ANTLR4

A (rather simple) overview

https://github.com/migulorama/AutoAnalyze

Transcript

1. ANTLR 4 A (RATHER SIMPLE) OVERVIEW By Duarte Duarte, Luís Cleto, Miguel Marques and Ruben Cordeiro COMP/MIEIC/FEUP - 2013/2014
2. WHAT'S ANTLR 4?
3. WHAT'S ANTLR 4? ANTLR 4 is a parser generator, generating parse trees automatically from a given grammar. It's a parser generator and a language tool.
4. DESIGN GOALS Ease-of-use over performance Simplicity over complexity
5. INDUSTRY ADOPTION Twitter search uses ANTLR for query parsing. Oracle uses ANTLR within the SQL Developer IDE. The NetBeans IDE parses C++ with ANTLR.
6. WHAT DOES PARSING WITH ANTLR4 MEAN?
7. WHAT DOES PARSING WITH ANTLR4 MEAN? Structure is described as a grammar. Recognizer is divided into lexer and parser. Parse tree can be auto generated.
8. WHAT DOES PARSING WITH ANTLR4 MEAN?
9. ANTLR4 ALL(*)
10. ANTLR4 ALL(*) Generates recursive descent parser like what you would build by hand Does dynamic grammar analysis while parsing saves results like a JIT does Allows for direct left recursion indirect left recursion not supported Parsing time typically near-linear
11. ANTLR4 ALL(*) By moving grammar analysis to parse-time, it's possible to avoid the common pitfalls of ambiguity and left-recursion.
12. GRAMMAR SPECIFICATION
13. GRAMMAR SPECIFICATION grammar assignment; stat: ID '=' expr ';' // match an assignment; can match "f();" | ID '=' expr ';' // oops! an exact duplicate of previous alternative | expr ';' // expression statement ; expr: ID '(' ')' | INT ; ID : [a‐z]+ ; // match one or more of any lowercase letter INT : [0‐9]+ ; // match integers The ANTLR tool generates recursive-descent parsers from grammar rules such as the ones above.
14. GENERATED PARSE TREE
15. Given this concrete structure, we can use the tree-walking mechanisms that ANTLR generates automatically, rather than writing the same tree-walking boilerplate code for each application. PARSE-TREE LISTENERS AND VISITORS
16. PARSE-TREE LISTENERS AND VISITORS Options for separation of actions from grammars in ANTLR4: Listeners (SAX style) Visitors (DOM style)
17. LISTENERS Tree traversal.
18. LISTENERS ParseTreeWalker call sequence.
19. VISITORS
20. PRACTICAL EXAMPLE BUILDING A CALCULATOR USING A VISITOR
21. PRACTICAL EXAMPLE Grammar Visitor Implementation Application
22. We built a calculator without having to insert raw actions into the grammar. The grammar is kept application independent and programming language neutral. PRATICAL EXAMPLE Java
23. SUMMARY
24. SUMMARY ANTLR4 is: Modular. Powerful.
25. SUMMARY ANTLR4's release codename is Honey Badger. It was named after the extremely resilient animal since ANTLR4 takes whatever you give it: it just doesn't give a damn!
26. SUMMARY
27. REFERENCES ANTLR4-Website Book (it's actually worth reading) http://www.antlr.org http://pragprog.com/book/tpantlr2/the-definitive-antlr-4- reference
28. QUESTIONS?