[BUGFIX] Fix failing specs on item repeats

Fix #7

* Use a C or Java extension
* Kill a lot of yucky, poorly performing and error prone code

.gitignore

@@ -1,6 +1,7 @@
 .DS_Store
 *.gem
-.bundle
+*.bundle
+*.jar
 Gemfile.lock
 pkg/*
 spec/reports

.travis.yml

@@ -7,5 +7,8 @@ rvm:
   - rbx-19mode
   - ruby-head
 
+before_install:
+  - sudo apt-get install libpcre3 libpcre3-dev
+
 notifications:
   irc: "irc.freenode.org#adhearsion-dev"

Guardfile

@@ -1,3 +1,7 @@
+guard 'rake', :task => 'compile' do
+  watch(%r{^ext/(.+)\.c$})
+end
+
 guard 'rspec', :cli => '--format documentation' do
   watch(%r{^spec/.+_spec\.rb$})
   watch(%r{^lib/(.+)\.rb$})     { |m| "spec/#{m[1]}_spec.rb" }

Rakefile

@@ -15,8 +15,20 @@ RSpec::Core::RakeTask.new(:rcov) do |spec|
   spec.rspec_opts = '--color'
 end
 
-task :default => :spec
-task :ci => ['ci:setup:rspec', :spec]
+task :default => [:compile, :spec]
+task :ci => ['ci:setup:rspec', :compile, :spec]
 
 require 'yard'
 YARD::Rake::YardocTask.new
+
+if RUBY_PLATFORM =~ /java/
+  require 'rake/javaextensiontask'
+  Rake::JavaExtensionTask.new 'ruby_speech' do |ext|
+    ext.lib_dir = 'lib/ruby_speech'
+  end
+else
+  require 'rake/extensiontask'
+  Rake::ExtensionTask.new 'ruby_speech' do |ext|
+    ext.lib_dir = 'lib/ruby_speech'
+  end
+end

ext/ruby_speech/RubySpeechGRXMLMatcher.java

@@ -0,0 +1,42 @@
+package com.benlangfeld.ruby_speech;
+
+import org.jruby.Ruby;
+import org.jruby.RubyClass;
+import org.jruby.RubyModule;
+import org.jruby.RubyObject;
+import org.jruby.anno.JRubyClass;
+import org.jruby.anno.JRubyMethod;
+import org.jruby.runtime.ObjectAllocator;
+import org.jruby.runtime.ThreadContext;
+import org.jruby.runtime.Visibility;
+import org.jruby.runtime.builtin.IRubyObject;
+import org.jruby.javasupport.util.RuntimeHelpers;
+
+import java.util.regex.*;
+
+@JRubyClass(name="RubySpeech::GRXML::Matcher")
+public class RubySpeechGRXMLMatcher extends RubyObject {
+
+  public RubySpeechGRXMLMatcher(final Ruby runtime, RubyClass rubyClass) {
+    super(runtime, rubyClass);
+  }
+
+  @JRubyMethod(visibility=Visibility.PRIVATE)
+  public IRubyObject check_potential_match(ThreadContext context, IRubyObject buffer)
+  {
+    Ruby runtime = context.getRuntime();
+
+    IRubyObject regex = getInstanceVariable("@regex");
+
+    Pattern p = Pattern.compile(regex.toString());
+    Matcher m = p.matcher(buffer.toString());
+
+    if (m.matches()) {
+    } else if (m.hitEnd()) {
+      RubyModule potential_match = runtime.getClassFromPath("RubySpeech::GRXML::PotentialMatch");
+      return RuntimeHelpers.invoke(context, potential_match, "new");
+    }
+    return runtime.getNil();
+  }
+
+}

ext/ruby_speech/RubySpeechService.java

@@ -0,0 +1,23 @@
+package com.benlangfeld.ruby_speech;
+
+import org.jruby.Ruby;
+import org.jruby.RubyClass;
+import org.jruby.RubyModule;
+import org.jruby.RubyObject;
+import org.jruby.runtime.ObjectAllocator;
+import org.jruby.runtime.builtin.IRubyObject;
+import org.jruby.runtime.load.BasicLibraryService;
+
+public class RubySpeechService implements BasicLibraryService {
+  public boolean basicLoad(Ruby ruby) {
+    RubyModule ruby_speech = ruby.defineModule("RubySpeech");
+    RubyModule grxml = ruby_speech.defineModuleUnder("GRXML");
+    RubyClass matcher = grxml.defineClassUnder("Matcher", ruby.getObject(), new ObjectAllocator() {
+      public IRubyObject allocate(Ruby runtime, RubyClass rubyClass) {
+        return new RubySpeechGRXMLMatcher(runtime, rubyClass);
+      }
+    });
+    matcher.defineAnnotatedMethods(RubySpeechGRXMLMatcher.class);
+    return true;
+  }
+}

ext/ruby_speech/extconf.rb

@@ -0,0 +1,7 @@
+require 'mkmf'
+
+$LIBS << " -lpcre"
+
+abort "-----\n#{lib} is missing.\n-----" unless find_header('pcre.h')
+
+create_makefile 'ruby_speech'

