Populate An Array Using Constexpr at Compile-time - c++

I would like to populate an array of enum using constexpr.
The content of the array follows a certain pattern.
I have an enum separating ASCII character set into four categories.
enum Type {
Alphabet,
Number,
Symbol,
Other,
};
constexpr Type table[128] = /* blah blah */;
I would like to have an array of 128 Type. They can be in a structure.
The index of the array will be corresponding to the ASCII characters and the value will be the Type of each character.
So I can query this array to find out which category an ASCII character belongs to. Something like
char c = RandomFunction();
if (table[c] == Alphabet)
DoSomething();
I would like to know if this is possible without some lengthy macro hacks.
Currently, I initialize the table by doing the following.
constexpr bool IsAlphabet (char c) {
return ((c >= 0x41 && c <= 0x5A) ||
(c >= 0x61 && c <= 0x7A));
}
constexpr bool IsNumber (char c) { /* blah blah */ }
constexpr bool IsSymbol (char c) { /* blah blah */ }
constexpr Type whichCategory (char c) { /* blah blah */ }
constexpr Type table[128] = { INITIALIZE };
where INITIALIZE is the entry point of some very lengthy macro hacks.
Something like
#define INITIALIZE INIT(0)
#define INIT(N) INIT_##N
#define INIT_0 whichCategory(0), INIT_1
#define INIT_1 whichCategory(1), INIT_2
//...
#define INIT_127 whichCategory(127)
I would like a way to populate this array or a structure containing the array without the need for this macro hack...
Maybe something like
struct Table {
Type _[128];
};
constexpr Table table = MagicFunction();
So, the question is how to write this MagicFunction?
Note: I am aware of cctype and likes, this question is more of a Is this possible? rather than Is this the best way to do it?.
Any help would be appreciated.
Thanks,

Ignoring ALL the issues, indices to the rescue:
template<unsigned... Is> struct seq{};
template<unsigned N, unsigned... Is>
struct gen_seq : gen_seq<N-1, N-1, Is...>{};
template<unsigned... Is>
struct gen_seq<0, Is...> : seq<Is...>{};
template<unsigned... Is>
constexpr Table MagicFunction(seq<Is...>){
return {{ whichCategory(Is)... }};
}
constexpr Table MagicFunction(){
return MagicFunction(gen_seq<128>{});
}
Live example.

In C++17 ::std::array has been updated to be more constexpr friendly and you can do the same as in C++14, but without some of the scary looking hacks to get around the lack of constexpr in crucial places. Here is what the code would look like there:
#include <array>
enum Type {
Alphabet,
Number,
Symbol,
Other,
};
constexpr ::std::array<Type, 128> MagicFunction()
{
using result_t = ::std::array<Type, 128>;
result_t result = {Other};
result[65] = Alphabet;
//....
return result;
}
const ::std::array<Type, 128> table = MagicFunction();
Again MagicFunction still needs to obey the rather loose constexpr rules. Mainly, it may not modify any global variables or use new (which implies modifying global state, namely the heap) or other such things.

IMHO the best way to do this is simply write a tiny setup program that will generate table for you. And then you can either throw out the setup program, or check it in alongside the generated source code.
The tricky part of this question is just a duplicate of this other one: Is it possible to create and initialize an array of values using template metaprogramming?
The trick is, it's impossible to write anything like
Type table[256] = some_expression();
at file scope, because global arrays can be initialized only with literal (source-level) initializer-lists. You can't initialize a global array with the result of a constexpr function, even if you could somehow get that function to return a std::initializer_list, which you can't because its constructor isn't declared constexpr.
So what you have to do is get the compiler to generate the array for you, by making it a static const data member of a template class. After one or two levels of metaprogramming that I'm too confused to write out, you'll bottom out in a line that looks something like
template <int... Indices>
Type DummyStruct<Indices...>::table[] = { whichCategory(Indices)... };
where Indices is a parameter-pack that looks like 0,1,2,... 254,255. You construct that parameter-pack using a recursive helper template, or maybe just using something out of Boost. And then you can write
constexpr Type (&table)[] = IndexHelperTemplate<256>::table;
...But why would you do all that, when the table is only 256 entries that will never change unless ASCII itself changes? The right way is the simplest way: precompute all 256 entries and write out the table explicitly, with no templates, constexpr, or any other magic.

