Brief Background
A regular expression consists of a character string where some characters are given special meaning with regard to pattern matching. Regular expressions have been in use from the early days of computing, and provide a powerful and efficient way to parse, interpret and search and replace text within an application.

Supported Syntax
Within a regular expression, the following characters have special meaning:

• Positional Operators
  

  ^ matches at the beginning of a line¹
  $ matches at the end of a line²
  \A matches the start of the entire string
  \Z matches the end of the entire string
  

• One-Character Operators
  

  . matches any single character³
  \d matches any decimal digit
  \D matches any non-digit
  \n matches a newline character
  \r matches a return character
  \s matches any whitespace character
  \S matches any non-whitespace character
  \t matches a horizontal tab character
  \w matches any word (alphanumeric) character
  \W matches any non-word (alphanumeric) character
  \x matches the character x, if x is not one of the above listed escape sequences.
  

• Character Class Operator
  

  [abc] matches any character in the set a, b or c
  [^abc] matches any character not in the set a, b or c
  [a-z] matches any character in the range a to z, inclusive
  A leading or trailing dash will be interpreted literally.
  

  Within a character class expression, the following sequences have special meaning if the syntax bit RE_CHAR_CLASSES is on:
  

  [:alnum:] Any alphanumeric character
  [:alpha:] Any alphabetical character
  [:blank:] A space or horizontal tab
  [:cntrl:] A control character
  [:digit:] A decimal digit
  [:graph:] A non-space, non-control character
  [:lower:] A lowercase letter
  [:print:] Same as graph, but also space and tab
  [:punct:] A punctuation character
  [:space:] Any whitespace character, including newline and return
  [:upper:] An uppercase letter
  [:xdigit:] A valid hexadecimal digit
  

• Subexpressions and Backreferences
  

  (abc) matches whatever the expression abc would match, and saves it as a subexpression. Also used for grouping.
  (?:...) pure grouping operator, does not save contents
  (?#...) embedded comment, ignored by engine
  \n where 0 < n < 10, matches the same thing the nth subexpression matched.
  

• Branching (Alternation) Operator
  

  a|b matches whatever the expression a would match, or whatever the expression b would match.
  

• Repeating Operators
  These symbols operate on the previous atomic expression.

  ? matches the preceding expression or the null string
  * matches the null string or any number of repetitions of the preceding expression
  + matches one or more repetitions of the preceding expression
  {m} matches exactly m repetitions of the one-character expression
  {m,n} matches between m and n repetitions of the preceding expression, inclusive
  {m,} matches m or more repetitions of the preceding expression
  

• Stingy (Minimal) Matching
  If a repeating operator (above) is immediately followed by a ?, the repeating operator will stop at the smallest number of repetitions that can complete the rest of the match.
  

Unsupported Syntax
Some flavors of regular expression utilities support additional escape sequences, and this is not meant to be an exhaustive list. In the future, gnu.regexp may support some or all of the following:

(?=...) positive lookahead operator (Perl5)
(?!...) negative lookahead operator (Perl5)
(?mods) inlined compilation/execution modifiers (Perl5)
\G end of previous match (Perl5)
\b word break positional anchor (Perl5)
\B non-word break positional anchor (Perl5)
\< start of word positional anchor (egrep)
\> end of word positional anchor (egrep)
[.symbol.] collating symbol in class expression (POSIX)
[=class=] equivalence class in class expression (POSIX)

Java Integration
In a Java environment, a regular expression operates on a string of Unicode characters, represented either as an instance of java.lang.String or as an array of the primitive char type. This means that the unit of matching is a Unicode character, not a single byte. Generally this will not present problems in a Java program, because Java takes pains to ensure that all textual data uses the Unicode standard.

Because Java string processing takes care of certain escape sequences, they are not implemented in gnu.regexp. You should be aware that the following escape sequences are handled by the Java compiler if found in the Java source:

\b backspace
\f form feed
\n newline
\r carriage return
\t horizontal tab
\" double quote
\' single quote
\\ backslash
\xxx character, in octal (000-377)
\uxxxx Unicode character, in hexadecimal (0000-FFFF)

The POSIX character classes (above), and the equivalent shorthand escapes (\d, \w and the like) are implemented to use the java.lang.Character static functions whenever possible. For example, \w and [:alnum:] (the latter only from within a class expression) will invoke the Java function Character.isLetterOrDigit() when executing. It is always better to use the POSIX expressions than a range such as [a-zA-Z0-9], because the latter will not match any letter characters in non-ISO 9660 encodings (for example, the umlaut character, "ü").

Reference Material

• Print Books and Publications
  Friedl, Jeffrey E.F., Mastering Regular Expressions. O'Reilly & Associates, Inc., Sebastopol, California, 1997.
  

• Software Manuals and Guides
  Berry, Karl and Hargreaves, Kathryn A., GNU Info Regex Manual Edition 0.12a, 19 September 1992.
  perlre(1) man page (Perl Programmer's Reference Guide)
  regcomp(3) man page (GNU C)
  gawk(1) man page (GNU utilities)
  sed(1) man page (GNU utilities)
  ed(1) man page (GNU utilities)
  grep(1) man page (GNU utilities)
  regexp(n) and regsub(n) man pages (TCL)
  

Notes
¹ but see the REG_NOTBOL and REG_MULTILINE flags
² but see the REG_NOTEOL and REG_MULTILINE flags
³ but see the REG_MULTILINE flag