i am writing a small interpreter in c++, but it returns me an error "vector subscript out of range"

Question

I am writing a very simple interpreter with these two commands: "print:value", to print and "printnl:value", to print in a new line and i use vectors to store the command and the value. the problem is that when i input something it returns ma an error "vector subscript out of range" and some random exit code. Here is the code:

#include <iostream>
#include <string>
#include <vector>
using namespace std;

vector<string> brk_token(string text, char breakon) {
    vector<string> tokens;
    string token = "";
    for (int i = 0;i < text.size(); i++) {
        if (text.at(i) != breakon) {
            token += text.at(i);
        }
        else {
            tokens.push_back(token);
            token = "";
        }


    }
    return tokens;
}

int main() {

    string program;
    while (true) {
        getline(cin, program);
        if (brk_token(program.c_str(), ':')[0] == "print") {
            cout << brk_token(program.c_str(), ':')[1];
        }
        else if (brk_token(program.c_str(), ':')[0] == "printnl") {
            cout << brk_token(program.c_str(), ':')[1] << "\n";
        }
        if (program == "break") {
            break;
        }
    }

    return 0;
}

I wrote: "print:hello" and "printnl:hello" and every time it returned that error, i wrote some random characters and the result is the same, can someone help?

Answer 1

Inside of brk_token(), when the loop reaches the end of the string, you are not pushing the final token into the vector. As such, inside of main(), you have undefined behavior when you try to access the vector[1] element that doesn't exist.

Also, some other issues with your code:

• the string::at() method performs bounds checking at runtime, but you are doing your own bounds checking in the for loop that iterates through a string, so using at() is useless overhead. You can safely use string::operator[] instead.

• your while loop in main() is not checking whether getline() is successful or not before processing the string.

• when calling brk_token(), you are converting the string into a const char* just to convert it back to a string again. You don't need to use string::c_str() in this code at all.

• calling brk_token() repeatedly is overkill. Call it one time per string and save the result into a local variable that you can then validate before using.

With all of that said, try this instead:

#include <iostream>
#include <string>
#include <vector>
using namespace std;

vector<string> brk_token(const string &text, char breakon) {
    vector<string> tokens;
    string token;
    for (size_t i = 0; i < text.size(); ++i) {
        if (text[i] != breakon) {
            token += text[i];
        }
        else {
            tokens.push_back(token);
            token = "";
        }
    }

    if (!token.empty())
        tokens.push_back(token);

    return tokens;
}

int main() {

    string program;
    vector<string> tokens;

    while (getline(cin, program) && (program != "break")) {
        tokens = brk_token(program, ':');
        if (tokens.size() != 2) {
            cerr << "Bad input\n";
        }
        else if (tokens[0] == "print") {
            cout << tokens[1];
        }
        else if (tokens[0] == "printnl") {
            cout << tokens[1] << '\n';
        }
        else {
            cerr << "Unknown input\n";
        }
    }

    return 0;
}