Copyright © 2003,2004,2005,2006,2009,2011 Michael E. Smoot
Table of Contents
Table of Contents
TCLAP has a few key classes to be aware of.
The first is the
CmdLine
(command line) class. This class parses
the command line passed to it according to the arguments that it
contains. Arguments are separate objects that are added to the
CmdLine
object one at a time. The six
argument classes are: ValueArg
,
UnlabeledValueArg
,
SwitchArg
, MultiSwitchArg
,
MultiArg
and
UnlabeledMultiArg
.
These classes are templatized, which means they can be defined to parse
a value of any type. Once you add the
arguments to the CmdLine
object, it parses the
command line
and assigns the data it finds to the specific argument objects it
contains. Your program accesses the values parsed by
calls to the getValue()
methods of the
argument objects.
Here is a simple example ...
#include <string> #include <iostream> #include <algorithm> #include <tclap/CmdLine.h> int main(int argc, char** argv) { // Wrap everything in a try block. Do this every time, // because exceptions will be thrown for problems. try { // Define the command line object, and insert a message // that describes the program. The "Command description message" // is printed last in the help text. The second argument is the // delimiter (usually space) and the last one is the version number. // The CmdLine object parses the argv array based on the Arg objects // that it contains. TCLAP::CmdLine cmd("Command description message", ' ', "0.9"); // Define a value argument and add it to the command line. // A value arg defines a flag and a type of value that it expects, // such as "-n Bishop". TCLAP::ValueArg<std::string> nameArg("n","name","Name to print",true,"homer","string"); // Add the argument nameArg to the CmdLine object. The CmdLine object // uses this Arg to parse the command line. cmd.add( nameArg ); // Define a switch and add it to the command line. // A switch arg is a boolean argument and only defines a flag that // indicates true or false. In this example the SwitchArg adds itself // to the CmdLine object as part of the constructor. This eliminates // the need to call the cmd.add() method. All args have support in // their constructors to add themselves directly to the CmdLine object. // It doesn't matter which idiom you choose, they accomplish the same thing. TCLAP::SwitchArg reverseSwitch("r","reverse","Print name backwards", cmd, false); // Parse the argv array. cmd.parse( argc, argv ); // Get the value parsed by each arg. std::string name = nameArg.getValue(); bool reverseName = reverseSwitch.getValue(); // Do what you intend. if ( reverseName ) { std::reverse(name.begin(),name.end()); std::cout << "My name (spelled backwards) is: " << name << std::endl; } else std::cout << "My name is: " << name << std::endl; } catch (TCLAP::ArgException &e) // catch any exceptions { std::cerr << "error: " << e.error() << " for arg " << e.argId() << std::endl; } }
The output should look like:
% test1 -n mike My name is: mike % test1 -n mike -r My name (spelled backwards) is: ekim % test1 -r -n mike My name (spelled backwards) is: ekim % test1 -r PARSE ERROR: One or more required arguments missing! Brief USAGE: test1 [-r] -n <string> [--] [-v] [-h] For complete USAGE and HELP type: test1 --help % test1 --help USAGE: test1 [-r] -n <string> [--] [-v] [-h] Where: -r, --reverse Print name backwards -n <string> --name <string> (required) (value required) Name to print --, --ignore_rest Ignores the rest of the labeled arguments following this flag. -v, --version Displays version information and exits. -h, --help Displays usage information and exits. Command description message
This example shows a number of different properties of the library...
help
, version
and --
SwitchArg
s
are specified automatically. Using either the -h
or
--help
flag will cause the USAGE message to be displayed,
-v
or --version
will cause
any version information to
be displayed, and --
or
--ignore_rest
will cause the
remaining labeled arguments to be ignored. These switches are
included by default on every command line. You can disable this functionality if desired (although we don't recommend it).
How we generate the behavior behind these flags is described
later.
-s=asdf
instead of
-s asdf
, you can do so.CmdLine
, constructing the Arg
s,
or parsing the command line will throw an
ArgException
.
Arguments, whatever their type, have a few common properties. These properties are set in the constructors of the arguments.
getopt_long()
].
