c++ 11 regex error

Question

Just an example code from C++ Primer 5th Edition: 17.3.3. Using the Regular Expression Library

Main file main.cpp:

#include <iostream>
#include "regexcase.h"
using namespace std;

int main() {
    using_regex();
    return 0;
}

Header file regexcase.h:

#ifndef REGEXCASE_H_
#define REGEXCASE_H_

#include <regex>
#include <string>

void using_regex();
std::string parseCode(std::regex_constants::error_type etype);

#endif /* REGEXCASE_H_ */

Source file regexcase.cpp:

#include "regexcase.h"
#include <iostream>
using namespace std;

void using_regex() {
    // look for words that violate a well-known spelling rule of thumb, "i before e, except after c":
    // find the characters ei that follow a character other than c
    string pattern("[^c]ei");
    // we want the whole word in which our pattern appears
    pattern = "[a-zA-Z]*" + pattern + "[a-zA-Z]*";  //[a-zA-Z]*   [[:alpha:]]*
    try {
        regex r(pattern, regex_constants::extended);    // construct a regex to find pattern                // , regex_constants::extended
        smatch results;     // define an object to hold the results of a search
        // define a string that has text that does and doesn't match pattern
        string test_str = "receipt freind theif receive";
        // use r to find a match to pattern in test_str
        if (regex_search(test_str, results, r)) // if there is a match
            cout << results.str() << endl;      // print the matching word
        else
            cout << "no match for " << pattern << endl;
    } catch (regex_error &e) {
        cout << "what: " << e.what() << "; code: " << parseCode(e.code()) << endl;
    }
}

string parseCode(regex_constants::error_type etype) {
    switch (etype) {
    case regex_constants::error_collate:
        return "error_collate: invalid collating element request";
    case regex_constants::error_ctype:
        return "error_ctype: invalid character class";
    case regex_constants::error_escape:
        return "error_escape: invalid escape character or trailing escape";
    case regex_constants::error_backref:
        return "error_backref: invalid back reference";
    case regex_constants::error_brack:
        return "error_brack: mismatched bracket([ or ])";
    case regex_constants::error_paren:
        return "error_paren: mismatched parentheses(( or ))";
    case regex_constants::error_brace:
        return "error_brace: mismatched brace({ or })";
    case regex_constants::error_badbrace:
        return "error_badbrace: invalid range inside a { }";
    case regex_constants::error_range:
        return "erro_range: invalid character range(e.g., [z-a])";
    case regex_constants::error_space:
        return "error_space: insufficient memory to handle this regular expression";
    case regex_constants::error_badrepeat:
        return "error_badrepeat: a repetition character (*, ?, +, or {) was not preceded by a valid regular expression";
    case regex_constants::error_complexity:
        return "error_complexity: the requested match is too complex";
    case regex_constants::error_stack:
        return "error_stack: insufficient memory to evaluate a match";
    default:
        return "";
    }
}

The output of calling using_regex(); is what: regex_error; code: error_brack: mismatched bracket([ or ])

It seems that the regex can't parse the bracket.

Refer to Answers in this question, I use regex_constants::extended to initialize the regex object, which then is regex r(pattern, regex_constants::extended);

Then the output is no match for [[:alpha:]]*[^c]ei[[:alpha:]]*

It seems that the regex can't match the pattern.

Then I use [a-zA-Z]* to replace character class [[:alpha:]]* (with regex_constants::extended still set). The output still is no match for [a-zA-Z]*[^c]ei[a-zA-Z]*

Platform: windows

Tools used: Eclipse for C/C++; MinGW (g++ --version: g++ 4.7.2)

EDIT: Thanks @sharth, add main file to complete the code.

Solution

I just did a test using libc++ and clang++. This works as expected. Here's my main:

int main() {
    string test_str = "receipt freind theif receive";
    string pattern = "[a-zA-Z]*[^c]ei[a-zA-Z]*";

    try {
        regex r(pattern, regex_constants::extended);
        smatch results;

        if (regex_search(test_str, results, r))
            cout << results.str() << endl;
        else
            cout << "no match for " << pattern << endl;
    } catch (regex_error &e) {
        cout << "what: " << e.what() << "; code: " << parseCode(e.code()) << endl;
    }
}

Output:

freind

---

On the other hand GCC 4.7.2, gives this result:

no match for [a-zA-Z]*[^c]ei[a-zA-Z]*

This is because in GCC 4.7.2's libstdc++, they still don't implement regex. Here's the implementation of regex_search:

template<typename _Bi_iter, typename _Allocator, typename _Ch_type, typename _Rx_traits>
inline bool regex_search(_Bi_iter __first, _Bi_iter __last, match_results<_Bi_iter, _Allocator>& __m, const basic_regex<_Ch_type, _Rx_traits>& __re, regex_constants::match_flag_type __flags) {
    return false;
}

---

And just to note, it is very helpful to include a small program that readers could compile. That way there is no confusion about what code is being run.

Comments

[zyy7259]

So what shall I do? Use boost::regex instead, shall I? (as @Maxim Yegorushkin said above)

[Bill Lynch]

Boost's implementation would work fine. In fact, I just did a quick test using Boost's 1.53.0 and GCC 4.7.2 and it worked the same as libc++.

[Admin]

What version of clang? The above doesn't work on 3.3 rc2.

[Bill Lynch]

@wvxvw: I was likely using whatever version of clang was included with OS X at the time. The larger issue, however, is that libstdc++ does not include support for regex. So, on OS X, you can do -stdlib=libc++ -std=c++11. On Linux, you will likely need to use boost's implementation.