How Does C 11 Enhance In-Class Initialization?

C 11 In-Class Initialization: Addressing Past Restrictions

Before C 11, in-class initialization was limited to static const members of integral or enumeration types. This restriction stemmed from the need to avoid complicated linker rules, ensuring each object has a unique definition.

However, C 11 introduced significant changes by allowing in-class initialization of non-static and non-const members (§12.6.2/8). This relaxation was possible due to improvements in the compiler, which now assumes the responsibility of sorting out any potential multiple definitions, allowing for more complex class declarations.

For non-static data members, brace-or-equal-initializers can be used to initialize during object construction. Variant members are not initialized unless explicitly specified, while default-initialization is performed when no initializer is provided.

In-class initialization of non-const static members is also now possible when they are declared with the constexpr specifier (Section 9.4.2).

These changes have improved code simplicity and efficiency by eliminating the need for out-of-class member initializers. However, they also introduce new rules for handling potential duplicate initializers within different class constructors. For example, if a member variable has an in-class initializer and is also initialized within a constructor, the in-class initializer will be ignored when using that constructor.

In summary, while the underlying implementation complexities have shifted from the linker to the compiler, C 11's in-class initialization greatly enhances class definition and member initialization capabilities, improving code readability and flexibility.

The above is the detailed content of How Does C 11 Enhance In-Class Initialization?. For more information, please follow other related articles on the PHP Chinese website!