Overview of Tatoo usage

The easiest way to write an analyzer specification file for Tatoo is to use the new Tatoo ebnf file format. An eclipse plugin for these file is available.

Throught the writing of a specification file one can:

• build a lexer
• build a complete lexer+parser analyzer
• produce an html document that describe the parser table (useful among other for teaching and students)

The ebnf file is a made of sections which consist of the line name: followed by its content. The valid sections are:

• tokens and blanks where the lexical part is given
• starts, error, versions and productions which describe the grammar

From the content of the ebnf file, Tatoo generates a lexer, a parser or both.

Though the generated lexer and parser can be used directly (which is advised if the analyzer is used in a small device), the glue between them, semantic stack managment or AST generation can be generated using the tools module of Tatoo. Its specification is given with an XML file or putting types in the ebnf file (the latter is dicouraged since it prevents clean separation between semantic and syntax).

The tools XML file consists in three kinds of directives:

• the tag rule which links a lexer rule (regular expression) to a terminal; these directives are usually not needed with the ebnf format. They can also specify which rules bypass the rule selection system (this is the case by default for blanks) or rules that are only part of a token (for instance strings or comments)
• the tag terminal gives the type of terminal attributes (which is void by default);
• the tag non-terminal gives the type of non-terminal attributes (which is also void by default).