Wrapper libraries (or library wrappers) consist of a thin layer of code (a "shim") which translates a library's existing interface into a compatible interface. This is done for several reasons:
To refine a poorly designed or complicated interface
Allow code to work together which otherwise cannot (e.g. incompatible data formats)
Enable cross language and/or runtime interoperability
The specific way in which a wrapper library is implemented is highly specific to the environment it is being written in and the scenarios which it intends to address. This is especially true in the case when cross-language/runtime interoperability is a consideration.
Example
The following provides a general illustration of a common wrapper library implementation. In this example, a C++ interface acts as a "wrapper" around a C interface.
The original C interface can be regarded as error prone, particularly in the case where users of the library forget to unlock an already locked mutex. The new interface effectively utilizes resource acquisition is initialization (RAII) in the new Mutex and Lock classes to ensure Mutexs are eventually unlocked and pthread_mutex_t objects are automatically released.
The above code closely mimics the implementation of boost::scoped_lock and boost::mutex which are part of the boost::thread library.
Some wrapper libraries exist to act as a bridge between a client application and a library written using an incompatible technology. For instance, a Java application may need to execute a system call. However system calls are typically exposed as C library functions. To resolve this issue Java implements wrapper libraries which make these system calls callable from a Java application.
In order to achieve this, languages like Java provide a mechanism called foreign function interface that makes this possible. Some examples of these mechanisms include: