Jay

Build a JSON decoder with leex and yecc.

leex

Documentation

A regular expression based lexical analyzer generator for Erlang, similar to lex or flex.

Reads a .xrl file with Definitions, Rules, and Erlang code headings. A .erl file is generated automatically if placed in the src directory.

Generates a string/1 function that will return a list of tokens.

.xrl file structure

<Header>

Definitions.

<Macro Definitions>

Rules.

<Token Rules>

Erlang code.

<Erlang code>

The headings are mandatory and you'll get an obtuse error if you misspell one or capitalize "code".

Definitions are a mapping of name to regex.

Rules are a mapping of regex to Erlang code. There are special predefined variables that can be used in the code portion.

• TokenChars
• TokenLen
• TokenLine

The code must either be a token literal or return a token.

• {token, Token}
• {end_token, Token}
• skip_token
• {error, ErrString}

Can also push back characters for the next token.

Each block is scoped and can include variables.

yecc

Documentation

An LALR-1 parser generator for Erlang, similar to yacc. Takes a Backus-Naur Form grammar definition as input, and produces Erlang code for a parser.

An LALR-1 parser is a Look-Ahead Left to right, Rightmost derivation parser. The "1" denotes one-token lookahead.

Reads a .yrl file. A .erl file is generated automatically if placed in the src directory.

Generates a parse/1 function that will transform a list of tokens.

.yrl file structure

<Header>

Nonterminals <list>.

Terminals <list>.

Rootsymbol <value>.

<Endsymbol <value>>.

<operator declarations>

<grammar rules>

<Erlang code.>

The grammar rules are in the form lhs -> rhs : code where lhs is a non-terminal category, rhs is a sequence of one or more symbols, and code is zero or more Erlang expressions. The Erlang expressions have access to pseudo variables '$n' where n is a positive integer relating to the position of the match (indexed starting at 1).