The way to do this in C++14 looks like this:
#include <array>
enum Type {
Alphabet,
Number,
Symbol,
Other,
};
constexpr ::std::array<Type, 128> MagicFunction()
{
using result_t = ::std::array<Type, 128>;
result_t result = {Other};
const result_t &fake_const_result = result;
const_cast<result_t::reference>(fake_const_result[65]) = Alphabet;
//....
return result;
}
const ::std::array<Type, 128> table = MagicFunction();
No clever template hackery required any longer. Though, because C++14 didn't really undergo a thorough enough review of what did and didn't have to be constexpr in the standard library, a horrible hack involving const_cast has to be used.
And, of course, MagicFunction had better not modify any global variables or otherwise violate the constexpr rules. But those rules are pretty liberal nowadays. You can, for example, modify all the local variables you want, though passing them by reference or taking their addresses may not work out so well.
See my other answer for C++17, which allows you to drop some of the ugly-looking hacks.

Related

Building data structures at compile time with template-metaprogramming, constexpr or macros

I want to optimize a little programm/library i'm writing and since 2 weeks i'm somewhat stuck and now wondering if what i had in mind is even possible like that.
(Please be gentle i don't have very much experience in meta-programming.)
My goal is of course to have certain computations be done by the compiler, so that the programmer - hopefully - only has to edit code at one point in the program and have the compiler "create" all the boilerplate. I do have a resonably good idea how to do what i want with macros, but it is wished that i do it with templates if possible.
My goal is:
Lets say i have a class that a using programmer can derive from. There he can have multiple incoming and outgoing datatypes that i want to register somehow so that the base class can do i'ts operations on them.
class my_own_multiply : function_base {
in<int> a;
in<float> b;
out<double> c;
// ["..."] // other content of the class that actually does something but is irrelevant
register_ins<a, b> ins_of_function; // example meta-function calls
register_outs<c> outs_of_function;
}
The meta-code i have up till now is this: (but it's not jet working/complete)
template <typename... Ts>
struct register_ins {
const std::array<std::unique_ptr<in_type_erasured>, sizeof...(Ts)> ins;
constexpr std::array<std::unique_ptr<in_type_erasured>, sizeof...(Ts)>
build_ins_array() {
std::array<std::unique_ptr<in_type_erasured>, sizeof...(Ts)> ins_build;
for (unsigned int i = 0; i < sizeof...(Ts); ++i) {
ins_build[i] = std::make_unique<in_type_erasured>();
}
return ins_build;
}
constexpr register_ins() : ins(build_ins_array()) {
}
template <typename T>
T getValueOf(unsigned int in_nr) {
return ins[in_nr]->getValue();
}
};
As you may see, i want to call my meta-template-code with a variable number of ins. (Variable in the sens that the programmer can put however many he likes in there, but they won't change at runtime so they can be "baked" in at compile time)
The meta-code is supposed to be creating an array, that is of the lengt of the number of ins and is initialized so that every field points to the original in in the my_own_multiply class. Basically giving him an indexable data structure that will always have the correct size. And that i could access from the function_base class to use all ins for certain functions wich are also iterable making things convinient for me.
Now i have looked into how one might do that, but i now am getting the feeling that i might not really be allowed to "create" this array at compile time in a fashion that allows me to still have the ins a and b be non static and non const so that i can mutate them. From my side they wouldn't have to be const anyway, but my compliler seems to not like them to be free. The only thing i need const is the array with the pointers. But using constexpr possibly "makes" me make them const?
Okay, i will clarify what i don't get:
When i'm trying to create an "instance" of my meta-stuff-structure then it fails because it expects all kinds of const, constexpr and so on. But i don't want them since i need to be able to mutate most of those variables. I only need this meta-stuff to create an array of the correct size already at compile time. But i don't want to sacrifice having to make everything static and const in order to achive this. So is this even possible under these kinds of terms?
I do not get all the things you have in mind (also regarding that std::unique_ptr in your example), but maybe this helps:
Starting from C++14 (or C++11, but that is strictly limited) you may write constexpr functions which can be evaluated at compile-time. As a precondition (in simple words), all arguments "passed by the caller" must be constexpr. If you want to enforce that the compiler replaces that "call" by the result of a compile-time computation, you must assign the result to a constexpr.
Writing usual functions (just with constexpr added) allows to write code which is simple to read. Moreover, you can use the same code for both: compile-time computations and run-time computations.
C++17 example (similar things are possible in C++14, although some stuff from std is just missing the constexpr qualifier):
http://coliru.stacked-crooked.com/a/154e2dfcc41fb6c7
#include <cassert>
#include <array>
template<class T, std::size_t N>
constexpr std::array<T, N> multiply(
const std::array<T, N>& a,
const std::array<T, N>& b
) {
// may be evaluated in `constexpr` or in non-`constexpr` context
// ... in simple man's words this means:
// inside this function, `a` and `b` are not `constexpr`
// but the return can be used as `constexpr` if all arguments are `constexpr` for the "caller"
std::array<T, N> ret{};
for(size_t n=0; n<N; ++n) ret[n] = a[n] * b[n];
return ret;
}
int main() {
{// compile-time evaluation is possible if the input data is `constexpr`
constexpr auto a = std::array{2, 4, 6};
constexpr auto b = std::array{1, 2, 3};
constexpr auto c = multiply(a, b);// assigning to a `constexpr` guarantees compile-time evaluation
static_assert(c[0] == 2);
static_assert(c[1] == 8);
static_assert(c[2] == 18);
}
{// for run-time data, the same function can be used
auto a = std::array{2, 4, 6};
auto b = std::array{1, 2, 3};
auto c = multiply(a, b);
assert(c[0] == 2);
assert(c[1] == 8);
assert(c[2] == 18);
}
return 0;
}

