Debug-Aiding Code in Visual C++.NET
Introductory Concepts
You can use four macros liberally in your code to verify assumptions and to alert you when problems arise in your application. These macros are executed only with Debug builds of your applications. When you make Release builds, these macros are removed.
New Term: Macro
A macro in C++ (also in the predecessor C programming language) is a pseudo-function created using the #define pre-compiler directive. Prior to compiling, the macro is replaced by all code used in the macro definition to define the macro functionality. One benefit of using macros is it allows the designers of MFC to define one implementation of a particular macro for Debug builds, and another for Release builds.
C++ Sidebar: Pre-Compiler Directives
In the C and C++ programming languages, all source code files are passed through a pre-compiler before compiling the code. The pre-compiler processes some specific directives, expanding the code to be compiled significantly. All pre-compiler directives are recognizable because they start with the pound symbol (#) as the very first character in the line, prior to any character other than white-space. The primary pre-compiler directives are listed in Table 2.1.
Table 2.1—C/C++ Pre-Compiler Directives
Directive |
Description |
#include file_name |
This directive instructs the pre-compiler to insert the contents of the specified file at this location. If the file name is enclosed with the less-than/greater-than brackets (<>), the configured directories are searched for the file specified. If quotes ("") are used, the path specified with the file name is used to find the file. The path can be either a relative path (from the project directory), or an absolute path (including the drive letter). |
#define CONSTANT value |
This directive is used to define a constant value. The constant is normally declared in all uppercase. The pre-compiler replaces every occurrence of the constant with the value prior to compiling. |
#define CONSTANT(par1) definition |
This directive is used to define a macro. The macro is normally declared in all upper-case. The pre-compiler replaces every occurrence of the macro with the definition of the macro prior to compiling. When the pre-compiler processes the macro, it goes through the definition and replaces all occurrences of the parameters with the actual variables or values that are supplied in each use of the macro. |
#ifdef CONSTANT |
This directive tells the pre-compiler to include the following section of code only if the specified constant has already been defined. |
#ifndef CONSTANT |
This directive tells the pre-compiler to include the following section of code only if the specified constant has not been defined. |
#else |
This directive is used with the prior two directives, marking the next section to be included depending on whether the condition was met for including the prior section. |
#endif |
This directive is used to mark the end of the sections to be conditionally included or excluded from the code to be compiled. |
#pragma option |
This directive is used to specify some compiler or machine specific option. There are many options available for use with the #pragma directive with most C/C++ compilers. To determine which #pragma options are available with a specific compiler, you need to reference the compiler's documentation (we'll look at some of the #pragma options available in Visual C++ from time to time as we make our way through this book). |
TIP
Start working with pre-compiler directives slowly. You should start by using only the #include directive to include needed header files, and the #define directive to define constants. Leave the other directives until you have a bit of experience under your belt; using them can result in confusing and unexpected results if you don't thoroughly think through how and when to use them.
Verifying Assumptions
There are four macros that you can use in your code to test various assumptions that you make while designing and building your applications. These are a series of variations on the first one of these macros, called ASSERT. The syntax for this macro is as follows:
ASSERT(bAssumption);
The parameter that is passed into this macro is any Boolean expression that should always evaluate to TRUE. For example, if you have a function with a parameter a that should always be positive in value, you can test it with an ASSERT as follows:
ASSERT(a >= 0);
If ever this function is called with a negative value passed in the a parameter, ASSERT will pop-up a warning like that shown in Figure 2.2, pointing you to the specific ASSERT that failed.
Figure 2.2 An ASSERT alerting you to a problem in your application.
CAUTION
Never place necessary functionality in the parameter being passed to the ASSERT macro.
One thing to keep in mind when using ASSERT is that when you create a Release build of your application, all uses of ASSERT are removed from the application. If you have used any program logic in the parameter being passed to this macro, it too will be removed from the application.
There are two variations on ASSERT that can be used to test classes and objects in your code. The first of these, ASSERT_VALID, is used as such:
ASSERT_VALID(pObject);
This version is used to test objects being used in your application to verify that they are valid objects, and that there are no problems with the internal state of the object. The one parameter passed into this macro is an instantiated C++ object.
NOTE
The ASSERT_VALID macro does not work on all C++ objects. The object must be inherited from the CObject base class, and must have overridden the AssertValid member function. For most standard MFC classes, these criteria are met. For testing your own classes, you'll need to keep these requirements in mind when designing the classes.
The second variation on ASSERT is the ASSERT_KINDOF macro, which is used as follows:
ASSERT_KINDOF(classname, pObject);
This macro validates that an object is a specific class, or is derived from that specific class. The first parameter is the class name that the second parameter needs to be an instance of.
NOTE
This function can only be used to verify the class of classes that meet other specific criteria, such as it has to be a descendent of the CObject class, and must have one of two other macros used in the class declaration. This is getting well ahead of ourselves if you don't already know the C++ programming language. You might have to look back at this note once you have a more thorough understanding of C++ and the MFC class library. The two macros that the class must have used one of in its declaration are DECLARE_DYNAMIC or DECLARE_SERIAL.
Like the ASSERT macro, both variations are completely removed from Release builds of applications, so don't use these in place of necessary logic. You want to use these macros only for verifying assumptions in your code.
