The use of dynamically allocated memory must be managed, because it has a tremendous effect on the performance of your programs. The current trend in handling dynamic memory seems to be shifting toward an automated approach. While C++ uses the manual approach for managing dynamic memory, this does not mean that it can't be automated in that language -- thus giving the C++ programmer the best of both worlds. This article explains how to do it. It is excerpted from chapter two of The Art of C++, written by Herbert Schildt (McGraw-Hill/Osborne, 2004; ISBN: 0072255129).
A Simple Garbage Collector for C++ (Page 1 of 12 )
Throughout the history of computing, there has been an ongoing debate concerning the best way to manage the use of dynamically allocated memory. Dynamically allocated memory is memory that is obtained during runtime from the heap, which is a region of free memory that is available for program use. The heap is also commonly referred to as free store or dynamic memory. Dynamic allocation is important because it enables a program to obtain, use, release, and then reuse memory during execution. Because nearly all real-world programs use dynamic allocation in some form, the way it is managed has a profound effect on the architecture and performance of programs.
In general, there are two ways that dynamic memory is handled. The first is the manual approach, in which the programmer must explicitly release unused memory in order to make it available for reuse. The second relies on an automated approach, commonly referred to as garbage collection, in which memory is automatically recycled when it is no longer needed. There are advantages and disadvantages to both approaches, and the favored strategy has shifted between the two over time.
C++ uses the manual approach to managing dynamic memory. Garbage collection is the mechanism employed by Java and C#. Given that Java and C# are newer languages, the current trend in computer language design seems to be toward garbage collection. This does not mean, however, that the C++ programmer is left on the “wrong side of history.” Because of the power built into C++, it is possible—even easy—to create a garbage collector for C++. Thus, the C++ programmer can have the best of both worlds: manual control of dynamic allocation when needed and automatic garbage collection when desired.
This chapter develops a complete garbage collection subsystem for C++. At the outset, it is important to understand that the garbage collector does not replace C++’s built-in approach to dynamic allocation. Rather, it supplements it. Thus, both the manual and garbage collection systems can be used within the same program.
Aside from being a useful (and fascinating) piece of code in itself, a garbage collector was chosen for the first example in this book because it clearly shows the unsurpassed power of C++. Through the use of template classes, operator overloading, and C++’s inherent ability to handle the low-level elements upon which the computer operates, such as memory addresses, it is possible to transparently add a core feature to C++. For most other languages, changing the way that dynamic allocation is handled would require a change to the compiler itself. However, because of the unparalleled power that C++ gives the programmer, this task can be accomplished at the source code level.
The garbage collector also shows how a new type can be defined and fully integrated into the C++ programming environment. Such typeextensibility is a key component of C++, and it’s one that is often overlooked. Finally, the garbage collector testifies to C++’s ability to “get close to the machine” because it manipulates and manages pointers. Unlike some other languages which prevent access to the low-level details, C++ lets the programmer get as close to the hardware as necessary.