mysql-parser 0.0.3

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: acd650741a0f7b5b7631b03b8fe200494a4265df
+  data.tar.gz: f543002b4a58e312a8934fd6610d2dbed5a36f59
+SHA512:
+  metadata.gz: b3968deb3c2ad1c145263086292524a48661d8e7e5952ce2da4d8117cfe64bd178f563c518d3321ac77bc2c5f369f7da1bcd1b8933229534a96295b5b1be02fc
+  data.tar.gz: 8da6a07a5ecf645c069373dc334e35223455d7f64693a1d8de7d69e6ecfe0bfd475bf6bc7908f073c5300ccc828614807af8f0dabd0e84a3c2f46cbebf7eed06

data/README.md

@@ -0,0 +1,131 @@
+MySQLParser
+===========
+
+This is a library to parse SQL commands. The only commands that are currently
+supported are ddl statements, specifically CREATE TABLE, ALTER TABLE, DROP VIEW,
+and DROP TABLE.
+
+Installation
+------------
+
+In command line:
+
+    > gem install mysql-parser.x.x.x.gem
+
+Usage
+-----
+
+In ruby:
+
+    > require 'mysql-parser'
+    > MySQLParser.new.parse "ALTER TABLE `table` DROP INDEX abc, DROP INDEX def"
+    => {:tree=><root: [<S: [" "]>, <r_commands: [<r_ALTER_TABLE: ["ALTER", <S: [" "]>, <r_ONLINE_OFFLINE: []>, <r_opt_IGNORE: []>, "TABLE", <S: [" "]>, <r_tbl_name: [<r_tbl_name_int: [<ident: ["`", <opt_ident_in_backtick: [<opt_ident_in_backtick: []>, "table"]>, "`", <S: [" "]>]>]>]>, <r_opt_alter_commands: [<r_comma_separated_alter_specification: [<r_comma_separated_alter_specification: [<r_alter_specification: ["DROP", <S: [" "]>, <r_INDEX_or_KEY: ["INDEX", <S: [" "]>]>, <r_index_name: [<ident: [<raw_ident: ["abc", <S: [" "]>]>]>]>]>]>, <comma: [",", <S: [" "]>]>, <r_alter_specification: ["DROP", <S: [" "]>, <r_INDEX_or_KEY: ["INDEX", <S: [" "]>]>, <r_index_name: [<ident: [<raw_ident: ["def", <S: [" "]>]>]>]>]>]>]>, <r_opt_after_alter: []>, <r_opt_PARTITION_options: []>]>]>]>, :state=>{}}
+
+Files
+-----
+
+### mysql.rex.rb
+This file is a lexer. It determines that `DROP` is a command, and
+that `'abcdef'` is a string. Most of this file is auto-generated
+by `bin/generate-literal` which reads `mysql.y.rb` and generates the required
+literals automatically. Therefore this file will not normally need to be edited.
+The following reasons might be reason to edit it:
+
+1. Creating a synonym
+2. Creating a long literal[1]
+3. Creating a literal which doesn't exist in the parser already
+
+[1] a long literal is a literal which is needed for a special purpose, for example,
+because it consists of spaces and needs a synonym to be assigned to it.
+Normally, however, we do not need them.
+
+#### Convention
+
+1. `S_...` means some symbol
+2. `A_...` means some state
+3. `L_...` means long literal
+4. Everything else is literal
+
+### mysql.y.rb
+This is the main file of this library. It contains all grammar and associated
+actions.
+
+#### Grammar
+
+The original MySQL grammar can be found at
+https://github.com/twitter/mysql/blob/master/sql/sql_yacc.yy. The documentation
+can be found under https://dev.mysql.com/doc/refman/5.6/en/
+
+##### Conflict
+`sql_yacc.yy` is not perfect. It contains a lot of conflicts. When translating
+to `mysql.y.rb`, please resolve those conflicts so that the rules can be
+predictable.
+
+#### Debugging
+
+Debugging `mysql.y.rb` can be done by setting `@yydebug` to `true` in
+`initialize`.
+
+#### Literals
+
+Introduction of a new literals can be done by just _using_ it. `bin/generate-literal.rb`
+will read `mysql.y.rb` and make sure that new literals are created.
+Similarly, removing any literal can be done by just _not using_ it.
+
+#### Convention
+
+Following is general convention.
+
+1. Space is `S`
+2. `opt_...` means that the rule is optional. The first branch should be
+empty.
+3. `{comma, space}_separated_...` means a collection of items separated by
+a separator (comma or space)
+
+### bin/generate-literal.rb
+
+This script scans `mysql.y.rb` (or actually, `parser.output` which is
+generated from `mysql.y.rb`) and adds new literals that don't exist, and
+removes literals that are unused.
+
+### bin/runner
+
+This is a REPL. Just input whatever and send an end-of-file character
+(`ctrl-d`) to let the script process the input. To exit, just terminate
+using `ctrl-c`.
+
+    > DROP TABLE
+
+
+    table
+    ^D
+    parse error on value "table" (TABLE)
+    >
+
+### bin/sanity_check
+
+This script makes sure that there is no skipped action and that all action
+names are correct.
+
+Development
+-----------
+
+After changing `mysql.rex.rb` or `mysql.y.rb`, run `rake generate` to
+generate the real lexer and parser. Run `rake spec` to run all test cases
+
+License
+=======
+
+    Copyright 2015 Square, Inc.
+
+    Licensed under the Apache License, Version 2.0 (the "License");
+    you may not use this file except in compliance with the License.
+    You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+    Unless required by applicable law or agreed to in writing, software
+    distributed under the License is distributed on an "AS IS" BASIS,
+    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+    See the License for the specific language governing permissions and
+    limitations under the License.

