Boost C++ Libraries

The first example is the simplest possible: it only handles two options. Here's the source code (the full program is in "example/first.cpp"):

// Declare the supported options.
po::options_description desc("Allowed options");
desc.add_options()
    ("help", "produce help message")
    ("compression", po::value<int>(), "set compression level")
;

po::variables_map vm;
po::store(po::parse_command_line(ac, av, desc), vm);
po::notify(vm);    

if (vm.count("help")) {
    cout << desc << "\n";
    return 1;
}

if (vm.count("compression")) {
    cout << "Compression level was set to " 
 << vm["compression"].as<int>() << ".\n";
} else {
    cout << "Compression level was not set.\n";
}

We start by declaring all allowed options using the options_description class. The add_options method of that class returns a special proxy object that defines operator(). Calls to that operator actually declare options. The parameters are option name, information about value, and option description. In this example, the first option has no value, and the second one has a value of type int.

After that, an object of class variables_map is declared. That class is intended to store values of options, and can store values of arbitrary types. Next, the calls to store, parse_command_line and notify functions cause vm to contain all the options found on the command line.

And now, finally, we can use the options as we like. The variables_map class can be used just like std::map, except that values stored there must be retrieved with the as method shown above. (If the type specified in the call to the as method is different from the actually stored type, an exception is thrown.)

It's now a good time to try compiling the code yourself, but if you're not yet ready, here's an example session:

$ bin/gcc/debug/first
Compression level was not set.
$ bin/gcc/debug/first --help
Allowed options:
  --help                 : produce help message
  --compression arg      : set compression level
$ bin/gcc/debug/first --compression 10
Compression level was set to 10.
    

An option value, surely, can have other types than int, and can have other interesting properties, which we'll discuss right now. The complete version of the code snipped below can be found in example/options_description.cpp.

Imagine we're writing a compiler. It should take the optimization level, a number of include paths, and a number of input files, and perform some interesting work. Let's describe the options:

int opt;
po::options_description desc("Allowed options");
desc.add_options()
    ("help", "produce help message")
    ("optimization", po::value<int>(&opt)->default_value(10), 
  "optimization level")
    ("include-path,I", po::value< vector<string> >(), 
  "include path")
    ("input-file", po::value< vector<string> >(), "input file")
;

The "help" option should be familiar from the previous example. It's a good idea to have this option in all cases.

The "optimization" option shows two new features. First, we specify the address of the variable(&opt). After storing values, that variable will have the value of the option. Second, we specify a default value of 10, which will be used if no value is specified by the user.

The "include-path" option is an example of the only case where the interface of the options_description class serves only one source -- the command line. Users typically like to use short option names for common options, and the "include-path,I" name specifies that short option name is "I". So, both "--include-path" and "-I" can be used.

Note also that the type of the "include-path" option is std::vector. The library provides special support for vectors -- it will be possible to specify the option several times, and all specified values will be collected in one vector.

The "input-file" option specifies the list of files to process. That's okay for a start, but, of course, writing something like:

compiler --input-file=a.cpp
    

is a little non-standard, compared with

compiler a.cpp
    

We'll address this in a moment.

The command line tokens which have no option name, as above, are called "positional options" by this library. They can be handled too. With a little help from the user, the library can decide that "a.cpp" really means the same as "--input-file=a.cpp". Here's the additional code we need:

po::positional_options_description p;
p.add("input-file", -1);

po::variables_map vm;
po::store(po::command_line_parser(ac, av).
          options(desc).positional(p).run(), vm);
po::notify(vm);
    

The first two lines say that all positional options should be translated into "input-file" options. Also note that we use the command_line_parser class to parse the command line, not the parse_command_line function. The latter is a convenient wrapper for simple cases, but now we need to pass additional information.

