Regex in C++11 vs PHP

Question

I'm new to regex and C++11. In order to match an expression like this :

TYPE SIZE NUMBER ("regina s x99");

I built a regex which looks like this one :

In my code I did this to try the regex :

std::string s("regina s x99");
std::regex rgx($RGX); //$RGX corresponds to the regex above
if (std::regex_match(s, rgx))
std::cout << "It works !" << std::endl;

This code throw a std::regex_error, but I don't know where it comes from..

Thanks,

Use GCC 4.9.0: http://stackoverflow.com/questions/15671536/why-does-this-c11-stdregex-example-throw-a-regex-error-exception — Wiktor Stribiżew, Apr 17 '15 at 10:47
@stribizhev I'm compiling with g++-4.9 and still the same error. — Nicolas Charvoz, Apr 17 '15 at 11:01
@Steephen I thought maybe the regex patern was a bit different between those two languages. — Nicolas Charvoz, Apr 17 '15 at 11:01
@NicolasCharvozKurzawa This works fine in gcc 4.9.2: http://ideone.com/REkhzI — Jonathan Mee, Apr 17 '15 at 11:03

score 1 · Answer 1 · answered Apr 17 '15 at 11:02

In C++ strings the \ character is special and needs to be escaped so that it gets passed to the regular expression engine, not interpreted by the compiler.

So you either need to use \\b:

std::regex rgx("\\b(regina|margarita|americaine|fantasia)\\b \\b(s|l|m|xl|xxl)\\b x([1-9])([0-9])");

or use a raw string, which means that \ is not special and doesn't need to be escaped:

 std::regex rgx(R"(\b(regina|margarita|americaine|fantasia)\b \b(s|l|m|xl|xxl)\b x([1-9])([0-9]))");

jpw · Accepted Answer · 2015-04-17T11:13:02.380

1

This works with g++ (4.9.2) in c++11 mode:

std::regex rgx("\\b(regina|margarita|americaine|fantasia)\\b\\s*(s|l|m|xl|xxl)\\b\\s*x([1-9]*[0-9])");

This will capture three groups: regina s 99 which matches the TYPE SIZE NUMBER pattern, while your original captured four groups regina s 9 9 and had the NUMBER as two values (maybe that was what you wanted though).

Demo on IdeOne

edited Apr 17 '15 at 11:13

answered Apr 17 '15 at 11:05

jpw

43,521
6
61
82

The last group can (and probably should) be tweaked a bit depending on the range of the NUMBER you want to match (if it has a leading 0 or not. Maybe it should be `x([1-9]+[0-9]*)` for `1 to N` – jpw Apr 17 '15 at 11:20

Steephen · Answer 3 · 2015-04-17T11:19:37.180

0

There was a typo in this line in original question:

if (std::reegex_match(s, rgx))

More over I am not sure what are you passing with this variable : $RGX

Corrected program as follows:

#include<regex>
#include<iostream>
int main()
{
    std::string s("regina s x99");
std::regex rgx("\\b(regina|margarita|americaine|fantasia)\\b \\s*(s|l|m|xl|xxl)\\b \\s*x([1-9])([0-9])"); //$RGX corresponds to the regex above
if (std::regex_match(s, rgx))
std::cout << "It works !" << std::endl;
else
  std::cout<<"No Match"<<std::endl;
}

edited Apr 17 '15 at 11:19

answered Apr 17 '15 at 10:56

Steephen

13,137
6
34
44

Yeah copy/paste does not work between VMWare and MAC Os X, there is no typo in my code, and the variable $RGX does not exist in my code, I didn't want to rewrite the regex again. I thought it was clearer this way. – Nicolas Charvoz Apr 17 '15 at 10:59
@NicolasCharvozKurzawa Thanks to know my answer helps you!! – Steephen Apr 17 '15 at 11:00
@Steephen You need to escape your slashes or this will never match. – Jonathan Mee Apr 17 '15 at 11:00
@Jonathan Mee I was just using the pattern he tried, I will correct it – Steephen Apr 17 '15 at 11:00
@JonathanMee corercted it is matching now – Steephen Apr 17 '15 at 11:02
1

@Steephen, since the OP said he gets a `regex_error` I think it's safe to assume the code compiled, and so there's no typo in the original. Also, you apparently didn't test your corrected code, it still doesn't match. – Jonathan Wakely Apr 17 '15 at 11:03
@JonathanWakely There was a typo, he corrected after see the answer!! I tested the code after correction before posting and it works – Steephen Apr 17 '15 at 11:05
I know there was a typo in **the question** but that obviously wasn't the cause of the `regex_error` exception. – Jonathan Wakely Apr 17 '15 at 11:06
@JonathanWakely I didn't mean typo is the reason for regex_error – Steephen Apr 17 '15 at 11:08

Regex in C++11 vs PHP

3 Answers3