PEG.js grammar: More JavaScript-like initializer/rule separation

Initializer and rules are now separated in a similar way as JavaScript
statements -- either by a semicolon or a line terminator, possibly with
whitespace and comments mixed in.

One consequence is that the grammars like this are now illegal:

  foo = "a" bar = "b"

A semicolon needs to be inserted between the rules:

  foo = "a";bar = "b"

I consider this a good change as the now-illegal syntax was somewhat
confusing.

spec/parser.spec.js

@@ -181,7 +181,7 @@ describe("PEG.js grammar parser", function() {
       initializer: null,
       rules:       [ruleA]
     });
-    expect('{ code } a = "abcd"').toParseAs({
+    expect('{ code }; a = "abcd"').toParseAs({
       type:        "grammar",
       initializer: { type: "initializer", code: " code " },
       rules:       [ruleA]
@@ -200,8 +200,7 @@ describe("PEG.js grammar parser", function() {
           code: " code "
         });
 
-    expect('{ code } start = "abcd"' ).toParseAs(grammar);
-    expect('{ code }\n; start = "abcd"').toParseAs(grammar);
+    expect('{ code }; start = "abcd"' ).toParseAs(grammar);
   });
 
   /* Canonical Rule is "a: \"abcd\"". */
@@ -382,6 +381,16 @@ describe("PEG.js grammar parser", function() {
     expect('start =/**/*/"abcd"').toFailToParse();
   });
 
+  // Canonical MultiLineCommentNoLineTerminator is "/* comment */".
+  it("parses MultiLineCommentNoLineTerminator", function() {
+    expect('start = "abcd"/**/\r\n'   ).toParseAs(trivialGrammar);
+    expect('start = "abcd"/*a*/\r\n'  ).toParseAs(trivialGrammar);
+    expect('start = "abcd"/*aaa*/\r\n').toParseAs(trivialGrammar);
+
+    expect('{ code }/**/*/\r\nstart = "abcd"').toFailToParse();
+    expect('{ code }/*\n*/\r\nstart = "abcd"').toFailToParse();
+  });
+
   /* Canonical SingleLineComment is "// comment". */
   it("parses SingleLineComment", function() {
     expect('start =//\n"abcd"'   ).toParseAs(trivialGrammar);
@@ -593,4 +602,25 @@ describe("PEG.js grammar parser", function() {
     expect('start =/* comment */"abcd"').toParseAs(trivialGrammar);
     expect('start =   "abcd"'          ).toParseAs(trivialGrammar);
   });
+
+  /* Canonical _ is " ". */
+  it("parses _", function() {
+    expect('start = "abcd"\r\n'             ).toParseAs(trivialGrammar);
+    expect('start = "abcd" \r\n'            ).toParseAs(trivialGrammar);
+    expect('start = "abcd"/* comment */\r\n').toParseAs(trivialGrammar);
+    expect('start = "abcd"   \r\n'          ).toParseAs(trivialGrammar);
+  });
+
+  /* Canonical EOS is ";". */
+  it("parses EOS", function() {
+    expect('start = "abcd"\n;'            ).toParseAs(trivialGrammar);
+    expect('start = "abcd" \r\n'          ).toParseAs(trivialGrammar);
+    expect('start = "abcd" // comment\r\n').toParseAs(trivialGrammar);
+    expect('start = "abcd"\n'             ).toParseAs(trivialGrammar);
+  });
+
+  /* Canonical EOF is the end of input. */
+  it("parses EOS", function() {
+    expect('start = "abcd"\n').toParseAs(trivialGrammar);
+  });
 });

src/parser.pegjs

@@ -41,7 +41,7 @@ Grammar
     }
 
 Initializer
-  = code:CodeBlock (__ ";")? {
+  = code:CodeBlock EOS {
       return {
         type: "initializer",
         code: code
@@ -52,7 +52,7 @@ Rule
   = name:IdentifierName __
     displayName:(StringLiteral __)?
     "=" __
-    expression:Expression (__ ";")? {
+    expression:Expression EOS {
       return {
         type:        "rule",
         name:        name,
@@ -200,6 +200,9 @@ Comment "comment"
 MultiLineComment
   = "/*" (!"*/" SourceCharacter)* "*/"
 
+MultiLineCommentNoLineTerminator
+  = "/*" (!("*/" / LineTerminator) SourceCharacter)* "*/"
+
 SingleLineComment
   = "//" (!LineTerminator SourceCharacter)*
 
@@ -504,3 +507,16 @@ WithToken       = "with"       !IdentifierPart
 
 __
   = (WhiteSpace / LineTerminatorSequence / Comment)*
+
+_
+  = (WhiteSpace / MultiLineCommentNoLineTerminator)*
+
+/* Automatic Semicolon Insertion */
+
+EOS
+  = __ ";"
+  / _ SingleLineComment? LineTerminatorSequence
+  / __ EOF
+
+EOF
+  = !.