c++11 正则表达式和 GCC

Question

根据 https://gcc.gnu.org/onlinedocs/libstdc++/manual/status.html#status.iso.2011 C++11标准的正则表达式引擎应该在GCC中完成。现在，有人可以向我解释为什么这个简单的例子

#include <iostream>
#include <string>
#include <regex>


using namespace std;


int main ()
{
    string string_array[] = {"http://www.cplusplus.com/reference/regex/regex_match/",
                             "tcp://192.168.2.1:1234/hello/how/are/you",
                             "https://mail.google.com/mail/u/0/?tab=wm#inbox/15178022db56df29?projector=1"};
    regex e("^(?:([A-Za-z]+):)?(\\/{0,3})([0-9.\\-A-Za-z]+)(?::(\\d+))?(?:\\/([^?#]*))?(?:\\?([^#]*))?(?:#(.*))?$");

    for(int i=0; i<3; i++)
    {
        smatch sm;
        regex_match (string_array[i],sm,e);

        for (unsigned i=0; i<sm.size(); ++i)
        {
            cout << "[" << sm[i] << "] ";
        }

        cout << endl;
    }
    return 0;
}

结果是这个输出(例如注意第二行错误解析的端口号，但似乎有很多错误)

[http://www.cplusplus.com/reference/regex/regex_match/] [http] [//] [www.cplusplus.com/reference/regex] [] [regex_match/] [] [] 
[tcp://192.168.2.1:1234/hello/how/are/you] [tcp] [//] [192.168.2.1:1234/hello/how/are/you] [] [] [] [] 
[https://mail.google.com/mail/u/0/?tab=wm#inbox/15178022db56df29?projector=1] [https] [//] [mail.google.com/mail/u/0/?tab=wm] [] [] [] [inbox/15178022db56df29?projector=1] 

而它的 python 对应物

import re

string_array = ["http://www.cplusplus.com/reference/regex/regex_match/",
                         "tcp://192.168.2.1:1234/hello/how/are/you",
                         "https://mail.google.com/mail/u/0/?tab=wm#inbox/15178022db56df29?projector=1"]
e = re.compile("^(?:([A-Za-z]+):)?(\\/{0,3})([0-9.\\-A-Za-z]+)(?::(\\d+))?(?:\\/([^?#]*))?(?:\\?([^#]*))?(?:#(.*))?$");

for i in range(len(string_array)):
    m = e.match(string_array[i])
    print(m.groups())

正确打印这个？

('http', '//', 'www.cplusplus.com', None, 'reference/regex/regex_match/', None, None)
('tcp', '//', '192.168.2.1', '1234', 'hello/how/are/you', None, None)
('https', '//', 'mail.google.com', None, 'mail/u/0/', 'tab=wm', 'inbox/15178022db56df29?projector=1')

我在 archlinux 上使用 gcc 5.3.0

编辑:

我把程序改成这样，检查正则表达式 syntax_option_type 标志

#include <iostream>
#include <string>
#include <regex>


using namespace std;


int main ()
{
    string string_array[] = {"http://www.cplusplus.com/reference/regex/regex_match/",
                             "tcp://192.168.2.1:1234/hello/how/are/you",
                             "https://mail.google.com/mail/u/0/?tab=wm#inbox/15178022db56df29?projector=1"};
    regex e("^(?:([A-Za-z]+):)?(\\/{0,3})([0-9.\\-A-Za-z]+)(?::(\\d+))?(?:\\/([^?#]*))?(?:\\?([^#]*))?(?:#(.*))?$");

    for(int i=0; i<3; i++)
    {
        smatch sm;
        cout << "match: " <<regex_match (string_array[i],sm,e) << endl;

        for (unsigned i=0; i<sm.size(); ++i)
        {
            cout << "[" << sm[i].str() << "] ";
        }
    }

    cout << endl;

    switch(e.flags())
    {
        case regex_constants::basic:
            cout << "POSIX syntax was used" << endl;
            break;
        case regex_constants::awk:
            cout << "POSIX awk syntax was used" << endl;
            break;
        case regex_constants::ECMAScript:
            cout << "ECMA syntax was used" << endl;
            break;
        case regex_constants::egrep:
            cout << "POSIX egrep syntax was used" << endl;
            break;
    }

    return 0;
}

令人惊讶的是我最后得到了

match: 1
[http://www.cplusplus.com/reference/regex/regex_match/] [http] [//] [www.cplusplus.com/reference/regex] [] [regex_match/] [] [] match: 1
[tcp://192.168.2.1:1234/hello/how/are/you] [tcp] [//] [192.168.2.1:1234/hello/how/are/you] [] [] [] [] match: 1
[https://mail.google.com/mail/u/0/?tab=wm#inbox/15178022db56df29?projector=1] [https] [//] [mail.google.com/mail/u/0/?tab=wm] [] [] [] [inbox/15178022db56df29?projector=1] 
ECMA syntax was used

这似乎真的是一个编译器错误..

Answer 1

这里有两个问题:

• Match object groups()与 match_results区别
• 不应转义字符类中的字符

match object groups()在 Python 中返回以组 1 开头的所有子匹配(捕获的子字符串):

Return a tuple containing all the subgroups of the match, from 1 up to however many groups are in the pattern.

match_results从第 0 组(整场比赛)开始枚举所有组:

If successful, it is not empty and contains a series of sub_match objects: the first sub_match element corresponds to the entire match, and, if the regex expression contained sub-expressions to be matched (i.e., parentheses-delimited groups), their corresponding sub-matches are stored as successive sub_match elements in the match_results object.

第 3 组中的 \ 没有转义连字符，它会被忽略，从而在 . 和 A 之间创建一个范围(= [.-A] 在正常的正则表达式世界中)。

您不能在 POSIX 正则表达式的字符类中使用转义符号，这被认为是一个错误。将连字符放在末尾以避免对其进行转义(如 [0-9.A-Za-z-]+)。

因此，在 Python 中，使用

e = re.compile("^(?:([A-Za-z]+):)?(\\/{0,3})([0-9.A-Za-z-]+)(?::(\\d+))?(?:\\/([^?#]*))?(?:\\?([^#]*))?(?:#(.*))?$");

在 C++ 中:

regex e("^(?:([A-Za-z]+):)?(\\/{0,3})([0-9.A-Za-z-]+)(?::(\\d+))?(?:\\/([^?#]*))?(?:\\?([^#]*))?(?:#(.*))?$");
// ...
for(int i=0; i<3; i++)
{
    smatch sm;
    regex_match (string_array[i],sm,e);

    for (unsigned i=1; i<sm.size(); ++i) // Here, start with the second element
    {
        cout << "[" << sm[i] << "] ";
    }

    cout << endl;
}

比较 Python demo和 C++ demo .