public final class Pattern

java.util.regex.Pattern for the Codename One runtime: a backtracking matcher written for this module, with the JDK’s syntax and the JDK’s answers for the constructs it has.

Supported

• Literals, ., and every escape for a character: \t \n \r \f \a \e, \0nn, \xhh, \x{h...}, backslash-u with four hex digits, \cX, and \Q...\E.
• Character classes with ranges, negation, nested unions and && intersections; \d \D \w \W \s \S \h \H \v \V; \R.
• The POSIX classes (\p{Alpha}, \p{Punct}, …), the java... classes, and the Unicode properties a device can answer: L, Lu, Ll, N/Nd, IsAlphabetic, IsLetter, IsDigit, IsUppercase, IsLowercase, IsWhite_Space.
• Greedy, reluctant and possessive quantifiers.
• Capturing, non-capturing and named groups, back references by number and by name, alternation, atomic groups.
• Lookahead and bounded lookbehind, positive and negative.
• ^ $ \A \z \Z \G \b \B.
• CASE_INSENSITIVE, MULTILINE, DOTALL, COMMENTS, LITERAL, UNIX_LINES and UNICODE_CASE, as arguments and inline ((?i), (?i:...), (?-i)).

Refused

These throw PatternSyntaxException (the flags an UnsupportedOperationException) naming the construct, when the pattern is compiled. Nothing is accepted and then matched differently.

• Unicode scripts, blocks and general categories other than the ones above (\p{IsGreek}, \p{InArrows}, \p{Sc}, …).
• \X, \N{name}, \b{g}.
• CANON_EQ and UNICODE_CHARACTER_CLASS ((?U)).
• A code point above U+FFFF inside a character class.

Differences

Text is matched one UTF-16 char at a time: . and a character class take half of a surrogate pair. Which characters are letters is what isLetter in this class answers, the scripts a device font is likely to have rather than all of Unicode. A group repeated many thousand times over a long text recurses once for each repetition, as the JDK’s does, on a stack that is smaller on a device.