QRegularExpression to Strip ANSI Control Codes

Question

I'm using the regular expression

"\x1B[([0-9]{1,2}(;[0-9]{1,2})?)?[m|K]"

to strip ANSI control codes out of a string. This works just fine with sed, but if I feed that string to a QRegularExpression it fails to match the control characters. I've tried using the QRegularExpression::escape() function; tried escaping just the leading '\x1B'; tried using '\e' in its place, all to no avail.

Try declaring it as `auto re = QRegularExpression { R"\x1B\[([0-9]{1,2}(;[0-9]{1,2})?)?[mK]" };`. If it does not work as expected, try using `\x{1B}` instead of `\x1B`. — Wiktor Stribiżew, May 02 '17 at 16:49
SixDegrees, I posted the answer with a bit of an explanation. Note that `[m|K]` may also match `|`. — Wiktor Stribiżew, May 03 '17 at 09:20
I posted the answer below. Once you have time, please feel free to accept/upvote (you will also get 2 points for accepting a working answer). — Wiktor Stribiżew, Aug 18 '17 at 09:52

score 1 · Answer 1 · answered May 02 '17 at 23:04

Try this

int main(int argc, char *argv[])
{
    QCoreApplication a(argc, argv);

    const char data[] = "some \x1b[31;1mtext";

    QRegularExpression re("\\x1B\\[([0-9]{1,2}(;[0-9]{1,2})?)?[m|K]");
    QRegularExpressionMatch match = re.match(data);

    if (!re.isValid()) {
        std::cout << "not valid" << std::endl;
    }
    if (match.hasMatch()) {
        std::cout << "matched" << std::endl;
    }

    return 0;
}

You should escape symbols like [ ("\\["). You can use both "\x1b" and "\\x1b" in QRegularExpression (in the first case it'll be just 0x1b symbol).

score 0 · Answer 2 · answered May 03 '17 at 09:19

You can use a raw string literal and forget about the problems with escaping regex escapes:

auto re = QRegularExpression { R"\x1B\[([0-9]{1,2}(;[0-9]{1,2})?)?[mK]" };

Here, each \ is treated as a literal backslash.

Also note that [m|K] matches 3 characters: m, K and |. Inside a positive character class, the characters are OR-ed, and | is parsed as a literal symbol.

QRegularExpression to Strip ANSI Control Codes

2 Answers2