b mut Unstructured<'a>,
    _marker: PhantomData<ElementType>,
}

impl<'a, 'b, ElementType: Arbitrary<'a>> Iterator for ArbitraryIter<'a, 'b, ElementType> {
    type Item = Result<ElementType>;
    fn next(&mut self) -> Option<Result<ElementType>> {
        let keep_going = self.u.arbitrary().unwrap_or(false);
        if keep_going {
            Some(Arbitrary::arbitrary(self.u))
        } else {
            None
        }
    }
}

/// Utility iterator produced by [`Unstructured::arbitrary_take_rest_iter`]
pub struct ArbitraryTakeRestIter<'a, ElementType> {
    u: Option<Unstructured<'a>>,
    size: usize,
    _marker: PhantomData<ElementType>,
}

impl<'a, ElementType: Arbitrary<'a>> Iterator for ArbitraryTakeRestIter<'a, ElementType> {
    type Item = Result<ElementType>;
    fn next(&mut self) -> Option<Result<ElementType>> {
        if let Some(mut u) = self.u.take() {
            if self.size == 1 {
                Some(Arbitrary::arbitrary_take_rest(u))
            } else if self.size == 0 {
                None
            } else {
                self.size -= 1;
                let ret = Arbitrary::arbitrary(&mut u);
                self.u = Some(u);
                Some(ret)
            }
        } else {
            None
        }
    }
}

/// A trait that is implemented for all of the primitive integers:
///
/// * `u8`
/// * `u16`
/// * `u32`
/// * `u64`
/// * `u128`
/// * `usize`
/// * `i8`
/// * `i16`
/// * `i32`
/// * `i64`
/// * `i128`
/// * `isize`
///
/// Don't implement this trait yourself.
pub trait Int:
    Copy
    + PartialOrd
    + Ord
    + ops::Sub<Self, Output = Self>
    + ops::Rem<Self, Output = Self>
    + ops::Shr<Self, Output = Self>
    + ops::Shl<usize, Output = Self>
    + ops::BitOr<Self, Output = Self>
{
    #[doc(hidden)]
    type Widest: Int;

    #[doc(hidden)]
    const ZERO: Self;

    #[doc(hidden)]
    const ONE: Self;

    #[doc(hidden)]
    fn as_widest(self) -> Self::Widest;

    #[doc(hidden)]
    fn from_widest(w: Self::Widest) -> Self;

    #[doc(hidden)]
    fn from_u8(b: u8) -> Self;

    #[doc(hidden)]
    fn from_usize(u: usize) -> Self;

    #[doc(hidden)]
    fn checked_add(self, rhs: Self) -> Option<Self>;

    #[doc(hidden)]
    fn wrapping_add(self, rhs: Self) -> Self;
}

macro_rules! impl_int {
    ( $( $ty:ty : $widest:ty ; )* ) => {
        $(
            impl Int for $ty {
                type Widest = $widest;

                const ZERO: Self = 0;

                const ONE: Self = 1;

                fn as_widest(self) -> Self::Widest {
                    self as $widest
                }

                fn from_widest(w: Self::Widest) -> Self {
                    let x = <$ty>::max_value().as_widest();
                    (w % x) as Self
                }

                fn from_u8(b: u8) -> Self {
                    b as Self
                }

                fn from_usize(u: usize) -> Self {
                    u as Self
                }

                fn checked_add(self, rhs: Self) -> Option<Self> {
                    <$ty>::checked_add(self, rhs)
                }

                fn wrapping_add(self, rhs: Self) -> Self {
                    <$ty>::wrapping_add(self, rhs)
                }
            }
        )*
    }
}

impl_int! {
    u8: u128;
    u16: u128;
    u32: u128;
    u64: u128;
    u128: u128;
    usize: u128;
    i8: i128;
    i16: i128;
    i32: i128;
    i64: i128;
    i128: i128;
    isize: i128;
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_byte_size() {
        let mut u = Unstructured::new(&[1, 2, 3, 4, 5, 6, 7, 8, 9, 6]);
        // Should take one byte off the end
        assert_eq!(u.arbitrary_byte_size().unwrap(), 6);
        assert_eq!(u.len(), 9);
        let mut v = vec![];
        v.resize(260, 0);
        v.push(1);
        v.push(4);
        let mut u = Unstructured::new(&v);
        // Should read two bytes off the end
        assert_eq!(u.arbitrary_byte_size().unwrap(), 0x104);
        assert_eq!(u.len(), 260);
    }

    #[test]
    fn int_in_range_of_one() {
        let mut u = Unstructured::new(&[1, 2, 3, 4, 5, 6, 7, 8, 9, 6]);
        let x = u.int_in_range(0..=0).unwrap();
        assert_eq!(x, 0);
        let choice = *u.choose(&[42]).unwrap();
        assert_eq!(choice, 42)
    }

    #[test]
    fn int_in_range_uses_minimal_amount_of_bytes() {
        let mut u = Unstructured::new(&[1]);
        u.int_in_range::<u8>(0..=u8::MAX).unwrap();

        let mut u = Unstructured::new(&[1]);
        u.int_in_range::<u32>(0..=u8::MAX as u32).unwrap();

        let mut u = Unstructured::new(&[1]);
        u.int_in_range::<u32>(0..=u8::MAX as u32 + 1).unwrap_err();
    }
}