data/bin/generate-literal

@@ -0,0 +1,72 @@
+#!/usr/bin/ruby
+
+require 'set'
+require_relative '../lib/helper'
+
+parser_filename = 'mysql.y.rb'
+lexer_filename = 'mysql.rex.rb'
+output_filename = 'parser.output'
+
+START = 0
+STOP = -1
+
+def prepare_autogen(filename)
+  FileUtil.read_three_parts(
+    filename,
+    '# BEGIN LITERAL (AUTO-GENERATED)',
+    '# END LITERAL (AUTO-GENERATED)'
+  )
+end
+
+def diff(new_list, old_list)
+  new_set = new_list.to_set
+  old_set = old_list.to_set
+  added = new_set - old_set
+  removed = old_set - new_set
+  if !added.empty?
+    puts "\n\n======= ADDED ======="
+    added.each do |t|
+      puts t
+    end
+    puts "\n\n"
+  end
+  if !removed.empty?
+    puts "\n\n======= REMOVED ======="
+    removed.each do |t|
+      puts t
+    end
+    puts "\n\n"
+  end
+end
+
+old_lex, lines_before_lex, lines_after_lex = prepare_autogen(lexer_filename)
+old_parse, lines_before_parse, lines_after_parse = prepare_autogen(parser_filename)
+arr = FileUtil.read_three_parts(
+  output_filename,
+  '**Terminals, with rules where they appear',
+  '--------- State ---------'
+)[0].map { |line|
+  result = /\s+(\w+)\s+\(/.match(line)
+  result ? result[1] : ''
+}.select { |line| !line.empty? }.sort.reverse.reject { |w| w == 'error' }
+
+literals = arr.reject { |w|
+  w == "S" || w.start_with?("S_") || w.start_with?("L_")
+}
+
+File.open(lexer_filename, 'w') do |f|
+  new_lex = literals.map{ |t| ":A_NIL #{t}\\b { [:#{t}, text] }\n" }
+  diff(new_lex, old_lex)
+  f.write((lines_before_lex + new_lex + lines_after_lex).join)
+end
+
+File.open(parser_filename, 'w') do |f|
+  name = 'dot'
+  new_parse = (["  #{name} :\n",
+                "    #{arr[0]} { call(:#{name}, :#{arr[0]}, val) }\n"] +
+               arr[START+1..STOP].map{ |t|
+                 "  | #{t} { call(:#{name}, :#{t}, val) }\n"
+               })
+  diff(new_parse, old_parse)
+  f.write((lines_before_parse + new_parse + lines_after_parse).join)
+end

data/bin/runner

@@ -0,0 +1,13 @@
+#!/usr/bin/ruby
+
+require_relative '../lib/mysql-parser'
+
+while true do
+  begin
+    print "> "
+    puts (MySQLParser.new.parse ARGF.read)
+  rescue => e
+    puts e
+    puts e.backtrace
+  end
+end

data/bin/sanity_check

@@ -0,0 +1,73 @@
+#!/usr/bin/ruby
+
+require_relative '../lib/helper'
+
+parser_filename = 'mysql.y.rb'
+output_filename = 'parser.output'
+
+arr = FileUtil.read_three_parts(
+  output_filename,
+  '**Terminals, with rules where they appear',
+  '--------- State ---------'
+)[0].map { |line|
+  result = /\s+(\w+)\s+\(/.match(line)
+  result ? result[1] : ''
+}.select { |line| !line.empty? }.sort.reverse.reject { |w| w == 'error' }
+
+literals = arr.reject { |w| /^[A-Z\d_]+$/ =~ w }
+if !literals.empty?
+  $stderr.puts literals
+  raise 'unrecognized literals'
+end
+
+lines = []
+
+File.open(parser_filename, 'r') do |f|
+  f.each_line do |line|
+    lines << line
+  end
+end
+
+raw_content = lines.join
+  .gsub(/^---- header.*/m, '') # elimate code section
+
+content = raw_content
+content = content.gsub(/#.*?$/m, '') # eliminate comments
+
+res = /:\s[^{]*?\|/m.match(content)
+if res
+  $stderr.puts res
+  raise 'First case should have action'
+end
+
+res = /\|[^{]*?\|/m.match(content)
+if res
+  $stderr.puts res
+  raise 'Middle case should have action'
+end
+
+res = /\|[^{]*?:\s/m.match(content)
+if res
+  $stderr.puts res
+  raise 'Last case should have action'
+end
+
+res = /:\s[^{]*?:/m.match(content)
+if res
+  $stderr.puts res
+  raise 'One case should have action'
+end
+
+content = content.gsub(/call\(:(.*?), (.*?), val\)/, '\1')
+
+# this scan will not match "dot" which is the last rule
+# this is okay as "dot" is auto generated and is always correct.
+content.scan(/(?=^\s*(\S*?)(\s*?:\s[^:]+))/m).each do |group|
+  all_branches = group[1].count "{"
+  count_matched = group[1].scan(/\{\s*#{group[0]}/).count
+  count_raise = group[1].scan(/\{\s*raise/).count
+  if all_branches != count_matched + count_raise
+    $stderr.puts "#{group}"
+    raise 'name does not match action symbol'
+  end
+end