How to initialize a map of arrays of different sizes?

Simple question, how to initialize a map of arrays (or some other container type) of different sizes? For example:
enum class code {A,B,C};
enum class res {X1,X2,X3,X4,X5};
std::map<code, ?> name {
{code:A, {res::X1,res::X2}},
{code:B, {res::X2,res::X3, res::X4}},
{code:C, {res::X5}}
};
I need to find at compile time if res::X2 is in map name at code::B
This expression should, I think, check for that using static_assert:
constexpr bool validate(code t, res p, int i = 0) {
return (name[t][i] == p ? true : ((sizeof(name[t]) == (i+1)) ? false :validate(t, p, ++i)));
}
Because validate is a constexpr a basic array would work but how to define it in a map argument as an array of type res? And that each array could be different in size?
So, I've made a mistake. I was under the impression that map can be accessed in constexpr function. What container type can you suggest me that would allow me to achieve what I have written above?
If it is ok for the array to be const, this works:
std::map<code, std::initializer_list<res>> name {
{code::A, {res::X1,res::X2}},
{code::B, {res::X2,res::X3, res::X4}},
{code::C, {res::X5}}
};
If you need to be able to write to the array, you'll need something like this:
std::map<code, std::vector<res>> name {
{code::A, {res::X1,res::X2}},
{code::B, {res::X2,res::X3, res::X4}},
{code::C, {res::X5}}
};
This will come at the cost of additional memory allocations, since vector allocated on the heap.
If you want it to be writable and are ok with a fixed size, this also works and avoids the additional allocations:
std::map<code, std::array<res, 3>> name {
{code::A, {res::X1,res::X2}},
{code::B, {res::X2,res::X3, res::X4}},
{code::C, {res::X5}}
};
As for the second part of your question, I'm not quite sure if any solution will work, given that you cannot access a map in a constexpr function.
You can't use a std::map in a constexpr expression, not least because it has a non-trivial destructor (~map(); specified in the standard).
If you search the map section of the standard for constexpr, you won't find it (http://eel.is/c++draft/map), whereas you will for std::array (http://eel.is/c++draft/array.syn)
Since std::map can't be used in constexpr expression I was forced to find another solution and thanks to #kfsone and #H. Guijt for pointing that out as well as mentioning std::initializer_list and std::array which lead me to this solution.
Since the data in the map in the question will be constant, i can do without an associative container. So I used a sorted array of std::initializer_lists together with a constexpr implementation of find function which works great for my needs.
template <class It, class T>
inline constexpr It sfind (It begin, It end, T const& value) noexcept
{
return ! (begin != end && *begin != value)? begin : sfind (begin+1, end, value);
}
enum class code {A,B,C};
enum class res {X1,X2,X3,X4,X5};
constexpr std::array<std::initializer_list<res>, 3> name {{
{res::X1,res::X2}, //code::A
{res::X2,res::X3, res::X4}, //code::B
{res::X5} // code::C
}};
const code c = code::A;
const res r = res::X3;
static_assert (name[static_cast<int>(c)].end () - sfind(name[static_cast<int>(c)].begin(), name[static_cast<int>(c)].end(), r) != 0,"Not found");

