Chapter 5: The Standard Library (C++20)_《C++20Get the details》_notes

Chapter 5: The Standard Library

• • • [5.1 The Ranges Library](#5.1 The Ranges Library)
    • [5.2 `std::span`](#5.2 std::span)
    • [5.3 Container Improvements](#5.3 Container Improvements)
    • [5.4 Arithmetic Utilities](#5.4 Arithmetic Utilities)
    • [5.5 Formatting Library](#5.5 Formatting Library)
    • [5.6 Calendar & Time Zones](#5.6 Calendar & Time Zones)
    • [5.7 Further Improvements](#5.7 Further Improvements)
    • [Compilation Notes](#Compilation Notes)
    • [Multiple-Choice Questions](#Multiple-Choice Questions)
    • [**Answers & Explanations**](#Answers & Explanations)
    • [Design Challenges (C++20 Standard Library & Concurrency)](#Design Challenges (C++20 Standard Library & Concurrency))
    • [Answer Key & Detailed Solutions](#Answer Key & Detailed Solutions)
    • Summary

---

5.1 The Ranges Library

Key Concepts

1. Ranges : Replace iterator pairs with contiguous sequences (containers, views).
2. Views : Non-owning, lazily evaluated ranges (e.g., filter, transform).
3. Range Adaptors : Pipelined operations using | operator.
4. Lazy Evaluation: Operations are deferred until iteration.

Code Example: Filter & Transform

#include <iostream>
#include <vector>
#include <ranges>

int main() {
    std::vector<int> nums{1, 2, 3, 4, 5};
    
    // Define a pipeline: filter even numbers, then square them
    auto result = nums 
        | std::views::filter([](int n) { return n % 2 == 0; })
        | std::views::transform([](int n) { return n * n; });
    
    for (int v : result) {
        std::cout << v << " "; // Output: 4 16
    }
}

• Explanation :
  • filter retains even numbers.
  • transform squares each element.
  • Operations are lazily evaluated when iterated.

---

5.2 std::span

Key Concepts

1. Non-Owowing View: Represents a contiguous sequence (array, vector).
2. Static vs. Dynamic Extent : Fixed size (std::span<int, 5>) or runtime size (std::span<int>).
3. Safety : Bounds checking (optional via at() in debug builds).

Code Example: Modifying Elements

#include <iostream>
#include <span>
#include <vector>

void print(std::span<int> s) {
    for (int v : s) {
        std::cout << v << " ";
    }
}

int main() {
    std::vector<int> vec{1, 2, 3, 4, 5};
    std::span<int> s(vec);
    
    s[2] = 99; // Modify the third element
    print(s);  // Output: 1 2 99 4 5
}

• Explanation :
  • span references vec's data.
  • Modifications via span affect the original container.

---

5.3 Container Improvements

Key Features

1. contains() for Associative Containers:

   #include <iostream>
   #include <set>
   
   int main() {
       std::set<int> s{1, 2, 3};
       if (s.contains(2)) {
           std::cout << "Found!"; // Output: Found!
       }
   }

   • Replaces s.find(2) != s.end().

2. std::erase/erase_if for Sequence Containers:

   #include <vector>
   #include <algorithm>
   
   int main() {
       std::vector<int> v{1, 2, 3, 4, 5};
       std::erase_if(v, [](int n) { return n % 2 == 0; });
       // v becomes [1, 3, 5]
   }

---

5.4 Arithmetic Utilities

Safe Integer Comparisons

#include <utility>
#include <iostream>

int main() {
    int a = -1;
    unsigned b = 0;
    
    // Safe comparison avoids integer overflow
    if (std::cmp_less(a, b)) {
        std::cout << "-1 < 0 (safely compared)";
    }
}

• Explanation :
  • std::cmp_less safely compares signed/unsigned values.

