Make lexer abstract

modules/lang-painless/ant.xml

@@ -1,7 +1,7 @@
 <?xml version="1.0"?>
 <project name="ant-stuff">
 
-<!-- 
+<!--
  grammar regeneration logic
  we do this with ant for several reasons:
  * remove generated tabs for forbidden-apis
@@ -9,7 +9,7 @@
  * fix CRLF line endings for windows consistency
  * ability to make classes package-private
  * keeping in source code control is easier on IDEs
- * regeneration should be rare, no reason to be religious about generated files 
+ * regeneration should be rare, no reason to be religious about generated files
  * all logic already written and battle tested in lucene build
 -->
   <target name="regenerate" description="Regenerate antlr lexer and parser" depends="run-antlr"/>
@@ -136,6 +136,10 @@
       <replaceregexp match="public ((interface|class) \Q@{grammar}\E\w+)" replace="\1" encoding="UTF-8">
         <fileset refid="grammar.fileset"/>
       </replaceregexp>
+      <!-- make the lexer abstract -->
+      <replaceregexp match="(class \Q@{grammar}\ELexer)" replace="abstract \1" encoding="UTF-8">
+        <fileset refid="grammar.fileset"/>
+      </replaceregexp>
       <!-- nuke timestamps/filenames in generated files -->
       <replaceregexp match="\Q// Generated from \E.*" replace="\/\/ ANTLR GENERATED CODE: DO NOT EDIT" encoding="UTF-8">
         <fileset refid="grammar.fileset"/>

modules/lang-painless/src/main/antlr/PainlessLexer.g4

@@ -23,12 +23,19 @@ lexer grammar PainlessLexer;
 }
 
 @members{
-  protected boolean isSimpleType(String name) {
-    throw new UnsupportedOperationException("Must be implemented in a subclass");
-  }
-  protected boolean slashIsRegex() {
-    throw new UnsupportedOperationException("Must be implemented in a subclass");
-  }
+/**
+ * Check against the current whitelist to determine whether a token is a type
+ * or not. Called by the {@code TYPE} token defined in {@code PainlessLexer.g4}.
+ * See also
+ * <a href="https://en.wikipedia.org/wiki/The_lexer_hack">The lexer hack</a>.
+ */
+protected abstract boolean isSimpleType(String name);
+
+/**
+ * Is the preceding {@code /} a the beginning of a regex (true) or a division
+ * (false).
+ */
+protected abstract boolean slashIsRegex();
 }
 
 WS: [ \t\n\r]+ -> skip;

modules/lang-painless/src/main/java/org/elasticsearch/painless/antlr/PainlessLexer.java

@@ -12,7 +12,7 @@ import org.antlr.v4.runtime.dfa.DFA;
 import org.antlr.v4.runtime.misc.*;
 
 @SuppressWarnings({"all", "warnings", "unchecked", "unused", "cast"})
-class PainlessLexer extends Lexer {
+abstract class PainlessLexer extends Lexer {
   static { RuntimeMetaData.checkVersion("4.5.1", RuntimeMetaData.VERSION); }
 
   protected static final DFA[] _decisionToDFA;
@@ -105,12 +105,19 @@ class PainlessLexer extends Lexer {
   }
 
 
-    protected boolean isSimpleType(String name) {
-      throw new UnsupportedOperationException("Must be implemented in a subclass");
-    }
-    protected boolean slashIsRegex() {
-      throw new UnsupportedOperationException("Must be implemented in a subclass");
-    }
+  /**
+   * Check against the current whitelist to determine whether a token is a type
+   * or not. Called by the {@code TYPE} token defined in {@code PainlessLexer.g4}.
+   * See also
+   * <a href="https://en.wikipedia.org/wiki/The_lexer_hack">The lexer hack</a>.
+   */
+  protected abstract boolean isSimpleType(String name);
+
+  /**
+   * Is the preceding {@code /} a the beginning of a regex (true) or a division
+   * (false).
+   */
+  protected abstract boolean slashIsRegex();
 
 
   public PainlessLexer(CharStream input) {