TCLAP is implemented entirely in header files which means you only need to include CmdLine.h to use the library.
#include <tclap/CmdLine.h>
You'll need to make sure that your compiler can see the header files. If you do the usual "make install" then your compiler should see the files by default. Alternatively, you can use the -I complier argument to specify the exact location of the libraries.
c++ -o my_program -I /some/place/tclap-1.X/include my_program.cpp
Where /some/place/tclap-1.X is the place you have unpacked the distribution.
Finally, if you want to include TCLAP as part of your software (which is perfectly OK, even encouraged) then simply copy the contents of /some/place/tclap-1.X/include (the tclap directory and all of the header files it contains) into your include directory. The necessary m4 macros for proper configuration are included in the config directory.
TCLAP was developed on Linux and MacOSX systems. It is also known to work on Windows, Sun and Alpha platforms. We've made every effort to keep the library compliant with the ANSI C++ standard so if your compiler meets the standard, then this library should work for you. Please let us know if this is not the case!
As we understand things, Visual C++ does not have the file
config.h
which is used to make platform
specific definitions. In this situation, we assume that you
have access to sstream
. Our understanding is that
this should not be a problem for VC++ 7.x. However, if this
is not the case and you need to use strstream
,
then simply tell your compiler to define the variable
HAVE_STRSTREAM
and undefine
HAVE_SSTREAM
That
should work. We think. Alternatively, just edit
the files ValueArg.h
and MultiArg.h
.
If your compiler doesn't support the using
syntax used
in UnlabeledValueArg
and
UnlabeledMultiArg
to support two stage name lookup,
then you have two options. Either comment out the statements if you don't
need two stage name lookup, or do a bunch of search and replace and use
the this
pointer syntax: e.g.
this->_ignoreable
instead
of just _ignorable
(do this for each variable
or method referenced by using
).
The CmdLine
class contains the arguments that define
the command line and manages the parsing of the command line. The
CmdLine
doesn't parse the command line itself it only
manages the parsing. The actual parsing of individual arguments occurs within
the arguments themselves. The CmdLine
keeps track of
of the required arguments, relationships
between arguments, and output generation.
SwitchArg
s are what the name implies:
simple, on/off, boolean switches. Use SwitchArg
s
anytime you want to turn
some sort of system property on or off. SwitchArg
s
don't parse a value. They return TRUE
or
FALSE
, depending on whether the switch has been found
on the command line and what the default value was defined as.
ValueArg
s are arguments that read a
value of some type
from the command line. Any time you need a file name, a number,
etc. use a ValueArg
or one of its variants.
All ValueArg
s are
templatized and will attempt to parse
the string its flag matches on the command line as the type it is
specified as. ValueArg<int>
will attempt to parse an
int, ValueArg<float>
will attempt to
parse a float, etc. If operator>>
for the specified type doesn't
recognize the string on the command line as its defined type, then
an exception will be thrown.
A MultiArg
is a ValueArg
that
can be specified more than once on a command line and instead of returning
a single value, returns a vector
of values.
Imagine a compiler that allows you to specify multiple directories to search for libraries...
% fooCompiler -L /dir/num1 -L /dir/num2 file.foo
Exceptions will occur if you try to do this
with a ValueArg
or a SwitchArg
.
In situations like this, you will want to use a
MultiArg
. A
MultiArg
is essentially a
ValueArg
that appends any
value that it matches and parses onto a vector of values. When the
getValue()
method is called, a vector of
values, instead of a single value is returned. A
MultiArg
is declared much like
a ValueArg
:
MultiArg<int> itest("i", "intTest", "multi int test", false,"int" ); cmd.add( itest );
Note that MultiArg
s can be added to the
CmdLine
in any order (unlike
UnlabeledMultiArg).
A MultiSwitchArg
is a SwitchArg
that can be specified more than once on a command line.
This can be useful
when command lines are constructed automatically from within other applications
or when a switch occurring
more than once indicates a value (-V means a little verbose -V -V -V means a lot
verbose), You can use a MultiSwitchArg
.
The call
to getValue()
for a MultiSwitchArg
returns the number (int) of times
the switch has been found on the command line in addition to the default value.