Math Constants

#include <numbers>
#include <iostream>

int main() {
    std::cout << std::numbers::pi; // Output: 3.14159...
}

---

5.5 Formatting Library

std::format Basics

#include <format>
#include <iostream>

int main() {
    std::cout << std::format("Hello, {}! π ≈ {:.2f}", "C++20", 3.14159);
    // Output: Hello, C++20! π ≈ 3.14
}

Custom Type Formatting

#include <format>
#include <iostream>

struct Point { int x, y; };

template <>
struct std::formatter<Point> {
    auto parse(auto& ctx) { return ctx.begin(); }
    auto format(const Point& p, auto& ctx) {
        return std::format_to(ctx.out(), "({}, {})", p.x, p.y);
    }
};

int main() {
    Point p{10, 20};
    std::cout << std::format("{}", p); // Output: (10, 20)
}

---

5.6 Calendar & Time Zones

Date & Time Example

#include <chrono>
#include <iostream>

using namespace std::chrono;

int main() {
    auto now = system_clock::now();
    std::cout << "Current time: " << now << "\n"; // ISO 8601 format

    auto date = 2023y/September/15;
    std::cout << "Date: " << date << "\n"; // Output: 2023-09-15
}

Time Zone Conversion

#include <chrono>
#include <iostream>

int main() {
    auto utc_time = std::chrono::system_clock::now();
    auto ny_tz = std::chrono::locate_zone("America/New_York");
    auto ny_time = std::chrono::zoned_time(ny_tz, utc_time);
    std::cout << "NY time: " << ny_time << "\n";
}

---

5.7 Further Improvements

std::bind_front

#include <functional>
#include <iostream>

void print(int a, int b) { std::cout << a << ", " << b << "\n"; }

int main() {
    auto f = std::bind_front(print, 42);
    f(10); // Output: 42, 10
}

std::is_constant_evaluated

#include <iostream>

constexpr int compute() {
    if (std::is_constant_evaluated()) {
        return 42; // Compile-time
    } else {
        return 0;  // Runtime
    }
}

int main() {
    constexpr int v1 = compute(); // 42
    int v2 = compute();           // 0
    std::cout << v1 << " " << v2; // Output: 42 0
}

---

Compilation Notes

• Use -std=c++20 (GCC/Clang) or /std:c++latest (MSVC).
• Link against <ranges>, <chrono>, and other C++20 headers.
• Ensure your compiler fully supports C++20 features (e.g., GCC 10+, Clang 12+).

---

Multiple-Choice Questions

---

Questions 1-25

1. Which statements about std::ranges::views are true?

A) Views own the data they operate on.

B) Views are lazily evaluated.

C) std::views::transform modifies the original container.

D) Views can be composed using the | operator.

2. What is true about std::span?

A) It owns the underlying data.

B) It can represent a runtime-sized range.

C) std::span<int, 5> has a static extent.

D) Accessing elements with operator[] always performs bounds checking.

3. Which C++20 features allow safe integer comparisons?

A) std::cmp_less

B) std::ssize

C) std::midpoint

D) std::in_range

4. Select valid ways to use std::format:

A) std::format("{}", 3.14)

B) std::format("{:.2f}", 3.14159)

C) std::format("{:%Y-%m-%d}", std::chrono::system_clock::now())

D) std::format("{0} {0}", "C++")

5. Which operations are valid for std::jthread?

A) Automatically joins on destruction.

B) Supports cooperative interruption.

C) Can replace std::thread without code changes.

D) Requires manual join() calls.

6. What applies to std::semaphore?

A) Supports timed waits.

B) Maximum count is fixed at compile time.

C) acquire() decrements the counter.

D) release(n) increments the counter by n.

7. Which std::chrono types are valid?

A) std::chrono::day

B) std::chrono::time_of_day

C) std::chrono::year_month_weekday

D) std::chrono::time_zone

8. What describes std::is_constant_evaluated()?

A) Returns true during compile-time evaluation.

B) Used to optimize runtime performance.

C) Valid only in constexpr functions.

D) Requires the <type_traits> header.

9. Which statements about std::ranges algorithms are true?

A) They accept iterator pairs.

B) They always return iterators.

C) std::ranges::sort can sort a std::vector directly.

D) They require explicit projection parameters.

10. What is true about std::bind_front?

A) Binds arguments to the front of a callable.

B) Supports partial function application.

