/** @file
$Id$
$Date$
$Author$
@copy © Marc Wäckerlin
@license LGPL, see file COPYING
$Log$
Revision 1.6 2005/01/28 07:52:00 marc
added #include to documentation
Revision 1.5 2004/12/20 13:21:49 marc
#ifndef forgotten
Revision 1.4 2004/12/20 07:40:36 marc
documentation improved, new grouping
Revision 1.3 2004/12/17 16:27:58 marc
error in documentation syntax
Revision 1.2 2004/12/16 13:09:31 marc
possibility to evaluate and extract sub expressions
Revision 1.1 2004/12/14 20:20:30 marc
initial version
*/
#ifndef __MRW__REGEXP_HPP__
#define __MRW__REGEXP_HPP__
#include
#include
#include
namespace mrw {
/** @defgroup regexp Regular Expressions
*/
//@{
/** @defgroup regexpregexp Regular Expressions
@pre #include
A simple wrapper around the C POSIX regular expression library
with a C++ Interface.
Usage sample:
@code
std::ifstream file(filename); // read from a file
mrw::RegExp findBrackets("^\\[.*\\]$"); // look for "[some text]"
for (std::string line;
mrw::getline(file, line);) // read line by line
if (findBrackets(line)) // check for regular expression
found(line); // found, do something
@endcode
*/
//@{
/** @brief a regular expression
This class manages a simple regular expression.
Usage sample:
@code
std::ifstream file(filename); // read from a file
mrw::RegExp findBrackets("^\\[.*\\]$"); // look for "[some text]"
for (std::string line;
mrw::getline(file, line);) // read line by line
if (findBrackets(line)) // check for regular expression
found(line); // found, do something
@endcode
*/
class RegExp {
public:
/// The maximum number of sub expressions that are evaluated.
static const unsigned int MAX_SUB = 99;
/** @brief flags that influence regular expressions
Flag @c newline treats a newline in the text to be compared as
dividing the text into multiple lines, so that @c $ can match
before the newline and @c ^ can match after. Also, don't
permit @c . to match a newline, and don't permit @c [^...] to
match a newline.
Otherwise, newline acts like any other ordinary character.
Flag @c nosub is used internally, don't specify it.
*/
enum Flags {
extended = REG_EXTENDED, ///< use extended regular expressions
icase = REG_ICASE, ///< ignore case in match
nosub = REG_NOSUB, ///< @b internal (report only success/fail)
newline = REG_NEWLINE ///< treat lines individually
};
/** @brief define a new regular expression
The regular expression is compiled on instanciation and can
then be matced several times on different texts.
@param pattern the regular expression pattern, see the @c man
page for POSIX regular expressions (on linux: @c
info 7 regex)
@param hassub pass @c true if you want to evaluate sub expressions
@param flags special flags, they default to extended and
should consist of the Flag values combined with @c |
flag @c nosub must not be used, because it is set
automatically if necessary
@throw mrw::invalid_argument if pattern compilation fails or @c nosub
was set in @c flags
*/
RegExp(const std::string& pattern, bool hassub=false, int flags=extended)
throw(std::exception);
/** @brief cleans up expression from memory */
~RegExp() throw();
/** @brief apply the expression, match a text against the expression
@param text a text that is matched against the regular expression
@return
- true if @c text matches
- false otherwise */
bool operator()(const std::string& text) throw(std::bad_exception);
/** @brief get the n-th sub expression of the last matched text
If the RegExp was instanciated with @c Regexp(pattern, true),
so that sub expressions are evaluated, then you can get the
n-th matched sub expression.
@param n the number of sub expression to get, get the n-th sub
expression
@throw mrw::invalid_argument if this subexpression is not
available
*/
std::string operator[](unsigned int n) const throw(std::exception);
private:
regex_t _regex;
bool _hassub;
regmatch_t _sub[MAX_SUB];
std::string _text;
};
//@}
//@}
}
#endif