Here is an example using the default initial value of 0:
MultiSwitchArg quiet("q","quiet","Reduce the volume of output"); cmd.add( quiet );
Alternatively, you can specify your own initial value:
MultiSwitchArg quiet("q","quiet","Reduce the volume of output",5); cmd.add( quiet );
An UnlabeledValueArg
is a ValueArg
that is not identified by a flag on the command line. Instead
UnlabeledValueArg
s are identified by their position in
the argv array.
To this point all of our arguments have had labels (flags)
identifying them on the command line, but there are some
situations where flags are burdensome and not worth the effort. One
example might be if you want to implement a magical command we'll
call copy. All copy does is
copy the file specified
in the first argument to the file specified in the second argument.
We can do this using UnlabeledValueArg
s which are pretty
much just ValueArg
s without the flag specified,
which tells
the CmdLine
object to treat them accordingly.
The code would look like this:
UnlabeledValueArg<float> nolabel( "name", "unlabeled test", 3.14, "nameString" ); cmd.add( nolabel );
Everything else is handled identically to what is seen above. The
only difference to be aware of, and this is important: the order
that UnlabeledValueArgs are added to the CmdLine
is the order that they will be parsed!!!!
This is not the case for normal
SwitchArg
s and ValueArg
s.
What happens internally is the first argument that the
CmdLine
doesn't recognize is assumed to be
the first UnlabeledValueArg
and
parses it as such. Note that you are allowed to intersperse labeled
args (SwitchArgs and ValueArgs) in between
UnlabeledValueArgs
(either on the command line
or in the declaration), but the UnlabeledValueArgs
will still be parsed in the order they are added. Just remember that order is
important for unlabeled arguments.
An UnlabeledMultiArg
is an UnlabeledValueArg
that allows more than one value to be specified. Only one
UnlabeledMultiArg
can be specified per command line.
The UnlabeledMultiArg
simply reads the remaining
values from argv up until -- or the end of the array is reached.
Say you want a strange command that searches each file specified for a given string (let's call it grep), but you don't want to have to type in all of the file names or write a script to do it for you. Say,
% grep pattern *.txt
First remember that the * is handled by the shell and expanded accordingly, so what the program grep sees is really something like:
% grep pattern file1.txt file2.txt fileZ.txt
To handle situations where multiple, unlabeled arguments are needed,
we provide the UnlabeledMultiArg
.
UnlabeledMultiArg
s
are declared much like everything else, but with only a description
of the arguments. By default, if an UnlabeledMultiArg
is specified, then at least one is required to be present or an
exception will be thrown. The most important thing to remember is,
that like UnlabeledValueArg
s: order matters!
In fact, an UnlabeledMultiArg must be the last argument added to the
CmdLine!. Here is what a declaration looks like:
// // UnlabeledMultiArg must be the LAST argument added! // UnlabeledMultiArg<string> multi("file names"); cmd.add( multi ); cmd.parse(argc, argv); vector<string> fileNames = multi.getValue();
You must only ever specify one (1) UnlabeledMultiArg
.
One UnlabeledMultiArg
will read every unlabeled
Arg that wasn't already processed by a
UnlabeledValueArg
into a
vector
of type T. Any
UnlabeledValueArg
or other
UnlabeledMultiArg
specified after the first
UnlabeledMultiArg
will be ignored, and if
they are required,
exceptions will be thrown. When you call the
getValue()
method of the UnlabeledValueArg
argument,
a vector
will be returned. If you can imagine a situation where there will
be multiple args of multiple types (stings, ints, floats, etc.)
then just declare the UnlabeledMultiArg
as type
string
and parse the different values yourself or use
several UnlabeledValueArg
s.
Table of Contents
Naturally, what we have seen to this point doesn't satisfy all of our needs.
Multiple SwitchArg
s can be combined into a
single argument on the command line. If you have switches -a, -b and -c
it is valid to do either:
% command -a -b -c
or
% command -abc
or
% command -ba -c
This is to make this library more in line with the POSIX and GNU standards (as I understand them).