C) Requires <functional> header.

D) Replaces std::bind entirely.

11. Which operations are valid for std::atomic<std::shared_ptr<T>>?

A) load()

B) store()

C) exchange()

D) wait()

12. What applies to std::latch?

A) Counts down to zero.

B) Reusable after reaching zero.

C) Threads block until the count reaches zero.

D) Supports arrive_and_wait().

13. Which are valid C++20 container improvements?

A) std::vector::contains()

B) std::erase_if(std::list, predicate)

C) std::map::merge()

D) std::string::starts_with()

14. What describes std::source_location?

A) Captures __FILE__ and __LINE__ at compile time.

B) Requires C++20 or later.

C) Defaults to the call site when used as a default argument.

D) Replaces the assert macro.

15. Which statements about coroutines are true?

A) co_await suspends execution.

B) co_return must return a value.

C) Coroutines require heap allocation.

D) co_yield resumes execution immediately.

16. What applies to std::format for user-defined types?

A) Requires specializing std::formatter<T>.

B) Uses format_as() function.

C) Supports custom format specifiers.

D) Requires inheritance from std::formattable.

17. Which are valid std::chrono literals?

A) 10ns

B) 3h

C) 2023_y

D) Sunday

18. What is true about std::jthread interruption?

A) Uses request_stop().

B) Checks get_stop_token().stop_requested().

C) Automatically interrupts running threads.

D) Requires std::stop_callback.

19. Which statements about std::span are true?

A) std::span<int> has a dynamic extent.

B) subspan() returns a smaller span.

C) size_bytes() returns the size in bytes.

D) Cannot be constructed from a std::array.

20. What describes std::ranges::to?

A) Converts a range to a container.

B) Requires explicit template arguments.

C) Part of C++20.

D) Works with all standard containers.

21. Which are valid uses of std::views?

A) std::views::iota(0, 10)

B) std::views::split("hello", 'l')

C) std::views::reverse

D) std::views::common

22. What applies to std::atomic_ref?

A) Provides atomic access to non-atomic objects.

B) Requires aligned memory.

C) Lifetime must not exceed the referenced object.

D) Supports all arithmetic types.

23. Which statements about std::barrier are true?

A) Reusable after completion.

B) Supports a completion function.

C) Threads block until all arrive.

D) Part of <barrier> header.

24. What is true about std::ranges::join?

A) Flattens a range of ranges.

B) Eagerly evaluates the input range.

C) Requires a bidirectional range.

D) Returns a view.

25. Which describe std::expected (C++23 preview)?

A) Represents a value or an error.

B) Part of C++20.

C) Similar to std::optional with error handling.

D) Requires <expected> header.

---

Answers & Explanations

1. B, D

• Views are non-owning and lazy. Composition uses |.

2. B, C

• std::span can have dynamic or static extent. Bounds checking is debug-only.

3. A, D

• std::cmp_less and std::in_range handle mixed signedness safely.

4. A, B, C, D

• All are valid: floating-point, date formatting, positional arguments.

5. A, B

• std::jthread auto-joins and supports stop tokens.

6. A, C, D

• Semaphores allow timed waits, acquire() reduces count, release(n) adds n.

7. A, C, D

• time_of_day is valid; time_zone is part of the chrono library.

8. A, B

• Returns true during compile-time evaluation. Usable in non-constexpr contexts.

9. A, C

• Ranges accept containers directly. Algorithms like sort work on ranges.

10. A, B, C

• bind_front binds leading arguments; requires <functional>.

11. A, B, C, D

• Atomic smart pointers support all atomic operations.

12. A, C, D

• Latches count down to zero, block until zero, and are single-use.

13. B, D

• std::erase_if and starts_with are C++20 additions.

14. A, B, C

• source_location captures compile-time context; defaults to call site.

15. A

• co_await suspends. co_return can return void. Heap allocation is optional.

16. A, C

• Specialize std::formatter<T> and define custom specifiers.

17. A, B, C

• Valid literals: 10ns, 3h, 2023y (C++20). Sunday is a type, not a literal.

18. A, B

• Interruption uses request_stop() and stop tokens.

19. A, B, C

• subspan() and size_bytes() are valid. std::span can wrap std::array.

20. A

• std::ranges::to (C++23) converts ranges to containers. Not part of C++20.

21. A, C, D

• iota, reverse, and common are valid views. split requires a range.

22. A, B, C

• atomic_ref references non-atomic objects with alignment requirements.

23. B, C, D

• Barriers support completion functions and reuse. Part of <barrier>.

24. A, D

• join flattens ranges and returns a view. Works with forward ranges.

25. A, C

• std::expected (C++23) is value/error type. Not part of C++20.

---

Design Challenges (C++20 Standard Library & Concurrency)

---

Question 1: Ranges Pipeline Composition
Task:

Create a pipeline using std::ranges that filters even numbers from a vector, squares them, reverses the order,

and returns only the first 3 elements. Ensure all operations are lazy-evaluated.

Test Case:

Input: {1, 2, 3, 4, 5, 6, 7, 8}

Expected Output: 64 36 16

---

Question 2: Custom Range View
Task:

Implement a custom view split_by_half that divides a range into two halves. Use this view with std::ranges::join to

flatten the result. Handle odd-sized ranges by making the first half larger.

Test Case:

Input: {1, 2, 3, 4, 5} → Split → {``{1,2,3}, {4,5}} → Join → {1,2,3,4,5}

---

Question 3: std::span Dynamic Bounds
Task:

Create a std::span over a C-style array and modify elements through the span. Implement bounds checking using
span::subspan to safely access elements, throwing std::out_of_range if invalid.

Test Case:

Array: int arr[] = {10, 20, 30, 40}

Access indices 1-3 → Modify to {25, 35}. Handle index 4 with exception.

---

Question 4: constexpr Vector Sorting
Task:

Use std::vector in a constexpr context to sort integers with std::sort (C++20 constexpr support). Verify the

sorted result at compile time using static_assert.

Test Case:

Input: {5, 3, 4, 1, 2} → Sorted: {1, 2, 3, 4, 5}

---

Question 5: Atomic Smart Pointers
Task:

Implement a thread-safe shared counter using std::atomic<std::shared_ptr<int>>. Use compare_exchange_weak to

atomically increment the counter across multiple threads.

Test Case:

3 threads incrementing 1000 times each → Final value: 3000.

---

Question 6: Custom Formatter for User-Defined Type
Task:

Create a std::formatter specialization for a struct Point3D { int x, y, z; } to display as (x,y,z). Support

format specifiers for hex (%x) and decimal (%d).

Test Case:
Point3D{15, 255, 16} with format "{:%x,%d,%x}" → Output: (f,255,10).

---

Question 7: Time Zone Conversions
Task:

Convert a UTC timestamp to New York's time zone accounting for daylight saving. Use std::chrono::zoned_time and print

the local time in YYYY-MM-DD HH:MM format.

Test Case:

UTC: 2023-07-01 12:00:00 → New York: 2023-07-01 08:00.

---

Question 8: Generator Coroutine with Exception Handling
Task:

Implement a coroutine Generator<T> that yields values until an error condition (e.g., negative value). Propagate

exceptions from the coroutine to the caller using std::rethrow_exception.

Test Case:

Input: {5, 3, -1, 2} → Output: 5 3 followed by std::runtime_error.

---

Question 9: Semaphore-Based Producer-Consumer
Task:

Implement a producer-consumer queue using std::counting_semaphore. Producers block when the buffer is full; consumers

block when empty. Use std::binary_semaphore for mutex.

Test Case:

2 producers, 2 consumers, buffer size 3. Verify no data races.

---

Question 10: Latch for Parallel Initialization
Task:

Use std::latch to synchronize 4 threads initializing different parts of a shared array. Each thread signals

completion, and the main thread aggregates results after all are ready.

Test Case:

Array initialized to {1, 2, 3, 4} by 4 threads → Final array: [1,2,3,4].

---

Question 11: Barrier for Cyclic Task Processing
Task:

Use std::barrier to implement 3 threads processing batches of data in phases. After each phase, threads wait at

the barrier before proceeding to the next phase.