There is one last variation on the ASSERT macro that is uniquely different, yet the same. This is the VERIFY macro. It is used just like the ASSERT macro, as follows:
VERIFY(bAssumption);
One key difference between the VERIFY and ASSERT macro is that any logic passed as the parameter to the VERIFY macro remains in the application when a Release build is made. This means that you can put actual program logic in the parameter to the VERIFY macro, and that logic will not be removed when you make a Release build. However, just like the ASSERT macro, the VERIFY macro will only test the expression and alert you if it's FALSE in Debug builds of your applications.
Following Flow and Execution
Sometimes you want to know the execution path that your application is taking, or certain things that are going on as it's executing, but you don't want to step through the application line-by-line. What you would really like to do is have a log of when the application reaches or passes certain points of code, and what the state of certain variables are. This is easy to do, adding your own code for writing logging messages to a file, but it's a pain to strip that code when you need to pass the application to someone else.
There is an easy way to accomplish this logging, without the pain of removing all of your logging messages. The way you should accomplish this is by using the TRACE macro. The TRACE macro sends any string message you pass to it to any debug output window you run on the computer the same time your application is running (see Figure 2.3). The syntax for this macro is as follows:
TRACE("This is my trace message.\n");
Figure 2.3 Trace messages from a debugging session.
TIP
The TRACE macro works just like the printf function in C, so you can pass it variables to be dynamically added to the message string just as you would with the printf function.
CAUTION
The output string from the TRACE macro is limited to 512 characters. If the formatted string you pass to the TRACE macro is longer than 512 characters (including the string terminating NULL character), it will trigger an ASSERT.
Just as with the ASSERT macro, all uses of the TRACE macro are automatically removed from Release builds.
C++ Sidebar: Formatting Strings
The C programming language was the predecessor to the C++ programming language. C++ was created for the purpose of making an object-oriented version of C. You can still write standard C applications using any C++ compiler since part of the C++ standard is full support for the C programming language.
In the C programming language, there is a family of string formatting/printing functions. The primary function, printf, is used to format and print strings to the primary output device (this is for character-mode applications that would be running in a DOS window on a Windows system). There are two other variations to this function. The first, fprintf, allows you to send your string message to a file or output device. In fact, the printf version calls the fprintf function, specifying the standard output device as where to send the string. Finally, there is also the sprintf function, which formats the string and places it into a character array. You occasionally still see these functions used in C++ code from time to time.
The printf function takes a string as its primary parameter, which can be followed by any number of variables to be included in the string. Each variable must have a placeholder in the string. These placeholders are specified by the percent symbol (%) followed by a character indicating the data type of the value to place in that location. Table 2.2 lists the characters and their corresponding data types.
Table 2.2—printf Data Type Specifiers
Placeholders |
Data Type |
%c |
A single-byte character. |
%C |
A wide or UNICODE character. |
%d or %i |
Decimal integer data type. This includes integers and longs. |
%u |
Unsigned integer data type. |
%f |
Floating-point data types, including doubles and floats. |
%s |
Null-terminated string. |
%S |
Wide or UNICODE string. |
%x |
Hexadecimal value (lower-case). |
%X |
Hexadecimal value (upper-case). |
%o |
Octal value. |
%% |
Percent sign. Because the percent sign is used to indicate that the next character in the string is a placeholder for a variable value, you have to include two percent signs in order to include a percent sign in your string. |
If you need to format a number variable to be a specific number of digits, you can insert the number of digits between the percent and the placeholder character, as follows:
printf("This number: %5d is five digits wide.", iNbr);
For floating-point numbers, you can also specify the precision by adding a decimal point and the precision following the number specifying the width, as follows:
printf("I have %3.2f dollars in my wallet.", fMoney);
If the fMoney variable's value was 123.456, the output from the above function call would be:
I have 123.46 dollars in my wallet.
Normally, when you format numbers, they are printed aligned on the right side. This means that if the fMoney value was 1.234, the output would look like the following:
I have 1.23 dollars in my wallet.
If you need the number to be aligned on the left side, add a minus sign ([ms]) to the formatting expression:
printf("I have %-3.2f dollars in my wallet.", fMoney);
The resulting output would look like the following:
I have 1.23 dollars in my wallet.
If you want the output to include the minus sign if the value is negative, place the plus sign (+) just after the percent sign:
printf("I have %+3.2f dollars in my wallet.", fMoney);
This will cause a minus sign to be printed if the value of the fMoney variable is negative.
You also can include non-printable characters in the strings by including a back-slash followed by the character that represents the non-printable character. This can be used to include tabs, new-lines, carriage-returns, and so on. Table 2.3 lists the primary non-printable characters that you're likely to include.
Table 2.3—printf Non-Printable Character Specifiers
Placeholders |
Description |
\n |
New-line character. Use this to break a string into multiple lines of text. |
\t |
Tab character. |
\r |
Carriage-return character, causes the text to scroll back to the first of the string (but does not scroll to the next line). |
\\ |
Back-slash character. Because the back-slash character is used to indicate that the next character in the string is a placeholder for a non-printable character, you have to include two back-slash characters in order to include a single back-slash character in your string. |