[ fromfile: clargs.xml id: clargs ]
main()
is a function that is called at program startup.
If you want the program to accept command-line arguments, you must define main
with
its full parameter list.
C and C++ permit some flexibility in the way that arguments are defined in main()
, so
you may see it defined in a variety of ways:
int main(int argc, char* argv[]) int main(int argCount, char ** argValues) int main(int argc, char * const argv[])
All the forms above are valid, and they each define two parameters: an int
to contain
the number of command-line arguments and an array of C-style strings to contain the actual arguments.
These parameters contain enough information to reconstruct the command-line arguments passed into the program from
the parent process.[15]
Example
1.19 is a simple main
program that prints its command-line arguments.
Example 1.19. src/clargs/clargs-iostream/clargs.cpp
argv
, the argument vector, is
an array (Section 21.4) that contains
all the command-line strings.
In other words, because each individual command-line string is itself an array of char
, argv
is an array of char
arrays.
argc
, the argument count, is
the number of char
arrays in argv
.
main()
needs to return an int
to the parent process.
In *nix systems, the return value of a process is called its exit status.
Presumably, the parent process can use the exit status to decide what to do next.[16]
The return value should be 0 if all went well, or a nonzero error code if something went wrong.
This can be accomplished in a few different ways.
With an explicit
return 0;
statement, as we do in most of our examples.
If there is no return statement in the main()
block, zero will be returned by
default when the closing brace } is reached. Example 1.20 demonstrates this.
For maximum portability, programs strictly adhering to the C++ standard should
#include <cstdlib>
and then explicitly terminate main()
with
return EXIT_SUCCESS;
as in Section 1.11.
Note | |
---|---|
Try not to confuse this interpretation of |
If you run this program with command-line arguments, you will see something like this in the output:
clargs> ./clargs spam eggs "space wars" 123 argc = 5 argv# 0 is ./clargs argv# 1 is spam argv# 2 is eggs argv# 3 is space wars argv# 4 is 123 246
The first argument is the name of the executable. The other arguments are taken from the command-line as strings separated by spaces or tabs. To pass a string that contains spaces as a single argument, you must enclose the string in quotes.
The last argument looks like the number 123. In fact, it is the string representation of that number. If you need to calculate with that number, it would be necessary to use a suitable function to convert the string "123" to the number 123.
In example Example 1.20, we have rewritten Example 1.19 to set up and access the command-line arguments using Qt types and avoiding the use of arrays. The two applications produce the same output.
Example 1.20. src/clargs/qt/clargs.cpp
#include <QTextStream> #include <QCoreApplication> #include <QStringList> int main (int argc, char* argv[]) { QCoreApplication app(argc, argv); QTextStream cout(stdout); QStringList arglst = app.arguments(); cout << "argc = " << argc << endl; for (int i=0; i<arglst.size(); ++i) { cout << QString("argv#%1 is %2").arg(i).arg(arglst[i]) << endl; } int num = arglst[argc - 1].toInt(); cout << num * 2 << endl; }
Most applications that employ Qt types should define an object of type QCoreApplication or QApplication as early as possible
in main()
. [17]
Section 8.3 discusses the
reasons for this and the distinctions between those two types in more detail.
The QCoreApplication app
is initialized with the argument count and the argument vector.
app
silently converts the char
arrays in argv
to QStrings and stores those strings in a QStringList (Section 4.2.1).
You can then access and process the command-line arguments by using app
to call the
arguments()
function.
Using higher-level data structures like these eliminates the need to work with char
arrays, which reduces the risk of memory corruption. [18]
Notice the use of the QString function toInt()
.
[15] The process that called main()
(e.g., a command-line shell, a window manager, etc)
[16] We discuss programs that control various separate processes in Section 17.1.
[17] Applications that use only types such as QString, QStringList, and QTextStream do not need a QCoreApplication.
[18] First read Section 1.15 then Chapter 21. Bjarne Stroustrup also has some helpful advice on this subject.
Generated: 2012-03-02 | © 2012 Alan Ezust and Paul Ezust. |