有没有一种干净的方法来使用范围拆分逗号和空格分隔的单词？

Question

I'm actually adapting this one from a deleted question from yesterday .

Input: a string containing comma separated words with space after the commas.

Output: the words.

By "clean" I mean something that doesn't rely on storing temporaries into variables which wouldn't have a reason to exist, based on the intent of the code.

My attempt

#include <iostream>
#include <range/v3/view/split.hpp>
#include <range/v3/view/split_when.hpp>
#include <range/v3/view/transform.hpp>
#include <range/v3/view/trim.hpp>
#include <string>
#include <vector>

auto constexpr is = [](char c){
    return [c](char c_){ return c_ == c; };
};
using namespace ranges::views;

int main()
{
  std::string ss{"cccccciao"};
  // correctly prints [i,a,o]
  std::cout << (ss | trim(is('c'))) << std::endl;

  std::string s = "blue, green, red";
  std::cout << split_when(s, is(',')) << std::endl;

  std::vector<std::string> vs;
  vs.push_back(ss);
  vs.push_back(ss);
  // correctly prints [[i,a,o],[i,a,o]]
  std::cout << (vs | transform(trim(is('c')))) << std::endl;

  // doesn't work
  //auto result0 = split_when(s, is(',')) | transform(trim(is(' ')));
  //auto result1 = transform(split_when(s, is(',')), trim(is(' ')));

  // doesn't work either
  auto sss = split_when(s, [](char c){ return c == ','; });
  //auto result2 = transform(sss, trim([](char c){ return c == ' '; }));
  //auto result3 = sss | transform(trim([](char c){ return c == ' '; }));
}

Based on the principle of least astonishment I expected that something like this oneliner should work,

auto result0 = s | split_when(is(',')) | transform(trim(is(' ')));

but it doesn't.

Answer 1

split can split on a pattern that is longer than one element. The one caveat is that ", " , considered as a range, is a pattern of three elements (including the terminating null character) - and in any event the adapter decays it to a pointer, which makes it not-a-range. So:

auto result = input | rv::split(", "sv);