C++ Regex not matching multiline strings

Question

I am having problems with C++0x regex when the string Im matching is a multiline string. Here is the code snippet Im trying to use:

std::smatch regMatch;
std::string data = "<key>id</key><string>1</string>\n<key>user</key><string>admin</string>";
if (std::regex_match(data, regMatch, std::regex("<key>user</key><string>(.*?)</string>"))) {
    std::cout << "Reg match: " << regMatch[1].str() << std::endl;
}

Answer 1

You should use regex_search instead of regex_match.

By the way, why not use (.*) instead of (.*?)?

Answer 2

The dot . does not match newline characters by default. You can add the switch (?s) to the beginning of the regex to switch on newline matching for the dot:

(?s)<key>user</key><string>(.*?)</string>

However, I'm not a huge fan of this because not all languages support this in their regex engines. Additionally, there might be another part of your regex pattern involving a dot that you don't want to match newlines. My preferred method is to just use a character set that includes a character class such as \s or \w along with its negated class. It's a pretty straightforward way of telling the regex to match this will match absolutely everything:

<key>user</key><string>([\w\W]*?)</string>

Maybe I'm misinterpreting how your XML is going to be parsed, but I've got to say that it's a bit odd how you intend to capture a string with the key name "user" that may or may not contain newlines (and other whitespace characters, and all other characters). Are you really okay with a user named

admin$#* &% '"; _____?