data/lib/ast.rb

@@ -0,0 +1,110 @@
+class AST
+  def initialize(a_name, a_subname, a_val)
+    @name = a_name
+    @subname = a_subname
+    @val = a_val
+  end
+  
+  def update(a_name, a_subname, a_val)
+    initialize(a_name, a_subname, a_val)
+  end
+  
+  def match(options)
+    (
+      {top: true}.merge(options)[:top] &&
+      @name == options[:name] &&
+      (options[:subname].nil? || @subname == options[:subname])
+    )
+  end
+
+  def find_all(options={})
+    ret = []
+    ret << self if match options
+    sub_options = options.merge top: true
+    @val.each do |v|
+      if v.is_a? AST
+        ret.concat (v.find_all sub_options)
+      end
+    end
+    ret
+  end
+
+  def find_left(options={})
+    return self if match options
+    sub_options = options.merge top: true
+    @val.each do |v|
+      if v.is_a? AST
+        ret = v.find_left sub_options
+        return ret if ret
+      end
+    end
+    nil
+  end
+
+  def find_top(options={})
+    return self if match options
+    sub_options = options.merge top: true
+    @val.each do |v|
+      if v.is_a? AST
+        return v if v.match sub_options
+      end
+    end
+    nil
+  end
+
+  def name
+    @name
+  end
+  
+  def subname
+    @subname
+  end
+
+  def val
+    @val
+  end
+
+  def val=(a_val)
+    @val = a_val
+  end
+
+  def eval
+    to_s.to_f
+  end
+
+  def to_list
+    to_list_helper.flatten
+  end
+
+  def to_s
+    @val.map { |v| v.to_s }.join
+  end
+  
+  def norm_name
+    s_all = to_s.strip
+    node = find_left(name: :r_tbl_name) || find_left(name: :ident)
+    s = node.to_s.strip
+    raise 'Internal Error: trying to normalize not-a-name' if s_all != s
+    if s.start_with? '`'
+      s[1..-2] # strip ` from both beginning and end
+    else
+      s
+    end
+  end
+
+  def inspect
+    "<#{@name}: #{@val}>"
+  end
+  
+  protected
+  
+  def to_list_helper
+    @val.map { |v|
+      if v.is_a?(AST) && v.name == @name
+        v.to_list_helper
+      else
+        [v]
+      end
+    }
+  end
+end

data/lib/helper.rb

@@ -0,0 +1,32 @@
+module FileUtil
+  def self.read_three_parts(filename, begin_middle, begin_after)
+    lines_before = []
+    lines_middle = []
+    lines_after = []
+    File.open(filename, 'r') do |f|
+      state = :BEFORE
+      f.each_line do |line|
+        case line.strip
+          when begin_middle
+            lines_before << line
+            state = :MIDDLE
+            next
+          when begin_after
+            lines_after << line
+            state = :AFTER
+            next
+        end
+        case state
+          when :BEFORE
+            lines_before << line
+          when :MIDDLE
+            lines_middle << line
+          when :AFTER
+            lines_after << line
+        end
+      end
+    end
+    
+    return lines_middle, lines_before, lines_after
+  end
+end

