Auto and Structured Binding

Posted on by Lucky Resistor

This article is just a short follow-up article for “Write Less Code using the ‘auto’ Keyword”. Structured binding is something handy, introduced in C++17. Therefore, only the latest compiler will support it.

If you are mainly write embedded code, you may skip this article, because it will take some years until C++17 support is available for your platform. If you also write desktop code, read on, you may like this feature.

Find a detailed reference about structural binding here. The following sections will explain the feature using several use cases.

Multiple Return Values using Tuples

Let us write a function to convert a string into an integer. For error handling, we like to use parameters or return values, but no exceptions. You may consider these variants:

uint32_t getInt(const std::string &text, bool *ok);
uint32_t getInt(const std::string &text, bool &ok);
bool getInt(const std::string &text, uint32_t &result);

The problem is, if you like to use the full range of the uint32_t, you need a second value to report of the conversion was successful or not. It may be a bool, or even better an enum to report this result.

#include <iostream>
#include <cstdint>
#include <string>
#include <limits>

uint32_t safeMultiplication(uint32_t a, uint32_t b, bool *ok = nullptr) {
    uint32_t result;
    if (__builtin_mul_overflow(a, b, &result)) {
        if (ok != nullptr) {
            *ok = false;
        }
        return 0;
    }
    if (ok != nullptr) {
        *ok = true;
    }
    return result;
}

uint32_t safeAdd(uint32_t a, uint32_t b, bool *ok = nullptr) {
    uint32_t result;
    if (__builtin_add_overflow(a, b, &result)) {
        if (ok != nullptr) {
            *ok = false;
        }
        return 0;
    }
    if (ok != nullptr) {
        *ok = true;
    }
    return result;
}

uint32_t getInt(const std::string &text, bool *ok = nullptr) {
    uint32_t result = 0;
    uint32_t factor = 1;
    bool factorValid = true;
    for (auto it = text.rbegin(); it != text.rend(); ++it) {
        if (*it < '0' || *it > '9') {
            if (ok != nullptr) {
                *ok = false;
            }
            return 0;
        }
        auto value = static_cast<uint32_t>(*it - '0');
        if (value > 0) {
            if (!factorValid) {
                if (ok != nullptr) {
                    *ok = false;
                }
                return 0;
            }
            bool multiplicationOk;
            value = safeMultiplication(factor, value, &multiplicationOk);
            if (!multiplicationOk) {
                if (ok != nullptr) {
                    *ok = false;
                }
                return 0;
            }
            bool additionOk;
            result = safeAdd(result, value, &additionOk);
            if (!additionOk) {
                if (ok != nullptr) {
                    *ok = false;
                }
                return 0;
            }
        }
        if (factorValid) {
            bool multiplicationOk;
            factor = safeMultiplication(factor, 10u, &multiplicationOk);
            if (!multiplicationOk) {
                factorValid = false;
            }
        }
    }
    if (ok != nullptr) {
        *ok = true;
    }
    return result;
}

void testText(const std::string &text) {
    bool ok;
    const auto number = getInt(text, &ok);
    std::cout << text << " -> " << number << " ok: " << (ok ? "yes" : "no") << std::endl;
}

int main(int argc, const char * argv[]) {
    testText("text");
    testText("4294967295");
    testText("4294967296");
    testText("00004294967295");
    testText("x123");
    return 0;
}

The example code works well. For simplicity, it uses two built-in functions __builtin_add_overflow and __builtin_mul_overflow from the GCC and clang compiler.

Using two return values would simplify the interface. It can be done using std::tuple and structured binding.

#include <iostream>
#include <cstdint>
#include <string>
#include <limits>
#include <tuple>

using Result = std::tuple<uint32_t, bool>;

Result safeMultiplication(uint32_t a, uint32_t b) {
    uint32_t result;
    if (__builtin_mul_overflow(a, b, &result)) {
        return std::make_tuple(0, false);
    }
    return std::make_tuple(result, true);
}

Result safeAdd(uint32_t a, uint32_t b) {
    uint32_t result;
    if (__builtin_add_overflow(a, b, &result)) {
        return std::make_tuple(0, false);
    }
    return std::make_tuple(result, true);
}

Result getInt(const std::string &text) {
    uint32_t result = 0;
    uint32_t factor = 1;
    bool factorValid = true;
    for (auto it = text.rbegin(); it != text.rend(); ++it) {
        if (*it < '0' || *it > '9') {
            return std::make_tuple(0, false);
        }
        const auto digitValue = static_cast<uint32_t>(*it - '0');
        if (digitValue > 0) {
            if (!factorValid) {
                return std::make_tuple(0, false);
            }
            const auto[value, multiplicationOk] = safeMultiplication(factor, digitValue);
            if (!multiplicationOk) {
                return std::make_tuple(0, false);
            }
            bool additionOk;
            std::tie(result, additionOk) = safeAdd(result, value);
            if (!additionOk) {
                return std::make_tuple(0, false);
            }
        }
        if (factorValid) {
            bool multiplicationOk;
            std::tie(factor, multiplicationOk) = safeMultiplication(factor, 10u);
            if (!multiplicationOk) {
                factorValid = false;
            }
        }
    }
    return std::make_tuple(result, true);
}

void testText(const std::string &text) {
    const auto[number, ok] = getInt(text);
    std::cout << text << " -> " << number << " ok: " << (ok ? "yes" : "no") << std::endl;
}

int main(int argc, const char * argv[]) {
    testText("text");
    testText("4294967295");
    testText("4294967296");
    testText("00004294967295");
    testText("x123");
    return 0;
}

It simplified the code with just a single downside. While the initial implementation made the ok optional, the new version will get it for each result.