Test Case:

3 phases with intermediate results printed after each barrier.

---

Question 12: jthread with Cooperative Cancellation
Task:

Use std::jthread and std::stop_token to run a task that periodically checks for cancellation. Stop the thread

after 5 iterations using request_stop().

Test Case:

Output: "Iteration 1"..."Iteration 5" then stops.

---

Question 13: Synchronized Output Streams
Task:

Use std::osyncstream to synchronize output from multiple threads writing to std::cout. Ensure lines from different

threads are not interleaved.

Test Case:

4 threads each printing 100 lines. Verify no interleaving.

---

Question 14: std::is_constant_evaluated
Task:

Implement a function square() that uses std::is_constant_evaluated to return a constexpr result when possible

and a runtime result otherwise. Verify with static_assert.

Test Case:
static_assert(square(5) == 25); and runtime square(5) == 25.

---

Question 15: std::source_location Logging
Task:

Create a logging macro LOG(msg) that captures std::source_location (file, line, function). Print logs in the

format [file:line] function: msg.

Test Case:
LOG("Error") → Output: [main.cpp:42] main: Error.

---

Answer Key & Detailed Solutions

---

Question 1 Solution: Ranges Pipeline

#include <vector>
#include <ranges>
#include <iostream>

int main() {
    std::vector<int> v = {1, 2, 3, 4, 5, 6, 7, 8};
    auto pipeline = v | std::views::filter([](int x) { return x % 2 == 0; })
                     | std::views::transform([](int x) { return x * x; })
                     | std::views::reverse
                     | std::views::take(3);
    for (int x : pipeline) std::cout << x << " ";
    // Output: 64 36 16
}

Explanation:

The pipeline chains multiple range adaptors. filter removes odd numbers, transform squares, reverse reverses

order, and take selects the first 3 elements. All operations are lazy.

---

Question 2 Solution: Custom View

#include <ranges>
#include <vector>
#include <iostream>

struct split_by_half : std::ranges::view_interface<split_by_half> {
    std::vector<int> vec;
    split_by_half(std::vector<int> v) : vec(std::move(v)) {}
    auto begin() const { 
        auto mid = vec.begin() + (vec.size() + 1) / 2;
        return std::array{vec.begin(), mid, mid, vec.end()}.begin();
    }
    auto end() const { return begin() + 2; }
};

int main() {
    std::vector<int> v = {1, 2, 3, 4, 5};
    auto halves = split_by_half(v) | std::views::join;
    for (int x : halves) std::cout << x << " "; // 1 2 3 4 5
}

Explanation:

The custom view split_by_half divides the vector into two halves. std::views::join flattens the split ranges.

---

Question 3 Solution: std::span Bounds

#include <span>
#include <iostream>
#include <stdexcept>

void modify_span(std::span<int> s, size_t index, int value) {
    if (index >= s.size()) throw std::out_of_range("Invalid index");
    s.subspan(index, 1)[0] = value;
}

int main() {
    int arr[] = {10, 20, 30, 40};
    std::span s(arr);
    try {
        modify_span(s, 1, 25); // OK
        modify_span(s, 4, 50); // Throws
    } catch (const std::out_of_range& e) {
        std::cout << e.what(); // "Invalid index"
    }
    // arr becomes {10,25,30,40}
}

Explanation:
subspan creates a subview with bounds checking. Accessing index 4 throws std::out_of_range.

---

Question 4 Solution: constexpr Vector Sort

#include <vector>
#include <algorithm>
#include <cassert>

consteval std::vector<int> constexpr_sort() {
    std::vector<int> v = {5,3,4,1,2};
    std::sort(v.begin(), v.end());
    return v;
}

int main() {
    constexpr auto sorted = constexpr_sort();
    static_assert(sorted[0] == 1); // Compile-time check
    // Runtime check omitted for brevity
}

Explanation:

C++20 allows std::sort in consteval functions. The sorted vector is verified at compile time.

