Java packages vs. C++ libraries

Question

In Java , there is what is called package . Does library in C++ represent the same meaning, especially in terms for example of containg relative classes and the use of protected members?

Thanks.

Answer 1

There are different dimensions of what a package means in Java. As a container that differentiates the names of the classes inside from the names of classes in other packages, its equivalent would be c++ namespaces .

As a unit that guarantees access to non-private members to classes in the same block, there is no equivalent in C++. The access level granted to a class is independent of the namespace where the class is defined.

As a way of ordering your sources in the disk, there is no equivalent, the C++ language has no requirements on how the code is stored in files.

Regarding c++ libraries, that is closer to jar files in Java. They bundle different classes that share some relation. A jar can contain more than one package, and more than one jar can contain classes from the same package. Similarly with libraries, they can contain classes from different namespaces and/or different libraries can contain classes from the same namespace.

Answer 2

The closest to Java packages are namespaces in C++.

They can be nested into one another, and you need to specifically declare that you are using them or a part of their contents. However, they do not enforce any physical file hierarchy like Java packages do.

Answer 3

严格来说，我认为C ++中的命名空间提供了相同的语义。

Answer 4

I guess it is more related to namespaces in C++. Java and C++ both use libraries. Library can be any independent set of classes[probably a framework] which can be accessed in our code.

Answer 5

External Libraries are there in both Java and C++. Just the formats vary, .jar in Java and .dll/.so in C++.

Purpose of Packages and Namespaces are different from Libraries. They avoid running out of names by allowing user to logically group the source.

Answer 6

A package in Java is a namespace for classes, interfaces and enums. Package name, a dot and the classname form the fully qualified classname of a class:

com.example.hello.HelloWorldApplication
^--packagename--^ ^-----classname-----^

The dots in the package name have a different meaning then the dot between the names: the first two dots of this example are part of the package name , the last one is a separator.

This should be kept in mind, because there's a common misunderstanding regarding package names: just because the names can be mapped to a hierarchical folder structure, some people think, package names have a hierarchy too - which is not the case: hello is not a "subpackage" of example !

But, to create a simple mapping to folders and files, a classloader can simply take the fully qualified class name, replace all dots with a slash and append .class to get a relative path to a class file.

But note again, that a folder/file mapping is not required t load classes - we can invent a class loader that gets classes from a database or a remote service - a folder/file mapping wouldn't make any sense in that case.