Suppose you have a command that must read input from one of two
possible locations, either a local file or a URL. The command
must read something, so one
argument is required, but
not both, yet neither argument is strictly necessary by itself.
This is called "exclusive or" or "XOR". To accommodate this
situation, there is now an option to add two or more
Arg
s to
a CmdLine
that are exclusively or'd with one another:
xorAdd()
. This means that exactly one of the
Arg
s must be set and no more.
xorAdd()
comes in two flavors, either
xorAdd(Arg& a, Arg& b)
to add just two Arg
s to be xor'd and
xorAdd( vector<Arg*> xorList )
to add more than two Arg
s.
ValueArg<string> fileArg("f","file","File name to read",true,"/dev/null", "filename"); ValueArg<string> urlArg("u","url","URL to load",true, "http://example.com", "URL"); cmd.xorAdd( fileArg, urlArg ); cmd.parse(argc, argv);
Once one Arg
in the xor list is matched on the
CmdLine
then the others in the xor list will be
marked as set. The question then, is how to determine which of the
Arg
s has been set? This is accomplished by calling the
isSet() method for each Arg
. If the
Arg
has been
matched on the command line, the isSet()
will return
TRUE
, whereas if the Arg
has been set as a result of matching the other Arg
that was xor'd isSet()
will
return FALSE
.
(Of course, if the Arg
was not xor'd and
wasn't matched, it will also return FALSE
.)
if ( fileArg.isSet() ) readFile( fileArg.getValue() ); else if ( urlArg.isSet() ) readURL( urlArg.getValue() ); else // Should never get here because TCLAP will note that one of the // required args above has not been set. throw("Very bad things...");
It is helpful to note that Arg
s of any type can be xor'd together.
This means that you can xor a SwitchArg
with a ValueArg
.
This is helpful in situations where one of several options is necessary and one of the options
requires additional information.
SwitchArg stdinArg("s", "stdin", "Read from STDIN", false); ValueArg<string> fileArg("f","file","File name to read",true,"/dev/null", "filename"); ValueArg<string> urlArg("u","url","URL to load",true, "http://example.com", "URL"); vector<Arg*> xorlist; xorlist.push_back(&stdinArg); xorlist.push_back(&fileArg); xorlist.push_back(&urlArg); cmd.xorAdd( xorlist );
Some commands have so many options that single flags no longer map
sensibly to the available options. In this case, it is desirable to
specify Arg
s using only long options. This one is easy to
accomplish, just make the flag value blank in the Arg
constructor. This will tell the Arg
that only the long
option should be matched and will force users to specify the long
option on the command line. The help output is updated accordingly.
ValueArg<string> fileArg("","file","File name",true,"homer","filename"); SwitchArg caseSwitch("","upperCase","Print in upper case",false);
Interface Change!!! Sorry folks, but we've changed
the interface since version 1.0.X for constraining Arg
s.
Constraints are now hidden behind the Constraint
interface. To
constrain an Arg
simply implement the interface
and specify the new class in the constructor as before.
You can still constrain Arg
s based on
a list of values. Instead of adding a vector
of
allowed values to the Arg
directly,
create a ValuesConstraint
object
with a vector
of values and add that to the
Arg
. The Arg
constructors
have been modified accordingly.
When the value for the
Arg
is parsed,
it is checked against the list of values specified in the
ValuesConstraint
.
If the value is in the list then it is accepted. If
not, then an exception is thrown. Here is a simple example:
vector<string> allowed; allowed.push_back("homer"); allowed.push_back("marge"); allowed.push_back("bart"); allowed.push_back("lisa"); allowed.push_back("maggie"); ValuesConstraint<string> allowedVals( allowed ); ValueArg<string> nameArg("n","name","Name to print",true,"homer",&allowedVals); cmd.add( nameArg );
When a ValuesConstraint
is specified,
instead of a type description being specified in the
Arg
, a
type description is created by concatenating the values in the
allowed list using operator<< for the specified type. The
help/usage for the Arg
therefore lists the
allowable values. Because of this, you might want to keep the list
relatively small, however there is no limit on this.