---

Question 5 Solution: Atomic Shared Pointer

#include <atomic>
#include <memory>
#include <thread>

struct Counter {
    std::atomic<std::shared_ptr<int>> ptr = std::make_shared<int>(0);
    void increment() {
        std::shared_ptr<int> current = ptr.load();
        std::shared_ptr<int> desired;
        do {
            desired = std::make_shared<int>(*current + 1);
        } while (!ptr.compare_exchange_weak(current, desired));
    }
};

int main() {
    Counter c;
    std::thread t1([&] { for (int i=0; i<1000; ++i) c.increment(); });
    std::thread t2([&] { for (int i=0; i<1000; ++i) c.increment(); });
    t1.join(); t2.join();
    std::cout << *c.ptr; // 2000 (assuming no data races)
}

Explanation:
compare_exchange_weak atomically updates the shared pointer. Thread safety is ensured via atomic operations.

---

Question 6 Solution: Custom Formatter

#include <format>
#include <iostream>

struct Point3D { int x, y, z; };

template<> struct std::formatter<Point3D> {
    char format = 'd';
    constexpr auto parse(std::format_parse_context& ctx) {
        auto it = ctx.begin();
        if (it != ctx.end()) format = *it++;
        return it;
    }
    auto format(const Point3D& p, std::format_context& ctx) const {
        if (format == 'x') 
            return std::format_to(ctx.out(), "({:x},{:x},{:x})", p.x, p.y, p.z);
        else
            return std::format_to(ctx.out(), "({},{},{})", p.x, p.y, p.z);
    }
};

int main() {
    Point3D p{15, 255, 16};
    std::cout << std::format("{:%x,%d,%x}", p); // (f,255,10)
}

Explanation:

The formatter specialization handles hex and decimal formatting using a custom specifier.

---

Question 7 Solution: Time Zone Conversion

#include <chrono>
#include <format>
#include <iostream>

int main() {
    using namespace std::chrono;
    auto utc_time = sys_days{2023y/July/1} + 12h;
    zoned_time ny_time{"America/New_York", utc_time};
    std::cout << std::format("{:%Y-%m-%d %H:%M}", ny_time); // 2023-07-01 08:00
}

Explanation:
zoned_time converts UTC to New York time, adjusting for daylight saving.

---

Question 8 Solution: Generator with Exceptions

#include <coroutine>
#include <stdexcept>
#include <iostream>

template<typename T>
struct Generator {
    struct promise_type {
        T value;
        std::exception_ptr eptr;
        auto yield_value(T val) { value = val; return suspend_always{}; }
        auto initial_suspend() { return suspend_never{}; }
        auto final_suspend() noexcept { return suspend_always{}; }
        void unhandled_exception() { eptr = std::current_exception(); }
        Generator get_return_object() { return Generator{handle_type::from_promise(*this)}; }
    };
    using handle_type = std::coroutine_handle<promise_type>;
    handle_type coro;
    explicit Generator(handle_type h) : coro(h) {}
    ~Generator() { if (coro) coro.destroy(); }
    T next() {
        coro.resume();
        if (coro.promise().eptr) std::rethrow_exception(coro.promise().eptr);
        return coro.promise().value;
    }
};

Generator<int> gen(std::vector<int> v) {
    for (int x : v) {
        if (x < 0) throw std::runtime_error("Negative value");
        co_yield x;
    }
}

int main() {
    try {
        auto g = gen({5, 3, -1, 2});
        std::cout << g.next() << " "; // 5
        std::cout << g.next() << " "; // 3
        g.next(); // Throws
    } catch (const std::exception& e) {
        std::cout << e.what(); // "Negative value"
    }
}

Explanation:

Exceptions in coroutines are captured via unhandled_exception and rethrown using std::rethrow_exception.