data/lib/lexer.rb

@@ -0,0 +1,724 @@
+#--
+# DO NOT MODIFY!!!!
+# This file is automatically generated by rex 1.0.5
+# from lexical definition file "mysql.rex.rb".
+#++
+
+require 'racc/parser'
+class MySQLParser < Racc::Parser
+  require 'strscan'
+
+  class ScanError < StandardError ; end
+
+  attr_reader   :lineno
+  attr_reader   :filename
+  attr_accessor :state
+
+  def scan_setup(str)
+    @ss = StringScanner.new(str)
+    @lineno =  1
+    @state  = nil
+  end
+
+  def action
+    yield
+  end
+
+  def scan_str(str)
+    scan_setup(str)
+    do_parse
+  end
+  alias :scan :scan_str
+
+  def load_file( filename )
+    @filename = filename
+    open(filename, "r") do |f|
+      scan_setup(f.read)
+    end
+  end
+
+  def scan_file( filename )
+    load_file(filename)
+    do_parse
+  end
+
+
+  def next_token
+    return if @ss.eos?
+    
+    # skips empty actions
+    until token = _next_token or @ss.eos?; end
+    token
+  end
+
+  def _next_token
+    text = @ss.peek(1)
+    @lineno  +=  1  if text == "\n"
+    token = case @state
+    when nil
+      case
+      when (text = @ss.scan(/\A/i))
+         action { @state = :A_NIL; nil }
+
+      else
+        text = @ss.string[@ss.pos .. -1]
+        raise  ScanError, "can not match: '" + text + "'"
+      end  # if
+
+    when :A_NIL
+      case
+      when (text = @ss.scan(/ *\/\*/i))
+         action { @state = :A_REM_MULTI; [:S_REM_IN, [text, ' /* ']] }
+
+      when (text = @ss.scan(/ *(\#|--)/i))
+         action { @state = :A_REM_INLINE; [:S_REM_IN, '-- '] }
+
+      when (text = @ss.scan(/`/i))
+         action { @state = :A_BACKTICK; [:S_BACKTICK_IN, text] }
+
+      when (text = @ss.scan(/"/i))
+         action { @state = :A_DOUBLEQUOTE; [:S_DOUBLEQUOTE_IN, text] }
+
+      when (text = @ss.scan(/'/i))
+         action { @state = :A_SINGLEQUOTE; [:S_SINGLEQUOTE_IN, text] }
+
+      when (text = @ss.scan(/TRUE\b/i))
+         action { [:S_ONE, text] }
+
+      when (text = @ss.scan(/FALSE\b/i))
+         action { [:S_ZERO, text] }
+
+      when (text = @ss.scan(/BOOLEAN\b/i))
+         action { [:TINYINT, text] }
+
+      when (text = @ss.scan(/CHARSET\b/i))
+         action { [:L_CHARACTER_SET, text] }
+
+      when (text = @ss.scan(/CHARACTER[ \t\n]+SET\b/i))
+         action { [:L_CHARACTER_SET, text] }
+
+      when (text = @ss.scan(/FROM\b/i))
+         action { [:FROM, text] }
+
+      when (text = @ss.scan(/WHERE\b/i))
+         action { [:WHERE, text] }
+
+      when (text = @ss.scan(/ZEROFILL\b/i))
+         action { [:ZEROFILL, text] }
+
+      when (text = @ss.scan(/YEAR\b/i))
+         action { [:YEAR, text] }
+
+      when (text = @ss.scan(/WITH\b/i))
+         action { [:WITH, text] }
+
+      when (text = @ss.scan(/VIEW\b/i))
+         action { [:VIEW, text] }
+
+      when (text = @ss.scan(/VARCHAR\b/i))
+         action { [:VARCHAR, text] }
+
+      when (text = @ss.scan(/VARBINARY\b/i))
+         action { [:VARBINARY, text] }
+
+      when (text = @ss.scan(/VALUES\b/i))
+         action { [:VALUES, text] }
+
+      when (text = @ss.scan(/UTF8MB4\b/i))
+         action { [:UTF8MB4, text] }
+
+      when (text = @ss.scan(/UTF8MB3\b/i))
+         action { [:UTF8MB3, text] }
+
+      when (text = @ss.scan(/UTF8\b/i))
+         action { [:UTF8, text] }
+
+      when (text = @ss.scan(/USING\b/i))
+         action { [:USING, text] }
+
+      when (text = @ss.scan(/UPDATE\b/i))
+         action { [:UPDATE, text] }
+
+      when (text = @ss.scan(/UNSIGNED\b/i))
+         action { [:UNSIGNED, text] }
+
+      when (text = @ss.scan(/UNIQUE\b/i))
+         action { [:UNIQUE, text] }
+
+      when (text = @ss.scan(/UNION\b/i))
+         action { [:UNION, text] }
+
+      when (text = @ss.scan(/UNDEFINED\b/i))
+         action { [:UNDEFINED, text] }
+
+      when (text = @ss.scan(/TRUNCATE\b/i))
+         action { [:TRUNCATE, text] }
+
+      when (text = @ss.scan(/TO\b/i))
+         action { [:TO, text] }
