Hosts file ANTLR grammar

Is there an existing, working hosts file grammar on the web?

I checked out list on http://www.antlr.org/grammar/list, but I didn't find it there.

I also checked the hosts file entry in Wikipedia, and it referenced RFC 952, but I don't think that is the same format used by /windows/system32/drivers/etc/hosts.

Any grammar format is better than none, but I would prefer one in ANTLR format. This is the first time I've used any grammar generators, and I want to keep my learning curve low. I'm already planning to use ANTLR for consuming other files.

Solution

From a Microsoft page:

The HOSTS file format is the same as the format for host tables in the Version 4.3 Berkeley Software Distribution (BSD) UNIX /etc/hosts file.

And the /etc/hosts file is described here.

An example file:

#
# Table of IP addresses and hostnames
#
172.16.12.2     peanut.nuts.com peanut
127.0.0.1       localhost
172.16.12.1     almond.nuts.com almond loghost
172.16.12.4     walnut.nuts.com walnut
172.16.12.3     pecan.nuts.com pecan
172.16.1.2      filbert.nuts.com filbert
172.16.6.4      salt.plant.nuts.com salt.plant salt

A hosts file looks to be formatted like this:

each table entry in /etc/hosts contains an IP address separated by whitespace(s) from a list of hostnames associated with that address
a table entry can optionally end with zero or more alias
comments begin with #

The bold words will be the rules in the ANTLR grammar, which may look like this:

grammar Hosts;

parse
  :  tableEntry* EOF
  ;

tableEntry
  :  address hostName aliases?
     {
       System.out.println("\n== Entry ==");
       System.out.println("  address  : " + $address.text);
       System.out.println("  hostName : " + $hostName.text);
       System.out.println("  aliases  : " + $aliases.text);
     }
  ;

address
  :  Octet '.' Octet '.' Octet '.' Octet
  ;

hostName
  :  Name
  ;

aliases
  :  Name+
  ;

Name
  :  Letter+ ('.' Letter+)*
  ;

Comment
  :  '#' ~('\r' | '\n')* {$channel=HIDDEN;}
  ;

Space
  :  (' ' | '\t' | '\r' | '\n') {$channel=HIDDEN;}
  ;

Octet
  :  Digit Digit Digit
  |  Digit Digit
  |  Digit
  ;

fragment Letter
  :  'a'..'z'
  |  'A'..'Z'
  ;

fragment Digit
  :  '0'..'9'
  ;

which can be tested with the class:

import org.antlr.runtime.*;

public class Main {
  public static void main(String[] args) throws Exception {
    String source = 
        "#                                                   \n" +
        "# Table of IP addresses and Hostnames               \n" +
        "#                                                   \n" +
        "172.16.12.2     peanut.nuts.com peanut              \n" +
        "127.0.0.1       localhost                           \n" +
        "172.16.12.1     almond.nuts.com almond loghost      \n" +
        "172.16.12.4     walnut.nuts.com walnut              \n" +
        "172.16.12.3     pecan.nuts.com pecan                \n" +
        "172.16.1.2      filbert.nuts.com filbert            \n" +
        "172.16.6.4      salt.plant.nuts.com salt.plant salt   ";
    ANTLRStringStream in = new ANTLRStringStream(source);
    HostsLexer lexer = new HostsLexer(in);
    CommonTokenStream tokens = new CommonTokenStream(lexer);
    HostsParser parser = new HostsParser(tokens);
    parser.parse();
  }
}

and will produce the following output:

bart@hades:~/Programming/ANTLR/Demos/Hosts$ java -cp antlr-3.3.jar org.antlr.Tool Hosts.g
bart@hades:~/Programming/ANTLR/Demos/Hosts$ javac -cp antlr-3.3.jar *.java
bart@hades:~/Programming/ANTLR/Demos/Hosts$ java -cp .:antlr-3.3.jar Main

== Entry ==
  address  : 172.16.12.2
  hostName : peanut.nuts.com
  aliases  : peanut

== Entry ==
  address  : 127.0.0.1
  hostName : localhost
  aliases  : null

== Entry ==
  address  : 172.16.12.1
  hostName : almond.nuts.com
  aliases  : almond loghost

== Entry ==
      address  : 172.16.12.4
  hostName : walnut.nuts.com
  aliases  : walnut

== Entry ==
  address  : 172.16.12.3
  hostName : pecan.nuts.com
  aliases  : pecan

== Entry ==
  address  : 172.16.1.2
  hostName : filbert.nuts.com
  aliases  : filbert

== Entry ==
  address  : 172.16.6.4
  hostName : salt.plant.nuts.com
  aliases  : salt.plant salt

Note that this is just a quick demo: host names can contain other characters than the ones I described, to name just one shortcoming.