Question

Suppose I want to read lines from the console and put those into a container until the user enters a blank line. I do not want that blank line ending up in my container, though. I can think of five different solutions:

a) break from loop

std::vector<std::string> container;
for (; ;)
{
    std::string line = get_input();
    if (line.empty()) break;
    container.push_back(line);
}

b) read before loop and inside loop

std::vector<std::string> container;
std::string line = get_input();
while (!line.empty())
{
    container.push_back(line);
    line = get_input();
}

c) read as part of loop condition, assignment version

std::vector<std::string> container;
std::string line;
while (!(line = get_input()).empty())
{
    container.push_back(line);
}

d) read as part of loop condition, sequence version

std::vector<std::string> container;
std::string line;
while (line = get_input(), !line.empty())
{
    container.push_back(line);
}

e) read too much, remove it after loop

std::vector<std::string> container;
std::string line;
do
{
    line = get_input();
    container.push_back(line);
}
while (!line.empty());
container.pop_back();

So, which solution would you prefer, and why? Which would be the easiest to understand for a beginner?

Answer 1

I prefer (a). simple and reads quite naturally.

(b) repeats the line that gets the input.

(c) and (d) both use syntax that may be confusing to beginners (specifically, the comma not within a for statement or a definition, and assignment within a conditional). I’d probably prefer (c) over (d) though.

(e) is… inefficient. What if that last push_back caused a reallocation?