How to construct a tuple from an array

I am designing a C++ library that reads a CSV file of reported data from some experiment and does some aggregation and outputs a pgfplots code. I want to make the library as generic and easy to use as possible. I also want to isolate it from the data types that are represented in the CSV file and leave the option to user to parse each column as she desires. I also want to avoid Boost Spirit Qi or other heavy duty parser.
The simple solution I have is for the user to create a type for each column, with a constructor that takes "char *". The constructor does its own parsing for the value it is given, which is one cell from the data. The user then passes me a list of types; the schema, representing the types in a line of data. I use this type list to create a tuple, in which every member of the tuple is responsible for parsing itself.
The problem now is how to initialise (construct) this tuple. Dealing with tuples is of course not straightforward since iterating over their elements is mostly a compile-time operation. I used Boost Fusion at first to achieve this task. However, the function I used (transform) although might take a tuple as input (with the appropriate adapter), it does not seem to return a tuple. I need the return value to be a tuple so some other code can use it as an associative type-to-value container (access it by type via std::get<T>), while using only standard tools, that is, without using Boost. So I had to convert whatever Fusion's transform returned into std::tuple.
My question is how to avoid this conversion, and better yet how to avoid Boost Fusion completely.
A simple solution that comes to mind is to use the constructor of std::tuple, and somehow pass each element its respective "const *" that it needs to construct. However, while this is possible using some complicated template-based enumeration techniques, I am wondering if there is a straightforward "parameter-pack"-like approach, or an even simpler way to pass the values to the constructors of the individual elements of a tuple.
To clarify what I am seeking, kindly take a look at this following code.
#include <cstdio>
#include <array>
template <typename...> struct format {};
template <typename...> struct file_loader {};
template <typename... Format>
struct
file_loader<format<Format...> > {
void load_file() {
size_t strsize = 500u;
char *str = new char[strsize]();
auto is = fopen("RESULT","r");
/* example of RESULT:
dataset2,0.1004,524288
dataset1,0.3253,4194304
*/
while(getline(&str, &strsize, is) >= 0) {
std::array<char*, 3> toks{};
auto s = str;
int i = 2;
while(i --> 0)
toks[i] = strsep (&s, ",");
toks[2] = strsep (&s, ",\n");
std::tuple<Format...> the_line{ /* toks */ } ; // <-- HERE
//// current solution:
// auto the_line{
// as_std_tuple( // <-- unnecessary conversion I'd like to avoid
// boost::fusion::transform(boost::fusion::zip(types, toks), boost::fusion::make_fused( CAST() ))
// )};
// do something with the_line
}
}
};
#include <string>
class double_type {
public:
double_type() {}
double_type(char const *token) { } // strtod
};
class int_type {
public:
int_type() {}
int_type(char const *token) { } // strtoul
};
int main(int argc, char *argv[]) {
file_loader< format< std::string,
double_type,
int_type > >
{}.load_file();
return 0;
}
I've highlighted the interesting line as "HERE" in a comment.
My question precisely is:
Is there a way to construct a std::tuple instance (of heterogeneous
types, each of which is implicitly convertible from "char *") with
automatic storage duration (on the stack) from a std::array<char *, N>,
where N equals the size of that tuple?
The answer I am seeking should
Avoid Boost Fusion
(Simplicity condition) Avoid using more than 5 lines of boilerplate template-based enumeration code
Alternatively, shows why this is not possible to do in the C++14 standard
The answer can use C++17 constructs, I wouldn't mind.
Thank you,
As with all questions involving std::tuple, use index_sequence to give you a parameter pack to index the array with:
template <class... Formats, size_t N, size_t... Is>
std::tuple<Formats...> as_tuple(std::array<char*, N> const& arr,
std::index_sequence<Is...>)
{
return std::make_tuple(Formats{arr[Is]}...);
}
template <class... Formats, size_t N,
class = std::enable_if_t<(N == sizeof...(Formats))>>
std::tuple<Formats...> as_tuple(std::array<char*, N> const& arr)
{
return as_tuple<Formats...>(arr, std::make_index_sequence<N>{});
}
Which you would use as:
std::tuple<Format...> the_line = as_tuple<Format...>(toks);

