| 1 | /*! |
| 2 | @file |
| 3 | Forward declares `boost::hana::Iterable`. |
| 4 | |
| 5 | Copyright Louis Dionne 2013-2022 |
| 6 | Distributed under the Boost Software License, Version 1.0. |
| 7 | (See accompanying file LICENSE.md or copy at http://boost.org/LICENSE_1_0.txt) |
| 8 | */ |
| 9 | |
| 10 | #ifndef BOOST_HANA_FWD_CONCEPT_ITERABLE_HPP |
| 11 | #define BOOST_HANA_FWD_CONCEPT_ITERABLE_HPP |
| 12 | |
| 13 | #include <boost/hana/config.hpp> |
| 14 | |
| 15 | |
| 16 | namespace boost { namespace hana { |
| 17 | //! @ingroup group-concepts |
| 18 | //! @defgroup group-Iterable Iterable |
| 19 | //! The `Iterable` concept represents data structures supporting external |
| 20 | //! iteration. |
| 21 | //! |
| 22 | //! Intuitively, an `Iterable` can be seen as a kind of container whose |
| 23 | //! elements can be pulled out one at a time. An `Iterable` also provides |
| 24 | //! a way to know when the _container_ is empty, i.e. when there are no |
| 25 | //! more elements to pull out. |
| 26 | //! |
| 27 | //! Whereas `Foldable` represents data structures supporting internal |
| 28 | //! iteration with the ability to accumulate a result, the `Iterable` |
| 29 | //! concept allows inverting the control of the iteration. This is more |
| 30 | //! flexible than `Foldable`, since it allows iterating over only some |
| 31 | //! part of the structure. This, in turn, allows `Iterable` to work on |
| 32 | //! infinite structures, while trying to fold such a structure would |
| 33 | //! never finish. |
| 34 | //! |
| 35 | //! |
| 36 | //! Minimal complete definition |
| 37 | //! --------------------------- |
| 38 | //! `at`, `drop_front` and `is_empty` |
| 39 | //! |
| 40 | //! |
| 41 | //! @anchor Iterable-lin |
| 42 | //! The linearization of an `Iterable` |
| 43 | //! ---------------------------------- |
| 44 | //! Intuitively, for an `Iterable` structure `xs`, the _linearization_ of |
| 45 | //! `xs` is the sequence of all the elements in `xs` as if they had been |
| 46 | //! put in a (possibly infinite) list: |
| 47 | //! @code |
| 48 | //! linearization(xs) = [x1, x2, x3, ...] |
| 49 | //! @endcode |
| 50 | //! |
| 51 | //! The `n`th element of the linearization of an `Iterable` can be |
| 52 | //! accessed with the `at` function. In other words, `at(xs, n) == xn`. |
| 53 | //! |
| 54 | //! Note that this notion is precisely the extension of the [linearization] |
| 55 | //! (@ref Foldable-lin) notion of `Foldable`s to the infinite case. This |
| 56 | //! notion is useful for expressing various properties of `Iterable`s, |
| 57 | //! and is used for that elsewhere in the documentation. |
| 58 | //! |
| 59 | //! |
| 60 | //! Compile-time `Iterable`s |
| 61 | //! ------------------------ |
| 62 | //! A _compile-time_ `Iterable` is an `Iterable` for which `is_empty` |
| 63 | //! returns a compile-time `Logical`. These structures allow iteration |
| 64 | //! to be done at compile-time, in the sense that the "loop" doing the |
| 65 | //! iteration can be unrolled because the total length of the structure |
| 66 | //! is kown at compile-time. |
| 67 | //! |
| 68 | //! In particular, note that being a compile-time `Iterable` has nothing |
| 69 | //! to do with being finite or infinite. For example, it would be possible |
| 70 | //! to create a sequence representing the Pythagorean triples as |
| 71 | //! `integral_constant`s. Such a sequence would be infinite, but iteration |
| 72 | //! on the sequence would still be done at compile-time. However, if one |
| 73 | //! tried to iterate over _all_ the elements of the sequence, the compiler |
| 74 | //! would loop indefinitely, in contrast to your program looping |