+
+      when (text = @ss.scan(/TINYTEXT\b/i))
+         action { [:TINYTEXT, text] }
+
+      when (text = @ss.scan(/TINYINT\b/i))
+         action { [:TINYINT, text] }
+
+      when (text = @ss.scan(/TINYBLOB\b/i))
+         action { [:TINYBLOB, text] }
+
+      when (text = @ss.scan(/TIMESTAMP\b/i))
+         action { [:TIMESTAMP, text] }
+
+      when (text = @ss.scan(/TIME\b/i))
+         action { [:TIME, text] }
+
+      when (text = @ss.scan(/THAN\b/i))
+         action { [:THAN, text] }
+
+      when (text = @ss.scan(/TEXT\b/i))
+         action { [:TEXT, text] }
+
+      when (text = @ss.scan(/TEMPTABLE\b/i))
+         action { [:TEMPTABLE, text] }
+
+      when (text = @ss.scan(/TEMPORARY\b/i))
+         action { [:TEMPORARY, text] }
+
+      when (text = @ss.scan(/TABLESPACE\b/i))
+         action { [:TABLESPACE, text] }
+
+      when (text = @ss.scan(/TABLE\b/i))
+         action { [:TABLE, text] }
+
+      when (text = @ss.scan(/SUBPARTITION\b/i))
+         action { [:SUBPARTITION, text] }
+
+      when (text = @ss.scan(/STORAGE\b/i))
+         action { [:STORAGE, text] }
+
+      when (text = @ss.scan(/SQL\b/i))
+         action { [:SQL, text] }
+
+      when (text = @ss.scan(/SPATIAL\b/i))
+         action { [:SPATIAL, text] }
+
+      when (text = @ss.scan(/SMALLINT\b/i))
+         action { [:SMALLINT, text] }
+
+      when (text = @ss.scan(/SIMPLE\b/i))
+         action { [:SIMPLE, text] }
+
+      when (text = @ss.scan(/SET\b/i))
+         action { [:SET, text] }
+
+      when (text = @ss.scan(/SELECT\b/i))
+         action { [:SELECT, text] }
+
+      when (text = @ss.scan(/SECURITY\b/i))
+         action { [:SECURITY, text] }
+
+      when (text = @ss.scan(/ROW_FORMAT\b/i))
+         action { [:ROW_FORMAT, text] }
+
+      when (text = @ss.scan(/RESTRICT\b/i))
+         action { [:RESTRICT, text] }
+
+      when (text = @ss.scan(/REPLACE\b/i))
+         action { [:REPLACE, text] }
+
+      when (text = @ss.scan(/REPAIR\b/i))
+         action { [:REPAIR, text] }
+
+      when (text = @ss.scan(/REORGANIZE\b/i))
+         action { [:REORGANIZE, text] }
+
+      when (text = @ss.scan(/RENAME\b/i))
+         action { [:RENAME, text] }
+
+      when (text = @ss.scan(/REMOVE\b/i))
+         action { [:REMOVE, text] }
+
+      when (text = @ss.scan(/REFERENCES\b/i))
+         action { [:REFERENCES, text] }
+
+      when (text = @ss.scan(/REDUNDANT\b/i))
+         action { [:REDUNDANT, text] }
+
+      when (text = @ss.scan(/REBUILD\b/i))
+         action { [:REBUILD, text] }
+
+      when (text = @ss.scan(/REAL\b/i))
+         action { [:REAL, text] }
+
+      when (text = @ss.scan(/PRIMARY\b/i))
+         action { [:PRIMARY, text] }
+
+      when (text = @ss.scan(/PASSWORD\b/i))
+         action { [:PASSWORD, text] }
+
+      when (text = @ss.scan(/PARTITIONING\b/i))
+         action { [:PARTITIONING, text] }
+
+      when (text = @ss.scan(/PARTITION\b/i))
+         action { [:PARTITION, text] }
+
+      when (text = @ss.scan(/PARTIAL\b/i))
+         action { [:PARTIAL, text] }
+
+      when (text = @ss.scan(/PARSER\b/i))
+         action { [:PARSER, text] }
+
+      when (text = @ss.scan(/PACK_KEYS\b/i))
+         action { [:PACK_KEYS, text] }
+
+      when (text = @ss.scan(/ORDER\b/i))
+         action { [:ORDER, text] }
+
+      when (text = @ss.scan(/OR\b/i))
+         action { [:OR, text] }
+
+      when (text = @ss.scan(/OPTION\b/i))
+         action { [:OPTION, text] }
+
+      when (text = @ss.scan(/OPTIMIZE\b/i))
+         action { [:OPTIMIZE, text] }
+
+      when (text = @ss.scan(/ONLINE\b/i))
+         action { [:ONLINE, text] }
+
+      when (text = @ss.scan(/ON\b/i))
+         action { [:ON, text] }
+
+      when (text = @ss.scan(/OFFLINE\b/i))
+         action { [:OFFLINE, text] }
+
+      when (text = @ss.scan(/NUMERIC\b/i))
+         action { [:NUMERIC, text] }
+
+      when (text = @ss.scan(/NULL\b/i))
