Translator's Note: Many programmers may be very familiar with C++, but they may not be very clear about the places where namespaces are frequently used. This article will describe the use of namespaces.
Namespaces were introduced into the C++ standard in 1995 and are typically defined as follows:
Namespaces define new scopes. They provide a way to avoid name conflicts.
Namespaces in C++ are typically used to avoid naming conflicts. Although namespaces are widely used in recent C++ code, most older code does not use this tool.
Based on the exploration and research of the source code of numerous C++ projects, some common reasons for using namespaces in these projects have been summarized.
Avoid name conflicts
As mentioned earlier, this is the most common reason for using namespaces in C++, where their use is only effective for the compiler and offers no value to developers in terms of code readability and maintainability.
Modular applications
Namespaces employ a "namespace-by-feature" approach to modularize code. "Namespace-by-feature" reflects a set of functionalities through namespaces. It groups related items into the same namespace. This creates highly cohesive, loosely coupled, and highly modular namespaces.
Boost is a prime example of feature-based grouping, containing thousands of namespaces, each used to group specific features.
Anonymous namespaces
Anonymous namespaces prevent the creation of global static variables. An anonymous namespace you create is only accessible within the file that created it.
Solving the problem of duplicate member names in enumeration types
If enumerations with the same name are defined within the same scope, "traditional" enumerations in C++ will export enumeration values within their scope, potentially leading to name conflicts.
In a large project, there is no guarantee that two different enumerations will not have the same name. This problem has been solved in C++11 by using enumeration classes, which implicitly define the enumeration values in the enumeration name.
Many years ago, this problem was solved by declaring the enumeration within a namespace, rather than declaring the enumeration like this.
enumstatus{
status_ok,
status_error
};
It is declared in the namespace:
namespacestatus{
enumstatus{
OK,
error
};
}
Many C++ projects use this technique; for example, the Unreal Engine source code extensively uses this technology.
Hidden implementation
For template libraries implemented in header files, developers don't need to use special data types when calling them because they only need to focus on functionality. Therefore, finding a suitable way for developers to call these libraries is quite interesting. In C#, the "internal" keyword does this, but in C++, there's no way to completely hide public data types from developers.
Separating definitions from implementations within modules is a C++ idiom created by the Boost library developers, but these definitions must be placed in a sub-namespace that can be called by developers, according to certain rules.
For example, in the boost::math documentation, it specifies:
Functions that are not applicable to the application are in boost::math::detail.
Disclaimer: This article is a reprint. If it involves copyright issues, please contact us promptly for deletion (QQ: 2737591964 ) . We apologize for any inconvenience.