Obviously, a list of allowed values isn't always the best way to
constrain things. For instance, one might wish to allow only
integers greater than 0. In this case, simply create a class that
implements the Constraint<int>
interface and
checks whether the value parsed is greater than 0 (done in the
check()
method) and create your
Arg
with your new Constraint
.
New constructors have been added for each Arg
that take a CmdLine
object as an argument.
Each Arg
then
add
s itself to the CmdLine
object. There is no difference in how the Arg
is handled between this method and calling the
add()
method directly. At the moment, there is
no way to do an xorAdd()
from the constructor. Here
is an example:
// Create the command line. CmdLine cmd("this is a message", '=', "0.99" ); // Note that the following args take the "cmd" object as arguments. SwitchArg btest("B","existTestB", "exist Test B", cmd, false ); ValueArg<string> stest("s", "stringTest", "string test", true, "homer", "string", cmd ); UnlabeledValueArg<string> utest("unTest1","unlabeled test one", "default","string", cmd ); // NO add() calls! // Parse the command line. cmd.parse(argc,argv);
It is straightforward to change the output generated by
TCLAP. Either subclass the
StdOutput
class and re-implement the methods you choose,
or write your own class that implements the
CmdLineOutput
interface. Once you have done this,
then use the CmdLine
setOutput
method to tell the CmdLine
to use your new output
class. Here is a simple example:
class MyOutput : public StdOutput { public: virtual void failure(CmdLineInterface& c, ArgException& e) { cerr << "My special failure message for: " << endl << e.what() << endl; exit(1); } virtual void usage(CmdLineInterface& c) { cout << "my usage message:" << endl; list<Arg*> args = c.getArgList(); for (ArgListIterator it = args.begin(); it != args.end(); it++) cout << (*it)->longID() << " (" << (*it)->getDescription() << ")" << endl; } virtual void version(CmdLineInterface& c) { cout << "my version message: 0.1" << endl; } }; int main(int argc, char** argv) { CmdLine cmd("this is a message", ' ', "0.99" ); // set the output MyOutput my; cmd.setOutput( &my ); // proceed normally ...
See test4.cpp
in the examples directory for the full
example. NOTE: if you supply your own Output object, we
will not delete it in the CmdLine
destructor. This
could lead to a (very small) memory leak if you don't take care of the object
yourself. Also note that the failure
method is
now responsible for exiting the application (assuming that is the desired
behavior).
Help and version information is useful for nearly all command line applications
and as such we generate flags that provide those options automatically.
However, there are situations when these flags are undesirable. For these
cases we've added we've added a forth parameter to the
CmdLine
constructor. Making this boolean parameter
false will disable automatic help and version generation.
CmdLine cmd("this is a message", ' ', "0.99", false );
The --
flag is automatically included in the
CmdLine
.
As (almost) per POSIX and GNU standards, any argument specified
after the --
flag is ignored.
Almost because if an
UnlabeledValueArg
that has not been set or an
UnlabeledMultiArg
has been specified, by default
we will assign any arguments beyond the --
to the those arguments as
per the rules above. This is primarily useful if you want to pass
in arguments with a dash as the first character of the argument. It
should be noted that even if the --
flag is
passed on the command line, the CmdLine
will
still test to make sure all of the required
arguments are present.
Of course, this isn't how POSIX/GNU handle things, they explicitly
ignore arguments after the --
. To accommodate this,
we can make both UnlabeledValueArg
s and
UnlabeledMultiArg
s ignoreable in their constructors.
See the API Documentation for details.
Sometimes it's desirable to read integers formatted in decimal, hexadecimal,
and octal format. This is now possible by #defining the TCLAP_SETBASE_ZERO
directive. Simply define this directive in your code and integer arguments will be parsed
in each base.