+         action { [:NULL, text] }
+
+      when (text = @ss.scan(/NOT\b/i))
+         action { [:NOT, text] }
+
+      when (text = @ss.scan(/NODEGROUP\b/i))
+         action { [:NODEGROUP, text] }
+
+      when (text = @ss.scan(/NO\b/i))
+         action { [:NO, text] }
+
+      when (text = @ss.scan(/MODIFY\b/i))
+         action { [:MODIFY, text] }
+
+      when (text = @ss.scan(/MIN_ROWS\b/i))
+         action { [:MIN_ROWS, text] }
+
+      when (text = @ss.scan(/MERGE\b/i))
+         action { [:MERGE, text] }
+
+      when (text = @ss.scan(/MEMORY\b/i))
+         action { [:MEMORY, text] }
+
+      when (text = @ss.scan(/MEDIUMTEXT\b/i))
+         action { [:MEDIUMTEXT, text] }
+
+      when (text = @ss.scan(/MEDIUMINT\b/i))
+         action { [:MEDIUMINT, text] }
+
+      when (text = @ss.scan(/MEDIUMBLOB\b/i))
+         action { [:MEDIUMBLOB, text] }
+
+      when (text = @ss.scan(/MAX_ROWS\b/i))
+         action { [:MAX_ROWS, text] }
+
+      when (text = @ss.scan(/MAXVALUE\b/i))
+         action { [:MAXVALUE, text] }
+
+      when (text = @ss.scan(/MATCH\b/i))
+         action { [:MATCH, text] }
+
+      when (text = @ss.scan(/LONGTEXT\b/i))
+         action { [:LONGTEXT, text] }
+
+      when (text = @ss.scan(/LONGBLOB\b/i))
+         action { [:LONGBLOB, text] }
+
+      when (text = @ss.scan(/LOCAL\b/i))
+         action { [:LOCAL, text] }
+
+      when (text = @ss.scan(/LIKE\b/i))
+         action { [:LIKE, text] }
+
+      when (text = @ss.scan(/LESS\b/i))
+         action { [:LESS, text] }
+
+      when (text = @ss.scan(/LATIN1\b/i))
+         action { [:LATIN1, text] }
+
+      when (text = @ss.scan(/LAST\b/i))
+         action { [:LAST, text] }
+
+      when (text = @ss.scan(/KEY_BLOCK_SIZE\b/i))
+         action { [:KEY_BLOCK_SIZE, text] }
+
+      when (text = @ss.scan(/KEYS\b/i))
+         action { [:KEYS, text] }
+
+      when (text = @ss.scan(/KEY\b/i))
+         action { [:KEY, text] }
+
+      when (text = @ss.scan(/INVOKER\b/i))
+         action { [:INVOKER, text] }
+
+      when (text = @ss.scan(/INTO\b/i))
+         action { [:INTO, text] }
+
+      when (text = @ss.scan(/INTEGER\b/i))
+         action { [:INTEGER, text] }
+
+      when (text = @ss.scan(/INT\b/i))
+         action { [:INT, text] }
+
+      when (text = @ss.scan(/INSERT_METHOD\b/i))
+         action { [:INSERT_METHOD, text] }
+
+      when (text = @ss.scan(/INNODB\b/i))
+         action { [:INNODB, text] }
+
+      when (text = @ss.scan(/INDEX\b/i))
+         action { [:INDEX, text] }
+
+      when (text = @ss.scan(/IN\b/i))
+         action { [:IN, text] }
+
+      when (text = @ss.scan(/IMPORT\b/i))
+         action { [:IMPORT, text] }
+
+      when (text = @ss.scan(/IGNORE\b/i))
+         action { [:IGNORE, text] }
+
+      when (text = @ss.scan(/IF\b/i))
+         action { [:IF, text] }
+
+      when (text = @ss.scan(/HASH\b/i))
+         action { [:HASH, text] }
+
+      when (text = @ss.scan(/FULLTEXT\b/i))
+         action { [:FULLTEXT, text] }
+
+      when (text = @ss.scan(/FULL\b/i))
+         action { [:FULL, text] }
+
+      when (text = @ss.scan(/FOREIGN\b/i))
+         action { [:FOREIGN, text] }
+
+      when (text = @ss.scan(/FLOAT\b/i))
+         action { [:FLOAT, text] }
+
+      when (text = @ss.scan(/FIXED\b/i))
+         action { [:FIXED, text] }
+
+      when (text = @ss.scan(/FIRST\b/i))
+         action { [:FIRST, text] }
+
+      when (text = @ss.scan(/EXISTS\b/i))
+         action { [:EXISTS, text] }
+
+      when (text = @ss.scan(/ENUM\b/i))
+         action { [:ENUM, text] }
+
+      when (text = @ss.scan(/ENGINE\b/i))
+         action { [:ENGINE, text] }
+
+      when (text = @ss.scan(/ENABLE\b/i))
+         action { [:ENABLE, text] }
+
+      when (text = @ss.scan(/DYNAMIC\b/i))
+         action { [:DYNAMIC, text] }
+
+      when (text = @ss.scan(/DROP\b/i))
+         action { [:DROP, text] }
+
+      when (text = @ss.scan(/DOUBLE\b/i))