| 75 | //! indefinitely if the sequence was a runtime one. |
| 76 | //! |
| 77 | //! __In the current version of the library, only compile-time `Iterable`s |
| 78 | //! are supported.__ While it would be possible in theory to support |
| 79 | //! runtime `Iterable`s, doing it efficiently is the subject of some |
| 80 | //! research. In particular, follow [this issue][1] for the current |
| 81 | //! status of runtime `Iterable`s. |
| 82 | //! |
| 83 | //! |
| 84 | //! Laws |
| 85 | //! ---- |
| 86 | //! First, we require the equality of two `Iterable`s to be related to the |
| 87 | //! equality of the elements in their linearizations. More specifically, |
| 88 | //! if `xs` and `ys` are two `Iterable`s of data type `It`, then |
| 89 | //! @code |
| 90 | //! xs == ys => at(xs, i) == at(ys, i) for all i |
| 91 | //! @endcode |
| 92 | //! |
| 93 | //! This conveys that two `Iterable`s must have the same linearization |
| 94 | //! in order to be considered equal. |
| 95 | //! |
| 96 | //! Secondly, since every `Iterable` is also a `Searchable`, we require |
| 97 | //! the models of `Iterable` and `Searchable` to be consistent. This is |
| 98 | //! made precise by the following laws. For any `Iterable` `xs` with a |
| 99 | //! linearization of `[x1, x2, x3, ...]`, |
| 100 | //! @code |
| 101 | //! any_of(xs, equal.to(z)) <=> xi == z |
| 102 | //! @endcode |
| 103 | //! for some _finite_ index `i`. Furthermore, |
| 104 | //! @code |
| 105 | //! find_if(xs, pred) == just(the first xi such that pred(xi) is satisfied) |
| 106 | //! @endcode |
| 107 | //! or `nothing` if no such `xi` exists. |
| 108 | //! |
| 109 | //! |
| 110 | //! Refined concepts |
| 111 | //! ---------------- |
| 112 | //! 1. `Searchable` (free model)\n |
| 113 | //! Any `Iterable` gives rise to a model of `Searchable`, where the keys |
| 114 | //! and the values are both the elements in the structure. Searching for |
| 115 | //! a key is just doing a linear search through the elements of the |
| 116 | //! structure. |
| 117 | //! @include example/iterable/searchable.cpp |
| 118 | //! |
| 119 | //! 2. `Foldable` for finite `Iterable`s\n |
| 120 | //! Every finite `Iterable` gives rise to a model of `Foldable`. For |
| 121 | //! these models to be consistent, we require the models of both `Foldable` |
| 122 | //! and `Iterable` to have the same linearization. |
| 123 | //! |
| 124 | //! @note |
| 125 | //! As explained above, `Iterable`s are also `Searchable`s and their |
| 126 | //! models have to be consistent. By the laws presented here, it also |
| 127 | //! means that the `Foldable` model for finite `Iterable`s has to be |
| 128 | //! consistent with the `Searchable` model. |
| 129 | //! |
| 130 | //! For convenience, finite `Iterable`s must only provide a definition of |
| 131 | //! `length` to model the `Foldable` concept; defining the more powerful |
| 132 | //! `unpack` or `fold_left` is not necessary (but still possible). The |
| 133 | //! default implementation of `unpack` derived from `Iterable` + `length` |
| 134 | //! uses the fact that `at(xs, i)` denotes the `i`th element of `xs`'s |
| 135 | //! linearization, and that the linearization of a finite `Iterable` must |
| 136 | //! be the same as its linearization as a `Foldable`. |
| 137 | //! |
| 138 | //! |
| 139 | //! Concrete models |
| 140 | //! --------------- |
| 141 | //! `hana::tuple`, `hana::string`, `hana::range` |
| 142 | //! |
| 143 | //! |
| 144 | //! [1]: https://github.com/boostorg/hana/issues/40 |
| 145 | template <typename It> |
| 146 | struct Iterable; |
| 147 | }} // end namespace boost::hana |
| 148 | |
| 149 | #endif // !BOOST_HANA_FWD_CONCEPT_ITERABLE_HPP |
| 150 | |