C++ STL面试题, STL

可以通过重载比较运算符或者提供自定义的比较函数。例如，在std::sort中可以通过第三个参数传递一个比较函数来自定义排序规则：std::sort(vec.begin(), vec.end(), [](int a, int b){ return a > b; });

STL中的智能指针（如std::unique_ptr, std::shared_ptr）用于自动管理动态分配的内存，防止内存泄漏。智能指针在生命周期结束时会自动释放所管理的内存。

仿函数（Functors）是行为类似于函数的对象，可以通过重载operator()来实现。仿函数通常用于STL算法中，提供自定义的操作逻辑。例如，可以定义一个仿函数来作为std::for_each的参数。

emplace_back在容器尾部直接构造元素，避免了不必要的拷贝或移动操作，而push_back则需要先构造一个临时对象，然后将其拷贝或移动到容器中。

C++11新增了许多特性，包括自动类型推导（auto）、范围for循环、lambda表达式、右值引用和移动语义、智能指针、线程支持库等。这些特性极大地提升了C++语言的易用性和效率。

虚函数表（vtable）是C++实现多态性的一种机制。当一个类中有虚函数时，编译器为该类生成一个虚函数表，表中存储了虚函数的地址。每个对象有一个指向虚函数表的指针，调用虚函数时通过这个指针找到正确的函数实现。

避免内存泄漏的关键是确保每个动态分配的内存都有相应的释放操作。可以通过RAII、智能指针或使用适当的容器来管理动态内存。此外，定期使用工具（如Valgrind）进行内存泄漏检测也是一种有效的手段。

std::vector是一个动态数组，支持随机访问，插入和删除操作复杂度在O(n)左右，而std::list是一个双向链表，插入和删除操作的复杂度为O(1)，但不支持随机访问。这两者的选择取决于具体的应用场景：当需要频繁的插入和删除时，list更合适，而需要频繁访问元素时，vector更有效。

内联函数是通过在函数定义前加上inline关键字来声明的，编译器会尝试将函数的调用替换为函数体，从而减少函数调用的开销。内联函数适用于小型、频繁调用的函数。但要注意，内联只是一个请求，编译器可以选择忽略它。

C++中的多态性通过虚函数表（vtable）和虚指针（vptr）实现。每个有虚函数的类在编译时会生成一个虚函数表，其中存储了虚函数的地址，虚指针指向这个表。通过基类指针或引用调用虚函数时，程序通过虚指针查找并调用对应派生类的函数。

map是一个有序容器，底层实现是红黑树，因此其元素总是按键值排序。unordered_map是一个无序容器，底层实现是哈希表，查找、插入和删除操作的平均时间复杂度为O(1)。选择使用哪一个取决于是否需要键值有序以及性能的需求。

auto关键字用于自动推导变量的类型，编译器在编译时会根据变量初始化的值推导出其类型。这使得代码更加简洁，但也需要开发者清楚推导的类型，以避免不必要的错误。

避免内存泄漏的常见方法包括使用智能指针（如std::shared_ptr和std::unique_ptr）、RAII模式，以及确保动态分配的内存在不再使用时被正确释放。智能指针在现代C++中是推荐的做法，因为它们自动管理内存的释放。

构造函数用于对象的初始化，而析构函数用于对象生命周期结束时的清理工作。构造函数在对象创建时自动调用，析构函数在对象销毁时自动调用。

std::shared_ptr is a smart pointer that manages shared ownership of a dynamically allocated object, meaning multiple std::shared_ptr instances can point to the same object, and the object is only deleted when the last std::shared_ptr is destroyed. std::unique_ptr, on the other hand, manages exclusive ownership, meaning only one std::unique_ptr can point to a given object, and the object is deleted when the std::unique_ptr is destroyed or reset. std::unique_ptr is preferred for exclusive ownership due to its simplicity and efficiency.

STL containers are exception-safe, meaning they will not leak resources or leave objects in an inconsistent state if an exception occurs during an operation. For example, if an exception is thrown during an insertion operation, the container will roll back to its previous state. Additionally, STL algorithms generally follow the strong exception safety guarantee, ensuring that if an operation fails, the program state remains unchanged.

Copy elision is an optimization technique where the compiler is allowed to omit the copying or moving of objects, typically when returning a local object from a function. This results in better performance by eliminating unnecessary temporary objects. In C++17, copy elision is mandatory in certain situations, such as returning a local object from a function, making it an important feature for performance optimization.

STL algorithms are divided into several categories: non-modifying sequence operations (e.g., std::find, std::accumulate), modifying sequence operations (e.g., std::copy, std::transform), sorting operations (e.g., std::sort, std::partial_sort), and set operations (e.g., std::set_union, std::set_intersection). Understanding these categories helps in effectively using the algorithms to manipulate containers and data structures in C++.

std::vector is a dynamic array that provides fast random access and is efficient for operations at the end of the sequence (push_back). std::list is a doubly-linked list that allows fast insertion and deletion from anywhere in the sequence but has slower access times due to lack of random access. std::vector is generally preferred unless frequent insertions/deletions in the middle of the sequence are required.

C++11 introduced the <thread> library, which provides a standard way to create and manage threads. It includes features like std::thread for creating threads, std::mutex and std::lock_guard for synchronization, and std::async for asynchronous task execution. These tools enable developers to write portable and efficient multithreaded applications.

'nullptr' is a keyword introduced in C++11 that represents a null pointer. Unlike NULL, which is typically defined as 0, nullptr is of type std::nullptr_t and provides type safety in pointer operations. It ensures that null pointers cannot be mistaken for integers, reducing the chance of subtle bugs in pointer arithmetic or overload resolution.

Variadic templates, introduced in C++11, allow functions and classes to accept an arbitrary number of template arguments. They are defined using the syntax template<typename... Args>. Variadic templates enable more flexible and generic code compared to traditional templates, which require each argument to be explicitly specified.

'constexpr' indicates that a function or variable can be evaluated at compile-time if given constant expressions as arguments. This allows for more powerful optimizations and guarantees that certain computations are done at compile time, improving runtime performance. 'constexpr' is often used in scenarios where you want to ensure that a value is constant and can be used in contexts requiring compile-time constants, such as array sizes or template arguments.