Getting Started with qmake

Let's assume that you have just finished a basic implementation of your application, and you have created the following files:

• hello.cpp

• hello.h

• main.cpp

You will find these files in the examples/qmake/tutorial directory of the Qt distribution. The only other thing you know about the setup of the application is that it's written in Qt. First, using your favorite plain text editor, create a file called hello.pro in examples/qmake/tutorial. The first thing you need to do is add the lines that tell qmake about the source and header files that are part of your development project.

We'll add the source files to the project file first. To do this you need to use the SOURCES variable. Just start a new line with SOURCES += and put hello.cpp after it. You should have something like this:

SOURCES += hello.cpp

We repeat this for each source file in the project, until we end up with the following:

SOURCES += hello.cpp
SOURCES += main.cpp

If you prefer to use a Make-like syntax, with all the files listed in one go you can use the newline escaping like this:

SOURCES = hello.cpp \
          main.cpp

Now that the source files are listed in the project file, the header files must be added. These are added in exactly the same way as source files, except that the variable name we use is HEADERS.

Once you have done this, your project file should look something like this:

HEADERS += hello.h
SOURCES += hello.cpp
SOURCES += main.cpp

The target name is set automatically. It is the same as the project filename, but with the suffix appropriate for the platform. For example, if the project file is called hello.pro, the target will be hello.exe on Windows and hello on Unix. If you want to use a different name you can set it in the project file:

TARGET = helloworld

The finished project file should look like this:

HEADERS += hello.h
SOURCES += hello.cpp
SOURCES += main.cpp

You can now use qmake to generate a Makefile for your application. On the command line, in your project directory, type the following:

qmake -o Makefile hello.pro

Then type make or nmake depending on the compiler you use.

For Visual Studio users, qmake can also generate Visual Studio project files. For example:

qmake -tp vc hello.pro

The release version of an application does not contain any debugging symbols or other debugging information. During development, it is useful to produce a debugging version of the application that has the relevant information. This is easily achieved by adding debug to the CONFIG variable in the project file.

For example:

CONFIG += debug
HEADERS += hello.h
SOURCES += hello.cpp
SOURCES += main.cpp

Use qmake as before to generate a Makefile. You will now obtain useful information about your application when running it in a debugging environment.

After a few hours of coding, you might have made a start on the platform-specific part of your application, and decided to keep the platform-dependent code separate. So you now have two new files to include into your project file: hellowin.cpp and hellounix.cpp. We cannot just add these to the SOURCES variable since that would place both files in the Makefile. So, what we need to do here is to use a scope which will be processed depending on which platform we are building for.

A simple scope that adds the platform-dependent file for Windows looks like this:

win32 {
    SOURCES += hellowin.cpp
}

When building for Windows, qmake adds hellowin.cpp to the list of source files. When building for any other platform, qmake simply ignores it. Now all that is left to be done is to create a scope for the Unix-specific file.

When you have done that, your project file should look something like this:

CONFIG += debug
HEADERS += hello.h
SOURCES += hello.cpp
SOURCES += main.cpp
win32 {
    SOURCES += hellowin.cpp
}
unix {
    SOURCES += hellounix.cpp
}

Use qmake as before to generate a Makefile.

You may not want to create a Makefile if a certain file does not exist. We can check if a file exists by using the exists() function. We can stop qmake from processing by using the error() function. This works in the same way as scopes do. Simply replace the scope condition with the function. A check for a file called main.cpp looks like this:

!exists( main.cpp ) {
    error( "No main.cpp file found" )
}

The ! symbol is used to negate the test. That is, exists( main.cpp ) is true if the file exists, and !exists( main.cpp ) is true if the file does not exist.

CONFIG += debug
HEADERS += hello.h
SOURCES += hello.cpp
SOURCES += main.cpp
win32 {
    SOURCES += hellowin.cpp
}
unix {
    SOURCES += hellounix.cpp
}
!exists( main.cpp ) {
    error( "No main.cpp file found" )
}

Use qmake as before to generate a makefile. If you rename main.cpp temporarily, you will see the message and qmake will stop processing.

Suppose you use Windows and you want to be able to see statement output with qDebug() when you run your application on the command line. To see the output, you must build your application with the appropriate console setting. We can easily put console on the CONFIG line to include this setting in the Makefile on Windows. However, let's say that we only want to add the CONFIG line when we are running on Windows and when debug is already on the CONFIG line. This requires using two nested scopes. First create one scope, then create the other inside it. Put the settings to be processed inside the second scope, like this:

win32 {
    debug {
        CONFIG += console
    }
}

Nested scopes can be joined together using colons, so the final project file looks like this:

CONFIG += debug
HEADERS += hello.h
SOURCES += hello.cpp
SOURCES += main.cpp
win32 {
    SOURCES += hellowin.cpp
}
unix {
    SOURCES += hellounix.cpp
}
!exists( main.cpp ) {
    error( "No main.cpp file found" )
}
win32:debug {
    CONFIG += console
}

That's it! You have now completed the tutorial for qmake, and are ready to write project files for your development projects.