---

Question 9 Solution: Semaphore Queue

#include <semaphore>
#include <queue>
#include <thread>

template<typename T>
class SafeQueue {
    std::queue<T> queue;
    std::mutex mtx;
    std::counting_semaphore<> items{0};
    std::counting_semaphore<> spaces{3}; // Buffer size 3
public:
    void enqueue(T item) {
        spaces.acquire();
        {
            std::lock_guard lock(mtx);
            queue.push(item);
        }
        items.release();
    }
    T dequeue() {
        items.acquire();
        T item;
        {
            std::lock_guard lock(mtx);
            item = queue.front();
            queue.pop();
        }
        spaces.release();
        return item;
    }
};

int main() {
    SafeQueue<int> q;
    std::jthread producer([&] { for (int i=0; i<5; ++i) q.enqueue(i); });
    std::jthread consumer([&] { for (int i=0; i<5; ++i) q.dequeue(); });
}

Explanation:
counting_semaphore controls buffer slots. Producers wait on spaces, consumers on items.

---

Question 10 Solution: Latch Synchronization

#include <latch>
#include <thread>
#include <vector>

int main() {
    std::vector<int> data(4);
    std::latch latch(4);
    auto worker = [&](int idx) {
        data[idx] = idx + 1;
        latch.count_down();
    };
    std::jthread t1(worker, 0);
    std::jthread t2(worker, 1);
    std::jthread t3(worker, 2);
    std::jthread t4(worker, 3);
    latch.wait(); // Wait for all threads
    // data becomes [1,2,3,4]
}

Explanation:
std::latch coordinates thread completion. count_down decrements the counter; wait blocks until zero.

---

Question 11 Solution: Barrier Phases

#include <barrier>
#include <iostream>
#include <thread>

int main() {
    std::barrier bar(3, []() noexcept { std::cout << "Phase completed\n"; });
    auto task = [&](int id) {
        for (int phase=0; phase<3; ++phase) {
            std::cout << "Thread " << id << " phase " << phase << "\n";
            bar.arrive_and_wait();
        }
    };
    std::jthread t1(task, 1);
    std::jthread t2(task, 2);
    std::jthread t3(task, 3);
}

Explanation:
std::barrier synchronizes threads at each phase. The completion function prints after each phase.

---

Question 12 Solution: jthread Cancellation

#include <stop_token>
#include <thread>
#include <iostream>

void task(std::stop_token stoken) {
    int count = 0;
    while (!stoken.stop_requested() && count < 10) {
        std::cout << "Iteration " << ++count << "\n";
        std::this_thread::sleep_for(std::chrono::milliseconds(100));
    }
}

int main() {
    std::jthread t(task);
    std::this_thread::sleep_for(std::chrono::milliseconds(550));
    t.request_stop(); // Stops after ~5 iterations
}

Explanation:
std::stop_token checks for cancellation requests. request_stop() signals the thread to exit.

---

Question 13 Solution: osyncstream

#include <syncstream>
#include <thread>

void log(int id) {
    for (int i=0; i<100; ++i)
        std::osyncstream(std::cout) << "Thread " << id << ": " << i << "\n";
}

int main() {
    std::jthread t1(log, 1);
    std::jthread t2(log, 2);
}

Explanation:
std::osyncstream ensures atomic writes to std::cout, preventing interleaved lines.

---

Question 14 Solution: is_constant_evaluated

#include <type_traits>

constexpr int square(int x) {
    if (std::is_constant_evaluated()) return x * x;
    else return x * x; // Runtime version (could differ)
}

int main() {
    static_assert(square(5) == 25); // Compile-time
    int x = square(5); // Runtime
}

Explanation:
std::is_constant_evaluated allows branching between compile-time and runtime logic.

---

Question 15 Solution: source_location Logging

#include <source_location>
#include <iostream>

#define LOG(msg) \
    std::cout << "[" << std::source_location::current().file_name() \
              << ":" << std::source_location::current().line() << "] " \
              << std::source_location::current().function_name() << ": " \
              << msg << "\n"

void demo() {
    LOG("Test message");
}

int main() {
    demo(); // Output: [main.cpp:10] demo: Test message
}

Explanation:
std::source_location captures context (file, line, function) at the point of the macro expansion.

---

Summary

These challenges cover C++20's critical features, emphasizing the Standard Library and Concurrency. Each solution

demonstrates idiomatic usage and highlights edge cases (e.g., exception handling, thread safety). Testing

ensures correctness across compile-time/runtime scenarios.