#define TCLAP_SETBASE_ZERO 1 #include "tclap/CmdLine.h" #include <iostream> using namespace TCLAP; using namespace std; int main(int argc, char** argv) { try { CmdLine cmd("this is a message", ' ', "0.99" ); ValueArg<int> itest("i", "intTest", "integer test", true, 5, "int"); cmd.add( itest ); // // Parse the command line. // cmd.parse(argc,argv); // // Set variables // int _intTest = itest.getValue(); cout << "found int: " << _intTest << endl; } catch ( ArgException& e ) { cout << "ERROR: " << e.error() << " " << e.argId() << endl; } }
The reason that this behavior is not the default behavior for TCLAP is that the use of
setbase(0)
appears to be something of a side effect and is not necessarily how
setbase()
is meant to be used. So while we're making this functionality
available, we're not turning it on by default for fear of bad things happening in different compilers.
If you know otherwise, please let us know.
The usual C++ types (int, long, bool, etc.) are supported by TCLAP out
of the box. As
long as operator>> and operator<< are supported, other types should work fine
too, you'll just need to specify the ArgTraits
which
tells TCLAP how you expect the type to be handled.
For example, assume that you'd like to read one argument on the command line in as a
std::pair
object. All you'll need to do is tell
TCLAP whether to treat std::pair
as a
String or Value. StringLike means to treat the string on the command line as a string
and use it directly, whereas ValueLike means that a value object should be extracted from the
string using operator>>. For std::pair
we'll choose ValueLike.
To accomplish this, add the following declaration to your file:
template<class T, class U> struct ArgTraits<std::pair<T, U>> { typedef ValueLike ValueCategory; };
For complete examples see the files test11.cpp
and test12.cpp
in the examples directory.
It is traditional in Posix environments that the "-" and "--" strings are used to signify
the beginning of argument flags and long argument names. However, other environments,
namely Windows, use different strings. TCLAP allows you to
control which strings are used with #define
directives. This allows
you to use different strings based on your operating environment. Here is an example:
// // This illustrates how to change the flag and name start strings for // Windows, otherwise the defaults are used. // // Note that these defines need to happen *before* tclap is included! // #ifdef WINDOWS #define TCLAP_NAMESTARTSTRING "~~" #define TCLAP_FLAGSTARTSTRING "/" #endif #include "tclap/CmdLine.h" using namespace TCLAP; using namespace std; int main(int argc, char** argv) { // Everything else is identical! ...
Table of Contents
Like all good rules, there are many exceptions....
Ideally this library would use RTTI to return a human readable name of the type declared for a particular argument. Unfortunately, at least for g++, the names returned aren't particularly useful.
Disclaimer: Almost no one will have any use for
Visitor
s, they were
added to provide special handling for default arguments. Nothing
that Visitor
s do couldn't be accomplished
by the user after the
command line has been parsed. If you're still interested, keep
reading...
Some of you may be wondering how we get the --help
,
--version
and --
arguments to do their thing without mucking up the
CmdLine
code with lots of if
statements and type checking. This is accomplished by using a
variation on the Visitor Pattern. Actually, it may not be a Visitor
Pattern at all, but that's what inspired me.
If we want some argument to do some sort of special handling,
besides simply parsing a value, then we add a Visitor
pointer to the Arg
. More specifically, we add a
subclass of the Visitor
class. Once the argument has been successfully parsed, the
Visitor
for that argument is
called. Any data that needs to be operated on is declared in the
Visitor
constructor and then operated on in the
visit()
method. A Visitor
is added to an Arg
as the last argument in its
declaration. This may sound
complicated, but it is pretty straightforward. Let's see an
example.
Say you want to add an --authors
flag to a program that
prints the names of the authors when present. First subclass
Visitor
:
#include "Visitor.h" #include <string> #include <iostream> class AuthorVisitor : public Visitor { protected: string _author; public: AuthorVisitor(const string& name ) : Visitor(), _author(name) {} ; void visit() { cout << "AUTHOR: " << _author << endl; exit(0); }; };
Now include this class definition somewhere and go about creating
your command line. When you create the author switch, add the
AuthorVisitor
pointer as follows:
SwitchArg author("a","author","Prints author name", false, new AuthorVisitor("Homer J. Simpson") ); cmd.add( author );
Now, any time the -a
or
--author
flag is specified,
the program will print the author name, Homer J. Simpson and exit
without processing any further (as specified in the
visit()
method).
For more information, look at the API Documentation and the examples included with the distribution.
Happy coding!