Lexical analysis is the first stage of a compilation process. It's used for getting a token sequence from source code. It gets an input character sequence and finds out what the token is in the start position, whether it's a language keyword, an identifier, a constant (also called a literal), or, maybe, some error. A lexical analyzer (also known as tokenizer) sends a stream of tokens further, into a parser, which builds an AST (abstract syntax tree).

It's possible to write a lexer from scratch, but much more convenient to use any lexer generator. If we define some parsing rules, corresponding to an input language syntax, we get a complete lexical analyzer (tokenizer), which can extract tokens from an input program text and pass them to a parser.

One of such generators is Flex. In this article, we'll examine how it works in general, and observe some nontrivial nuances of developing a lexer with Flex.