CPP Tuples

A Nim wrapper for C++'s std::tuple, the variadic class. However, it is not wrapped as a type with variadic generic parameters (because user-defined variadic generic types are currently not supported in Nim), but as a type with no generic parameter. The actual type is inferred by the C++ compiler.

Example

import cpptuples

proc main =
  var x = initCPPTuple(cint(1), cfloat(2), clong(3))
  assert get[cint](x) == 1

main()

This creates a CPPTuple. The constructor takes any number of arguments.

If the tuple contains two or more cint, you won't be able to access the cint element with the above syntax, so this module provides a new syntax:

import cpptuples

proc main =
  var x = initCPPTuple(cint(10), cint(20))
  assert x[cint, 0] == 10

main()

The index (csize_t, starts from 0) must be known at the compile-time. In other words, it must be static. If there's no index given (e.g. x[cint]), it's equivalent to get[cint](x).

Note

• This can only be used within a procedure or function. Initializing a global CPPTuple variable is NOT allowed.

• This module relies on the type-inferring feature of C++. For some reason it only works with C++17 or later (Nim currently uses C++14), so if you're with a backend compiler other than GCC or Clang, you have to pass a flag to it (with --passC), telling it to use the new standard. For GCC and Clang, it is automatically done by this module (It passes -std=gnu++17), so you don't have to worry about that.

• In a project that uses this module, the {.register.} pragma cannot be used, as it's incompatible with C++17.

• Please always use types from C or C++ (like cint) as element types for CPPTuple. Using a Nim type (e.g. int64) for it may lead to incorrect code generation (and compile-time errors). If you really need a 64-bit integer, use C++'s std::int64_t from <cstdint>. For Nim's int, use std::intptr_t instead, as it always has the same size as a pointer. Never use GC'ed types (like string or seq) as an element.