+         action { [:DOUBLE, text] }
+
+      when (text = @ss.scan(/DISK\b/i))
+         action { [:DISK, text] }
+
+      when (text = @ss.scan(/DISCARD\b/i))
+         action { [:DISCARD, text] }
+
+      when (text = @ss.scan(/DISABLE\b/i))
+         action { [:DISABLE, text] }
+
+      when (text = @ss.scan(/DIRECTORY\b/i))
+         action { [:DIRECTORY, text] }
+
+      when (text = @ss.scan(/DESC\b/i))
+         action { [:DESC, text] }
+
+      when (text = @ss.scan(/DELETE\b/i))
+         action { [:DELETE, text] }
+
+      when (text = @ss.scan(/DELAY_KEY_WRITE\b/i))
+         action { [:DELAY_KEY_WRITE, text] }
+
+      when (text = @ss.scan(/DEFINER\b/i))
+         action { [:DEFINER, text] }
+
+      when (text = @ss.scan(/DEFAULT\b/i))
+         action { [:DEFAULT, text] }
+
+      when (text = @ss.scan(/DECIMAL\b/i))
+         action { [:DECIMAL, text] }
+
+      when (text = @ss.scan(/DATETIME\b/i))
+         action { [:DATETIME, text] }
+
+      when (text = @ss.scan(/DATE\b/i))
+         action { [:DATE, text] }
+
+      when (text = @ss.scan(/DATA\b/i))
+         action { [:DATA, text] }
+
+      when (text = @ss.scan(/CURRENT_USER\b/i))
+         action { [:CURRENT_USER, text] }
+
+      when (text = @ss.scan(/CURRENT_TIMESTAMP\b/i))
+         action { [:CURRENT_TIMESTAMP, text] }
+
+      when (text = @ss.scan(/CREATE\b/i))
+         action { [:CREATE, text] }
+
+      when (text = @ss.scan(/CONVERT\b/i))
+         action { [:CONVERT, text] }
+
+      when (text = @ss.scan(/CONSTRAINT\b/i))
+         action { [:CONSTRAINT, text] }
+
+      when (text = @ss.scan(/CONNECTION\b/i))
+         action { [:CONNECTION, text] }
+
+      when (text = @ss.scan(/COMPRESSED\b/i))
+         action { [:COMPRESSED, text] }
+
+      when (text = @ss.scan(/COMPACT\b/i))
+         action { [:COMPACT, text] }
+
+      when (text = @ss.scan(/COMMENT\b/i))
+         action { [:COMMENT, text] }
+
+      when (text = @ss.scan(/COLUMN_FORMAT\b/i))
+         action { [:COLUMN_FORMAT, text] }
+
+      when (text = @ss.scan(/COLUMN\b/i))
+         action { [:COLUMN, text] }
+
+      when (text = @ss.scan(/COLLATE\b/i))
+         action { [:COLLATE, text] }
+
+      when (text = @ss.scan(/COALESCE\b/i))
+         action { [:COALESCE, text] }
+
+      when (text = @ss.scan(/CHECKSUM\b/i))
+         action { [:CHECKSUM, text] }
+
+      when (text = @ss.scan(/CHECK\b/i))
+         action { [:CHECK, text] }
+
+      when (text = @ss.scan(/CHAR\b/i))
+         action { [:CHAR, text] }
+
+      when (text = @ss.scan(/CHANGE\b/i))
+         action { [:CHANGE, text] }
+
+      when (text = @ss.scan(/CASCADED\b/i))
+         action { [:CASCADED, text] }
+
+      when (text = @ss.scan(/CASCADE\b/i))
+         action { [:CASCADE, text] }
+
+      when (text = @ss.scan(/BY\b/i))
+         action { [:BY, text] }
+
+      when (text = @ss.scan(/BTREE\b/i))
+         action { [:BTREE, text] }
+
+      when (text = @ss.scan(/BLOB\b/i))
+         action { [:BLOB, text] }
+
+      when (text = @ss.scan(/BIT\b/i))
+         action { [:BIT, text] }
+
+      when (text = @ss.scan(/BINARY\b/i))
+         action { [:BINARY, text] }
+
+      when (text = @ss.scan(/BIGINT\b/i))
+         action { [:BIGINT, text] }
+
+      when (text = @ss.scan(/AVG_ROW_LENGTH\b/i))
+         action { [:AVG_ROW_LENGTH, text] }
+
+      when (text = @ss.scan(/AUTO_INCREMENT\b/i))
+         action { [:AUTO_INCREMENT, text] }
+
+      when (text = @ss.scan(/ASC\b/i))
+         action { [:ASC, text] }
+
+      when (text = @ss.scan(/AS\b/i))
+         action { [:AS, text] }
+
+      when (text = @ss.scan(/ANALYZE\b/i))
+         action { [:ANALYZE, text] }
+
+      when (text = @ss.scan(/ALTER\b/i))
+         action { [:ALTER, text] }
+
+      when (text = @ss.scan(/ALL\b/i))
+         action { [:ALL, text] }
+
+      when (text = @ss.scan(/ALGORITHM\b/i))
+         action { [:ALGORITHM, text] }
+
+      when (text = @ss.scan(/AFTER\b/i))