By now, all options are described and parsed. We'll save ourselves the trouble of implementing the rest of the compiler logic and only print the options:

if (vm.count("include-path"))
{
    cout << "Include paths are: " 
         << vm["include-path"].as< vector<string> >() << "\n";
}

if (vm.count("input-file"))
{
    cout << "Input files are: " 
         << vm["input-file"].as< vector<string> >() << "\n";
}

cout << "Optimization level is " << opt << "\n";                

Here's an example session:

$ bin/gcc/debug/options_description --help
Usage: options_description [options]
Allowed options:
  --help                 : produce help message
  --optimization arg     : optimization level
  -I [ --include-path ] arg : include path
  --input-file arg       : input file
$ bin/gcc/debug/options_description
Optimization level is 10
$ bin/gcc/debug/options_description --optimization 4 -I foo -I another/path --include-path third/include/path a.cpp b.cpp
Include paths are: foo another/path third/include/path
Input files are: a.cpp b.cpp
Optimization level is 4

Oops, there's a slight problem. It's still possible to specify the "--input-file" option, and usage message says so, which can be confusing for the user. It would be nice to hide this information, but let's wait for the next example.

It's quite likely that specifying all options to our compiler on the command line will annoy users. What if a user installs a new library and wants to always pass an additional command line element? What if he has made some choices which should be applied on every run? It's desirable to create a config file with common settings which will be used together with the command line.

Of course, there will be a need to combine the values from command line and config file. For example, the optimization level specified on the command line should override the value from the config file. On the other hand, include paths should be combined.

Let's see the code now. The complete program is in "examples/multiple_sources.cpp". The option definition has two interesting details. First, we declare several instances of the options_description class. The reason is that, in general, not all options are alike. Some options, like "input-file" above, should not be presented in an automatic help message. Some options make sense only in the config file. Finally, it's nice to have some structure in the help message, not just a long list of options. Let's declare several option groups:

// Declare a group of options that will be 
// allowed only on command line
po::options_description generic("Generic options");
generic.add_options()
    ("version,v", "print version string")
    ("help", "produce help message")    
    ;
    
// Declare a group of options that will be 
// allowed both on command line and in
// config file
po::options_description config("Configuration");
config.add_options()
    ("optimization", po::value<int>(&opt)->default_value(10), 
          "optimization level")
    ("include-path,I", 
         po::value< vector<string> >()->composing(), 
         "include path")
    ;

// Hidden options, will be allowed both on command line and
// in config file, but will not be shown to the user.
po::options_description hidden("Hidden options");
hidden.add_options()
    ("input-file", po::value< vector<string> >(), "input file")
    ;        

Note the call to the composing method in the declaration of the "include-path" option. It tells the library that values from different sources should be composed together, as we'll see shortly.

The add method of the options_description class can be used to further group the options:

po::options_description cmdline_options;
cmdline_options.add(generic).add(config).add(hidden);

po::options_description config_file_options;
config_file_options.add(config).add(hidden);

po::options_description visible("Allowed options");
visible.add(generic).add(config);
      

The parsing and storing of values follows the usual pattern, except that we additionally call parse_config_file, and call the store function twice. But what happens if the same value is specified both on the command line and in config file? Usually, the value stored first is preferred. This is what happens for the "--optimization" option. For "composing" options, like "include-file", the values are merged.

Here's an example session:

$ bin/gcc/debug/multiple_sources
Include paths are: /opt
Optimization level is 1
$ bin/gcc/debug/multiple_sources --help
Allows options:

Generic options:
  -v [ --version ]       : print version string
  --help                 : produce help message

Configuration:
  --optimization n       : optimization level
  -I [ --include-path ] path : include path

$ bin/gcc/debug/multiple_sources --optimization=4 -I foo a.cpp b.cpp
Include paths are: foo /opt
Input files are: a.cpp b.cpp
Optimization level is 4

The first invocation uses values from the configuration file. The second invocation also uses values from command line. As we see, the include paths on the command line and in the configuration file are merged, while optimization is taken from the command line.