Main design goals of JFlex

The main design goals of JFlex are:


List of JFlex features

See also release history

Compatible with JLex

JFlex file specification syntax is compatible with the syntax of JLex, itself inspired by lex on Unix.

The specification file can contain comments in any part.

JFlex has all the features of JLex, and more.

JFLex is fast

The lexers generated by JFlex are based on deterministic finite automates (DFA). They are fast, without expensive backtracking.

JFLex itself is fast to generate the lexer.

The scanner can have up to 64,000 DFA states.

Regexps, Macros and Predefined character classes

Regular expressions in JFlex are powerful, they support repetition, negation, union, intersection, set difference, symmetric difference, etc.

Macros are abbreviations for regular expressions, they can be used to make lexical specifications easier to read and understand. Cycles in macro definitions are detected and reported at generation time.

JFlex comes with useful predefined character classes, such as \R for any new-line character, [:uppercase:] for an upper-case word, or [:jletter:] that corresponds to Character.isJavaIdentifierStart()

Unicode and Platform independent

JFlex can scan Unicode files, with support for Unicode 7.0.

The end-of-line operator “$” is platform independent, and detects end-of-line on Windows or Unix automatically.

Easy integration with parsers

JFlex has builtin support for several parser generators:

It can also integrate with ANTLR.

Easy to use in your project

JFlex can be used

Works as expected

Easy debugging

JFlex has column and line counters.

JFlex has debugging support.

JFlex can generate a standalone-lexer, independent of any parser.