What is correct regex pattern to parse HTTP's digest authentication?

Question

I have a program to parse HTTP digest's components like this:

#include "stdafx.h"
#include <iostream>
#include <string>
#include <regex>
#include <unordered_map>

int main()
{
    std::string nsInput = R"(Digest realm = "http-auth@example.org",
        qop= " auth, auth-int ", algorithm = MD5 ,
        nonce ="7ypf/xlj9XXwfDPEoM4URrv/xwf94BcCAzFZH4GiTo0v"    ,
        opaque="FQhe/qaU925kfnzjCev0ciny7QMkPqMAFRtzCUYo5tdS"
    )";
    //  Spaces are inserted into some places of the input intentionally

    std::smatch mat_opt, mat_val;
    std::unordered_map<std::string, std::string> mapDigest;

    try {
        std::regex rex_opt(R"(\s*([A-Za-z]{3,})\s*=)");
        std::regex rex_val(R"(\s*\"\s*(.{3,})\s*\"|\s*(.{3,})\s*,)");

        auto& str = nsInput;
        while (std::regex_search(nsInput, mat_opt, rex_opt))
        {
            if (mat_opt.size() >= 2) {
                auto& field = mat_opt[1].str();
                std::string& next = mat_opt.suffix().str();

                if (std::regex_search(next, mat_val, rex_val) && mat_val.size() >= 2) {
                    auto& value = mat_val[1].str();
                    mapDigest[field] = value;
                }

                str = mat_opt.suffix().str();
            }
        }

        for (auto& itr : mapDigest) {
            std::cout << itr.first << ":" << itr.second << ".\n";
        }
    }
    catch (std::regex_error& e) {
        std::cout << "regex_search failed" << e.what() << "\n";
    }

    return 0;
}

The output:

nonce:7ypf/xlj9XXwfDPEoM4URrv/xwf94BcCAzFZH4GiTo0v.
realm:http-auth@example.org.
qop:auth, auth-int .
algorithm:.
opaque:FQhe/qaU925kfnzjCev0ciny7QMkPqMAFRtzCUYo5tdS.

What I am trying to solve are:

1) The spaces are still appeared at the end of "qop"'s value.

2) The value of "algorithm" can't be matched.

May someone shine the obscure cause and how to fix it?

Thanks

Answer 1

First, your code cannot compile because you are trying to bind a non-const lvalue reference to a temporary object in the following lines:

// ...
auto& field = mat_opt[1].str();
// ...
std::string& next = mat_opt.suffix().str();
// ...
auto& value = mat_val[1].str();
// ...

I recommend to remove the reference, and use auto or std::string instead. Because of RVO, it has little performance loss.

To remove the spaces at the end of the values, you can use .{3,}? instead of .{3,} in your regex pattern. .{3,} without ? will match greedily, thus will match all characters followed (including white spaces).

The string MD5 is matched by the second parenthesis in your regex pattern, so you should access it by mat_val[2] instead of mat_val[1]. You can use conditional expression as follows:

auto value = mat_val[1].matched ? mat_val[1].str() : mat_val[2].str();

BTW, since you are using raw string literal, you needn't write an extra \ before the character " in your regex pattern.

Answer 2

As stated by others, regex may not be the weapon of choice to parse HTTP digests.

Nevertheless, I found the pattern challenging. What makes it harder than necessary is the fact that you have separators in quotes that should be ignored (in the qop-part). Your other problems stem from greedy matches (e.g. the {3,}-part).

Anyway, this is what I got after 15 minutes:

=\s*((?:[^,"]|"\s*([^"]*?)\s?")+?)(?=\s*,|$)

Demo

Update: I went the extra mile - just to prove my point.

#include <iostream>
#include <string>
#include <regex>
#include <unordered_map>

int main()
{
    std::string nsInput = R"(Digest realm = "http-auth@example.org",
        qop= " auth, auth-int ", algorithm = MD5 ,
        nonce ="7ypf/xlj9XXwfDPEoM4URrv/xwf94BcCAzFZH4GiTo0v"    ,
        opaque="FQhe/qaU925kfnzjCev0ciny7QMkPqMAFRtzCUYo5tdS"
    )";
    //  Spaces are inserted into some places of the input intentionally

    std::smatch mat_opt, mat_val;
    std::unordered_map<std::string, std::string> mapDigest;

    try {
        std::regex rex_opt(R"(\s*([A-Za-z]{3,})\s*=)");
        std::regex rex_val("=\\s*((?:[^,\"]|\"\\s*([^\"]*?)\\s?\")+?)(?=\\s*,|$)");

        auto& str = nsInput;
        while (std::regex_search(nsInput, mat_opt, rex_opt))
        {
            if (mat_opt.size() >= 2) {
                auto field = mat_opt[1].str();

                if (std::regex_search(nsInput, mat_val, rex_val)) {
                    auto value = mat_val[2].matched ? mat_val[2].str() : mat_val[1].str();
                    mapDigest[field] = value;
                }

                str = mat_opt.suffix().str();
            }
        }

        for (auto& itr : mapDigest) {
            std::cout << itr.first << ":" << itr.second << ".\n";
        }
    }
    catch (std::regex_error& e) {
        std::cout << "regex_search failed" << e.what() << "\n";
    }

    return 0;
}

Output:

opaque:FQhe/qaU925kfnzjCev0ciny7QMkPqMAFRtzCUYo5tdS.
nonce:7ypf/xlj9XXwfDPEoM4URrv/xwf94BcCAzFZH4GiTo0v.
algorithm:MD5.
realm:http-auth@example.org.
qop:auth, auth-int.