ext/ruby_speech/ruby_speech.c

@@ -0,0 +1,41 @@
+#include "ruby.h"
+#include "pcre.h"
+#include <stdio.h>
+
+static VALUE method_check_potential_match(VALUE self, VALUE buffer)
+{
+  int erroffset = 0;
+  const char *errptr = "";
+  int options = 0;
+  VALUE regex_string = rb_funcall(rb_iv_get(self, "@regex"), rb_intern("to_s"), 0);
+  const char *regex = StringValueCStr(regex_string);
+
+  pcre *compiled_regex = pcre_compile(regex, options, &errptr, &erroffset, NULL);
+
+  int result = 0;
+  int ovector[30];
+  int workspace[1024];
+  char *input = StringValueCStr(buffer);
+  result = pcre_dfa_exec(compiled_regex, NULL, input, strlen(input), 0, PCRE_PARTIAL,
+    ovector, sizeof(ovector) / sizeof(ovector[0]),
+    workspace, sizeof(workspace) / sizeof(workspace[0]));
+  pcre_free(compiled_regex);
+
+  if (result == PCRE_ERROR_PARTIAL) {
+    VALUE RubySpeech      = rb_const_get(rb_cObject, rb_intern("RubySpeech"));
+    VALUE GRXML           = rb_const_get(RubySpeech, rb_intern("GRXML"));
+    VALUE PotentialMatch  = rb_const_get(GRXML, rb_intern("PotentialMatch"));
+
+    return rb_class_new_instance(0, NULL, PotentialMatch);
+  }
+  return Qnil;
+}
+
+void Init_ruby_speech()
+{
+  VALUE RubySpeech  = rb_define_module("RubySpeech");
+  VALUE GRXML       = rb_define_module_under(RubySpeech, "GRXML");
+  VALUE Matcher     = rb_define_class_under(GRXML, "Matcher", rb_cObject);
+
+  rb_define_method(Matcher, "check_potential_match", method_check_potential_match, 1);
+}

lib/ruby_speech/grxml/element.rb

@@ -24,23 +24,6 @@ def self.module
       def regexp_content # :nodoc:
         children.map(&:regexp_content).join
       end
-
-      def potential_match?(other)
-        false
-      end
-
-      def max_input_length
-        0
-      end
-
-      def longest_potential_match(input)
-        input.dup.tap do |longest_input|
-          begin
-            return longest_input if potential_match? longest_input
-            longest_input.chop!
-          end until longest_input.length.zero?
-        end
-      end
     end # Element
   end # GRXML
 end # RubySpeech

lib/ruby_speech/grxml/item.rb

@@ -139,27 +139,6 @@ def regexp_content # :nodoc:
           super
         end
       end
-
-      def potential_match?(other)
-        tokens = children
-        return false if other.length > max_input_length
-        other.chars.each_with_index do |digit, index|
-          index -= tokens.size until index < tokens.size if repeat
-          return false unless tokens[index].potential_match?(digit)
-        end
-        true
-      end
-
-      def max_input_length # :nodoc:
-        case repeat
-        when Range
-          children.size * repeat.max
-        when Integer
-          children.size * repeat
-        else
-          children.size
-        end
-      end
     end # Item
   end # GRXML
 end # RubySpeech

lib/ruby_speech/grxml/matcher.rb

@@ -1,3 +1,10 @@
+require 'ruby_speech/ruby_speech'
+
+if RUBY_PLATFORM =~ /java/
+  require 'jruby'
+  com.benlangfeld.ruby_speech.RubySpeechService.new.basicLoad(JRuby.runtime)
+end
+
 module RubySpeech
   module GRXML
     class Matcher
@@ -104,17 +111,6 @@ def check_full_match(buffer)
                   :interpretation => interpret_utterance(buffer)
       end
 
-      def check_potential_match(buffer)
-        grammar.root_rule.children.each do |token|
-          p "Checking buffer #{buffer} against token #{token} which has a longest potential match #{token.longest_potential_match(buffer)}"
-          break if buffer.length.zero?
-          longest_potential_match = token.longest_potential_match buffer
-          return if longest_potential_match.length.zero?
-          buffer.gsub! /^#{Regexp.escape longest_potential_match}/, ''
-        end
-        buffer.length.zero? ? PotentialMatch.new : nil
-      end
-
       def regexp_content
         grammar.root_rule.children.map &:regexp_content
       end

lib/ruby_speech/grxml/one_of.rb

@@ -26,10 +26,6 @@ def <<(arg)
       def regexp_content # :nodoc:
         "(#{children.map(&:regexp_content).join '|'})"
       end
-
-      def potential_match?(input)
-        children.any? { |c| c.potential_match? input }
-      end
     end # OneOf
   end # GRXML
 end # RubySpeech

lib/ruby_speech/grxml/token.rb

@@ -26,10 +26,6 @@ def normalize_whitespace
       def regexp_content # :nodoc:
         Regexp.escape content
       end
-
-      def potential_match?(other)
-        other == content
-      end
     end # Token
   end # GRXML
 end # RubySpeech

ruby_speech.gemspec

@@ -18,6 +18,13 @@ Gem::Specification.new do |s|
   s.executables   = `git ls-files -- bin/*`.split("\n").map{ |f| File.basename(f) }
   s.require_paths = ["lib"]
 
+  if RUBY_PLATFORM =~ /java/
+    s.platform = "java"
+    s.files << "lib/ruby_speech/ruby_speech.jar"
+  else
+    s.extensions = ['ext/ruby_speech/extconf.rb']
+  end
+
   s.add_runtime_dependency %q<niceogiri>, ["~> 1.1", ">= 1.1.1"]
   s.add_runtime_dependency %q<nokogiri>, ["~> 1.5", ">= 1.5.6"]
   s.add_runtime_dependency %q<activesupport>, [">= 3.0.7"]
@@ -32,4 +39,6 @@ Gem::Specification.new do |s|
   s.add_development_dependency %q<guard>, [">= 0.9.0"]
   s.add_development_dependency %q<guard-rspec>, [">= 0"]
   s.add_development_dependency %q<ruby_gntp>, [">= 0"]
+  s.add_development_dependency %q<guard-rake>, [">= 0"]
+  s.add_development_dependency %q<rake-compiler>, [">= 0"]
 end