Bind Elements of a Struct

You can bind to members of any struct using the same syntax:

struct Result {
    uint32_t value;
    bool ok;
}

Result result = {10, true}

int main() {
    auto[value, ok] = result;
}

If we use a struct in our example, it simplifies several lines, but requires some additions elsewhere.

#include <iostream>
#include <cstdint>
#include <string>
#include <limits>
#include <tuple>

struct Result {
    uint32_t value;
    bool ok;
    constexpr static Result error() { return {0, false}; }
    constexpr static Result success(uint32_t value) { return {value, true}; }
};

Result safeMultiplication(uint32_t a, uint32_t b) {
    uint32_t result;
    if (__builtin_mul_overflow(a, b, &result)) {
        return Result::error();
    }
    return Result::success(result);
}

Result safeAdd(uint32_t a, uint32_t b) {
    uint32_t result;
    if (__builtin_add_overflow(a, b, &result)) {
        return Result::error();
    }
    return Result::success(result);
}

Result getInt(const std::string &text) {
    uint32_t result = 0;
    uint32_t factor = 1;
    bool factorValid = true;
    for (auto it = text.rbegin(); it != text.rend(); ++it) {
        if (*it < '0' || *it > '9') {
            return Result::error();
        }
        const auto digitValue = static_cast<uint32_t>(*it - '0');
        if (digitValue > 0) {
            if (!factorValid) {
                return Result::error();
            }
            const auto[value, multiplicationOk] = safeMultiplication(factor, digitValue);
            if (!multiplicationOk) {
                return Result::error();
            }
            const auto[immediateResult, additionOk] = safeAdd(result, value);
            if (!additionOk) {
                return Result::error();
            }
            result = immediateResult;
        }
        if (factorValid) {
            const auto[newFactor, multiplicationOk] = safeMultiplication(factor, 10u);
            if (!multiplicationOk) {
                factorValid = false;
            }
            factor = newFactor;
        }
    }
    return Result::success(result);
}

void testText(const std::string &text) {
    const auto[number, ok] = getInt(text);
    std::cout << text << " -> " << number << " ok: " << (ok ? "yes" : "no") << std::endl;
}

int main(int argc, const char * argv[]) {
    testText("text");
    testText("4294967295");
    testText("4294967296");
    testText("00004294967295");
    testText("x123");
    return 0;
}

The struct made the status optional, because we can write a line like this:

const auto value = getInt(text).value;

Or only check if the value is valid:

if (getInt(text).ok) {
    // ...
}

Binding to an Array

Binding values from an array is another interesting use:

#include <iostream>

int table[4][4] = {
    { 1,  2,  3,  4},
    { 5,  6,  7,  8},
    { 9, 10, 11, 12},
    {13, 14, 15, 16},
};

int main(int argc, const char * argv[]) {
    for (const auto &row : table) {
        const auto [a, b, c, d] = row;
        std::cout << a << ", " << b << ", " << c << ", " << d << std::endl;
    }
    return 0;
}

References

I found structural binding also useful creating references to values for actual modification, or const references to avoid copy operations.

#include <iostream>

int table[4][4] = {
    { 1,  2,  3,  4},
    { 5,  6,  7,  8},
    { 9, 10, 11, 12},
    {13, 14, 15, 16},
};

int main(int argc, const char * argv[]) {
    for (auto &row : table) {
        auto& [a, b, c, d] = row;
        std::swap(a, d);
        std::swap(b, c);
    }
    for (const auto &row : table) {
        const auto& [a, b, c, d] = row;
        std::cout << a << ", " << b << ", " << c << ", " << d << std::endl;
    }
    return 0;
}

Conclusion

If you already use a C++17 capable compiler, like the GCC, clang or Visual Studio, try to use structural binding in your project. You can find a table with the current compiler support in this document.

  • It can create cleaner interfaces, where all values are returned with the return statement.
  • Returned objects can be directly split into variables.
  • Working with std::tuple values is straightforward.
  • Array tables can be accessed in a column like way.

If you have questions, miss some information or have any feedback, feel free to add a comment below.

Learn More

Use Enum with More Class!

Use Enum with More Class!

You may be familiar with enum values, but do you know about enum classes? This great feature was introduced with C++11 to solve several problems with the regular enum declaration. Now I will explain the ...
Read More
Event-based Firmware (Part 1/2)

Event-based Firmware (Part 1/2)

You start with small, simple firmware. But with each added feature, the complexity grows and grows. Soon, you need a good design in order to maintain the firmware and ensure the code remains clean and ...
Read More
Bit Manipulation using Templates

Bit Manipulation using Templates

Did you read my article about C++ templates for embedded code? You learned how to use function templates. This post adds a practical example to this topic. Bit Manipulation You may be familiar with bit ...
Read More
Write Less Code using the "auto" Keyword

Write Less Code using the “auto” Keyword

The auto keyword was introduced with C++11. It reduces the amount of code you have to write, reduces repetitive code and the number of required changes. Sadly, many C++ developers are not aware of how ...
Read More
Class or Module for Singletons?

Class or Module for Singletons?

Should you use a class or a module with a namespace for a singleton interface in your firmware? I found there are many misunderstandings which lead beginners to make a wrong decision in this matter ...
Read More
Make your Code Safe and Readable with Flags

Make your Code Safe and Readable with Flags

Flags play an important role in embedded software development. Microcontrollers and chips are using registers where single bits or combinations of bits play a big role in the configuration. All the bits and their role ...
Read More

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: