Regex to match ResourceBundle

I need a regular expression that will match the file name of a ResourceBundle, which follows the format name_lo_CA_le.properties. It should only match bundles that have a locale portion in their file names, and the name portion shall have no underscores.

After hours of experimentation I came up with the following:

^[a-zA-Z]+(_([a-z]{2}(_[A-Z]{0,2})?|[a-z]{0,2}(_[A-Z]{2})?){1}(_\\w*)?){1}\\.properties$

It doesn't seem to work for all cases:

"bundle.properties".match(...);               // false - correct
"bundle_.properties".match(...);              // false - correct
"bundle_en.properties".match(...);            // true - correct
"bundle__US.properties".match(...);           // true - correct
"bundle_en_US.properties".match(...);         // true - correct
"bundle_en__Windows.properties".match(...);              // false!
"bundle__US_Windows.properties".match(...);   // true - correct
"bundle_en_US_Windows.properties".match(...); // true - correct

开发者_Go百科

I have absolutely no idea how to proceed from here. Here's my reasoning behind the parenthesized part:

(...){1} matches exactly one locale portion.

(_([a-z]{2}(_[A-Z]{0,2})?|[a-z]{0,2}(_[A-Z]{2})?){1} matches exactly one of either a two-character language code and a possibly-zero-and-at-most-2-character country code or the other way around.

(_\\w*)? matches one or no variant.

Any idea how to fix and/or improve this regular expression?

---

This matched all the examples:

^[a-zA-Z\_\.]+[A-Z]{0,2}[a-zA-Z\_\.]*.properties$

---

You can try something like:

^[a-zA-Z\_\.]+[A-Z]{2}[a-zA-Z\_\.]*.properties$

---

This works for me:

public class Test {

  public static void main(String[] args) {
    String regex = "^[a-zA-Z]+(_)([a-z]{2})?(_)?([A-Z]{2})(_)?(\\w*)(\\.properties)$";

    assert "bundle.properties".matches(regex) == false;               // false - correct
    assert "bundle_.properties".matches(regex) == false;              // false - correct
    assert "bundle_en.properties".matches(regex) == false;            // false!
    assert "bundle__US.properties".matches(regex) == true;           // true - correct
    assert "bundle_en_US.properties".matches(regex) == true;         // true - correct
    assert "bundle_en__Windows".matches(regex) == false;             // false!
    assert "bundle__US_Windows.properties".matches(regex) == true;   // true - correct
    assert "bundle_en_US_Windows.properties".matches(regex) == true; // true - correct
  }
}

---

The Regex I ended up using:

^[a-zA-Z.]+(_([a-z]{2}(_[A-Z]{0,2})?|[a-z]{0,2}(_[A-Z]{2})?){1}(_\w*)?)\.properties$

It still doesn't match a locale portion with no country, as in bundle_en__Windows.properties, but it's the best I could come up with.