strongly enum to integer and vice-versa

Here is my source (as answer) , how to implement underlying_value, and to_enum functions.
How to automatically convert strongly typed enum into int?
underlying_value - no problem.
but, to_enum - has problem.
see:
enum class E{ a = 1, b = 3, c = 5 };
auto e_a = utils::underlying_value(E::a); //OK
E t = utils::to_enum<E>( 2 ) ; // compiled, but it's incorrect. I think here must throws exception?
Q: how to implement to_enum correctly?
Despite the comments on the question, this can be done in C++11, though to do it without code repetition you will eventually have to wrap the enum class declaration in a macro. That may make my answer unsuitable, depending on your needs. Either way, doing the checked conversion requires some machinery, so I will get to the macro last.
The basic idea is to use constexpr functions to scan an array:
#include <iostream>
#include <stdexcept>
enum class E { a = 1, b = 3, c = 5 };
constexpr E values[] = {E::a, E::b, E::c};
constexpr size_t count = sizeof(values) / sizeof(E);
constexpr E to_enum(int value, size_t index = 0)
{
return
index >= count ? throw std::runtime_error("invalid integer") :
static_cast<int>(values[index]) == value ? values[index] :
to_enum(value, index + 1);
}
constexpr E converted = to_enum(3);
// Will not compile if uncommented.
// constexpr E bad_converted = to_enum(2);
int main()
{
std::cout << static_cast<int>(converted) << std::endl;
return 0;
}
This prints 3. If the line with bad_converted is uncommented, this code will not compile at all, as it says. The checked conversion can be done either at runtime or during compilation. It will be done during compilation if the argument to to_enum is a compile-time constant. Also, as you can probably see, this does a linear scan of values, but that can be replaced with another algorithm if it becomes a performance problem for a very large enum.
The code I just showed is a sketch that shows the underlying method. In order to make this less of a pain to use, you should wrap the declaration of E in a macro that will automatically generate the values[] array and the associated functions. I will show and rationalize the contents of this macro one point at a time.
The basic macro looks like this
// Declaration header
#define ENUM_CLASS(TypeName, __VA_ARGS__)
// Use
ENUM_CLASS(E, a = 1, b = 3, c = 5);
So, in this example, __VA_ARGS__ will be the tokens a = 1, b = 3, c = 5. So, we can declare the enum itself inside the macro as follows:
enum class TypeName { __VA_ARGS__ };
However, we cannot simply declare:
constexpr TypeName values[] = { __VA_ARGS__ };
because that expands to
constexpr TypeName values[] = { a = 1, b = 3, c = 5 };
which is not scoped (missing TypeName:: in front of each value), and is not valid C++ because of the extra assignment operators inside the array initializer. I will solve the second problem first. You need to define a class like this one:
template <typename E>
class swallow_assignment {
public:
E _value;
constexpr explicit swallow_assignment(E value) : _value(value)
{
}
template <typename Any>
constexpr const swallow_assignment& operator =(Any other) const
{
return *this;
}
constexpr operator E() const
{
return _value;
}
};
Now, you can write (swallow_assignment<E>)E::a = 1. What will happen is, at compile time, E::a will get converted to the assignable value (swallow_assignment<E>)E::a, which has the same internal representation as E::a. That value will then ignore the assignment of 1, and then will be converted back to E::a.
What's left is to prefix each of the declared constants so that we get
constexpr TypeName values[] =
{(swallow_assignment<E>)E::a = 1,
(swallow_assignment<E>)E::b = 3,
(swallow_assignment<E>)E::c = 5})
which will now be a valid initializer. This can be done with a mapping macro. I won't go into the details here, because that is a whole separate topic, but such a macro can be found here https://github.com/aantron/better-enums/blob/e28177b11a9e3d7152c5216d84fdf8939aff0eba/enum_preprocessor_map.h. Boost might also have a better one. Whatever macro you are using, I will assume that its signature is PP_MAP(prefix, __VA_ARGS__). The sketch for the final macro definition for the whole enum then becomes:
#define ENUM_CLASS(TypeName, __VA_ARGS__) \
enum class TypeName { __VA_ARGS__ }; \
constexpr TypeName values[] = \
{ PP_MAP((swallow_assignment<TypeName>)TypeName::, \
__VA_ARGS__) }; \
constexpr size_t count = sizeof(values) / sizeof(TypeName);
You will probably want to stuff these definitions into a specialization of a traits type, so that you can use this macro with more than one enum class (otherwise the arrays named values will collide). You may have to use weak symbols to avoid linking problems if you make values a static member of a traits class, however.
These last points are left as an exercise, because this answer is already way too long :) I have a library which does all of the above, though it wraps an enum instead of providing a traits specialization for an enum class. There is an unpublished branch with a combination of enum class/traits, however. You can see the library here: http://aantron.github.io/better-enums. The library's ::_from_integral() method corresponds to the to_enum function in your question, and it does both run-time and compile-time conversions.