+         action { [:AFTER, text] }
+
+      when (text = @ss.scan(/ADD\b/i))
+         action { [:ADD, text] }
+
+      when (text = @ss.scan(/ACTION\b/i))
+         action { [:ACTION, text] }
+
+      when (text = @ss.scan(/,/i))
+         action { [:S_COMMA        , text] }
+
+      when (text = @ss.scan(/@/i))
+         action { [:S_AT           , text] }
+
+      when (text = @ss.scan(/0+\b/i))
+         action { [:S_ZERO         , text] } # this must come before S_NAT
+
+      when (text = @ss.scan(/1\b/i))
+         action { [:S_ONE          , text] } # this must come before S_NAT
+
+      when (text = @ss.scan(/\d+/i))
+         action { [:S_NAT          , text] } # definitely not 0, 1
+
+      when (text = @ss.scan(/-?\d+\.\d+/i))
+         action { [:S_FLOAT        , text] } # this must come before S_DOT
+
+      when (text = @ss.scan(/[\$a-zA-Z0-9_]+/i))
+         action { [:S_IDENT_NORMAL , text] }
+
+      when (text = @ss.scan(/=/i))
+         action { [:S_EQUAL        , text] }
+
+      when (text = @ss.scan(/\(/i))
+         action { [:S_LEFT_PAREN   , text] }
+
+      when (text = @ss.scan(/\)/i))
+         action { [:S_RIGHT_PAREN  , text] }
+
+      when (text = @ss.scan(/-/i))
+         action { [:S_MINUS        , text] }
+
+      when (text = @ss.scan(/\./i))
+         action { [:S_DOT          , text] }
+
+      when (text = @ss.scan(/[ \t\n]+/i))
+         action { [:S_SPACE, ' '] } # set to one space
+
+      else
+        text = @ss.string[@ss.pos .. -1]
+        raise  ScanError, "can not match: '" + text + "'"
+      end  # if
+
+    when :A_REM_MULTI
+      case
+      when (text = @ss.scan(/\*\/ */i))
+         action { @state = :A_NIL; [:S_REM_OUT, ' */ '] }
+
+      when (text = @ss.scan(/(.+)(?=\*\/ *)/i))
+         action { [:S_COMMENT, text] }
+
+      else
+        text = @ss.string[@ss.pos .. -1]
+        raise  ScanError, "can not match: '" + text + "'"
+      end  # if
+
+    when :A_REM_INLINE
+      case
+      when (text = @ss.scan(/\n/i))
+         action { @state = :A_NIL; [:S_REM_OUT, text] }
+
+      when (text = @ss.scan(/.*(?=$)/i))
+         action { [:S_COMMENT, text] }
+
+      else
+        text = @ss.string[@ss.pos .. -1]
+        raise  ScanError, "can not match: '" + text + "'"
+      end  # if
+
+    when :A_BACKTICK
+      case
+      when (text = @ss.scan(/``/i))
+         action { [:S_IDENT_IN_BACKTICK, text] }
+
+      when (text = @ss.scan(/`/i))
+         action { @state = :A_NIL; [:S_BACKTICK_OUT, text] }
+
+      when (text = @ss.scan(/[^`]+/i))
+         action { [:S_IDENT_IN_BACKTICK, text] }
+
+      else
+        text = @ss.string[@ss.pos .. -1]
+        raise  ScanError, "can not match: '" + text + "'"
+      end  # if
+
+    when :A_DOUBLEQUOTE
+      case
+      when (text = @ss.scan(/""/i))
+         action { [:S_STRING_IN_QUOTE, text] }
+
+      when (text = @ss.scan(/"/i))
+         action { @state = :A_NIL; [:S_DOUBLEQUOTE_OUT, text] }
+
+      when (text = @ss.scan(/[^"]*/i))
+         action { [:S_STRING_IN_QUOTE, text] }
+
+      else
+        text = @ss.string[@ss.pos .. -1]
+        raise  ScanError, "can not match: '" + text + "'"
+      end  # if
+
+    when :A_SINGLEQUOTE
+      case
+      when (text = @ss.scan(/''/i))
+         action { [:S_STRING_IN_QUOTE, text] }
+
+      when (text = @ss.scan(/'/i))
+         action { @state = :A_NIL; [:S_SINGLEQUOTE_OUT, text] }
+
+      when (text = @ss.scan(/[^']*/i))
+         action { [:S_STRING_IN_QUOTE, text] }
+
+      else
+        text = @ss.string[@ss.pos .. -1]
+        raise  ScanError, "can not match: '" + text + "'"
+      end  # if
+
+    else
+      raise  ScanError, "undefined state: '" + state.to_s + "'"
+    end  # case state
+    token
+  end  # def _next_token
+
+  def tokenize(code)
+    scan_setup(code)
+    tokens = []
+    while token = next_token
+      tokens << token
+    end
+    tokens
+  end
+end # class