Why doesn't the icase flag for regex work correctly

Question

This function is supposed to return the first instance of a vowel in a string, but it only works on lowercase. I thought regex_constants::icase was supposed to handle uppercase letters as well. Also, feel free to suggest a better way to do this.

The following returns 0 (it now returns -1) for "HAPPY" and 3 for "GREaT":

#include <regex>

int firstVowel(std::string str)
{
    std::smatch match;
    std::regex pattern("[aeiou]", std::regex_constants::icase);
    while (std::regex_search(str, match, pattern))
    {
        return match.position();
    }
    return -1;
}

I'm getting the returns through main:

std::cout << firstVowel("HAPPY") << "\n";
std::cout << firstVowel("GREaT") << "\n";

A simple way is `"[aAeEiIoOuU]"`. Also, use `if`, not `while`. And it seems [that works](https://ideone.com/lI74AA) — Wiktor Stribiżew, Jul 12 '20 at 23:00
@WiktorStribiżew Any idea why this code doesn't work (I tested `if` as well)? — Eric Xu, Jul 12 '20 at 23:05
You are probably using a "bad" compiler with poor `std::regex` implementation. — Wiktor Stribiżew, Jul 12 '20 at 23:09
I'm compiling with clang and `HAPPY` generate `1` as return value, and `2` for `GREaT` — Alberto Sinigaglia, Jul 12 '20 at 23:10
My bad, looks like match_flags as a parameter directs the match. Hopefully `(?i)` can give you an option to test which overides the icase option compiled with the regex. — , Jul 12 '20 at 23:46

score 2 · Accepted Answer · answered Jul 12 '20 at 23:24

2

Please do not use that library icase: C++11 regex::icase inconsistent behavior it is terrible....

Instead switch the string to lowercase like so:

std::string data = "Abc";
std::transform(data.begin(), data.end(), data.begin(),
[](unsigned char c){ return std::tolower(c); });

If you are using non-standard ASCII chars than you are on your own, there is no upper to lower for Greek Letters, Chinese Characters, Romanian Letters, and the like....

answered Jul 12 '20 at 23:24

Yunfei Chen

630
1
8
20

Is there a difference in using `unsigned char` vs `char`? – Eric Xu Jul 12 '20 at 23:51
1

https://stackoverflow.com/questions/75191/what-is-an-unsigned-char, In your case either one works, but just for safety reasons, I use unsigned char.... – Yunfei Chen Jul 12 '20 at 23:53

Why doesn't the icase flag for regex work correctly

1 Answers1