Check sizes are the same when compiling

I have a dictionary and an array neither of which change size during the program but could often be extended pre compilation. The number of keys and the array length should always be the same size. Is there a way to check this when compiling as it'd be easy to add the key but not to the array or visa versa?
There are ways to check compile time constants. In C++11 it has been cemented with static_assert but it's possible with templates as well.
For example given:
enum Key {
K_Zero,
K_One,
K_Two,
K_NUMBER_ELEMENTS
};
static char const Dictionary[] = { ... };
You would do it C++11:
static_assert(K_NUMBER_ELEMENTS == ARRAY_SIZE(Dictionary),
"Keys / Dictionary mismatch");
Where ARRAY_SIZE is defined as:
template <typename T, unsigned N>
char (&ComputeArraySize(T (&)[N]))[N];
#define ARRAY_SIZE(Array) sizeof(ComputeArraySize(Array))
If you are still in C++03 (which is probably the case if you don't know the version), then you ought to be a little more clever and replace the static_assert with:
template <unsigned M, unsigned N> struct mp_equal;
template <unsigned N> struct mp_equal<N,N> {};
namespace {
mp_equal<K_NUMBER_ELEMENTS, ARRAY_SIZE(Dictionary)>
AssertKeysAndDictionarySizeMatch = {};
}
Which will trigger a compile time error if they do not match.
Assuming by dictionary you mean map or unordered_map there's no immediate way to do it at compile time. You could runtime assert in main OR you could force the map to be always initialized from an array of pairs, and then static_assert that the length of the pair array is the same as your main array.