Simplify scanner and generate better error message for legacy octals in templates

LOG=N
BUG=
R=arv@chromium.org, dslomov@chromium.org

Review URL: https://codereview.chromium.org/812163003

Cr-Commit-Position: refs/heads/master@{#25895}

src/messages.js

@@ -154,6 +154,7 @@ var kMessages = {
   too_many_variables:            ["Too many variables declared (only 4194303 allowed)"],
   strict_param_dupe:             ["Strict mode function may not have duplicate parameter names"],
   strict_octal_literal:          ["Octal literals are not allowed in strict mode."],
+  template_octal_literal:        ["Octal literals are not allowed in template strings."],
   strict_duplicate_property:     ["Duplicate data property in object literal not allowed in strict mode"],
   accessor_data_property:        ["Object literal may not have data and accessor property with the same name"],
   accessor_get_set:              ["Object literal may not have multiple get/set accessors with the same name"],

src/parser.cc

@@ -936,7 +936,7 @@ FunctionLiteral* Parser::DoParseProgram(CompilationInfo* info, Scope** scope,
                         &ok);
 
     if (ok && strict_mode() == STRICT) {
-      CheckOctalLiteral(beg_pos, scanner()->location().end_pos, &ok);
+      CheckStrictOctalLiteral(beg_pos, scanner()->location().end_pos, &ok);
     }
 
     if (ok && allow_harmony_scoping() && strict_mode() == STRICT) {
@@ -3803,9 +3803,8 @@ FunctionLiteral* Parser::ParseFunctionLiteral(
                                            CHECK_OK);
     }
     if (strict_mode() == STRICT) {
-      CheckOctalLiteral(scope->start_position(),
+      CheckStrictOctalLiteral(scope->start_position(), scope->end_position(),
-                        scope->end_position(),
+                              CHECK_OK);
-                        CHECK_OK);
     }
     if (allow_harmony_scoping() && strict_mode() == STRICT) {
       CheckConflictingVarDeclarations(scope, CHECK_OK);

src/preparser.cc

@@ -125,7 +125,7 @@ PreParser::PreParseResult PreParser::PreParseLazyFunction(
     DCHECK_EQ(Token::RBRACE, scanner()->peek());
     if (scope_->strict_mode() == STRICT) {
       int end_pos = scanner()->location().end_pos;
-      CheckOctalLiteral(start_position, end_pos, &ok);
+      CheckStrictOctalLiteral(start_position, end_pos, &ok);
     }
   }
   return kPreParseSuccess;
@@ -937,7 +937,7 @@ PreParser::Expression PreParser::ParseFunctionLiteral(
     }
 
     int end_position = scanner()->location().end_pos;
-    CheckOctalLiteral(start_position, end_position, CHECK_OK);
+    CheckStrictOctalLiteral(start_position, end_position, CHECK_OK);
   }
 
   return Expression::Default();

src/preparser.h

@@ -385,17 +385,26 @@ class ParserBase : public Traits {
   }
 
   // Checks whether an octal literal was last seen between beg_pos and end_pos.
-  // If so, reports an error. Only called for strict mode.
+  // If so, reports an error. Only called for strict mode and template strings.
-  void CheckOctalLiteral(int beg_pos, int end_pos, bool* ok) {
+  void CheckOctalLiteral(int beg_pos, int end_pos, const char* error,
+                         bool* ok) {
     Scanner::Location octal = scanner()->octal_position();
     if (octal.IsValid() && beg_pos <= octal.beg_pos &&
         octal.end_pos <= end_pos) {
-      ReportMessageAt(octal, "strict_octal_literal");
+      ReportMessageAt(octal, error);
       scanner()->clear_octal_position();
       *ok = false;
     }
   }
 
+  inline void CheckStrictOctalLiteral(int beg_pos, int end_pos, bool* ok) {
+    CheckOctalLiteral(beg_pos, end_pos, "strict_octal_literal", ok);
+  }
+
+  inline void CheckTemplateOctalLiteral(int beg_pos, int end_pos, bool* ok) {
+    CheckOctalLiteral(beg_pos, end_pos, "template_octal_literal", ok);
+  }
+
   // Validates strict mode for function parameter lists. This has to be
   // done after parsing the function, since the function can declare
   // itself strict.
@@ -1489,7 +1498,8 @@ class PreParser : public ParserBase<PreParserTraits> {
     if (!ok) {
       ReportUnexpectedToken(scanner()->current_token());
     } else if (scope_->strict_mode() == STRICT) {
-      CheckOctalLiteral(start_position, scanner()->location().end_pos, &ok);
+      CheckStrictOctalLiteral(start_position, scanner()->location().end_pos,
+                              &ok);
     }
     if (materialized_literals) {
       *materialized_literals = function_state_->materialized_literal_count();
@@ -2802,7 +2812,8 @@ typename ParserBase<Traits>::ExpressionT ParserBase<
 
     // Validate strict mode.
     if (strict_mode() == STRICT) {
-      CheckOctalLiteral(start_pos, scanner()->location().end_pos, CHECK_OK);
+      CheckStrictOctalLiteral(start_pos, scanner()->location().end_pos,
+                              CHECK_OK);
     }
 
     if (allow_harmony_scoping() && strict_mode() == STRICT)
@@ -2846,6 +2857,7 @@ ParserBase<Traits>::ParseTemplateLiteral(ExpressionT tag, int start, bool* ok) {
   if (peek() == Token::TEMPLATE_TAIL) {
     Consume(Token::TEMPLATE_TAIL);
     int pos = position();
+    CheckTemplateOctalLiteral(pos, peek_position(), CHECK_OK);
     typename Traits::TemplateLiteralState ts = Traits::OpenTemplateLiteral(pos);
     Traits::AddTemplateSpan(&ts, true);
     return Traits::CloseTemplateLiteral(&ts, start, tag);
@@ -2897,6 +2909,7 @@ ParserBase<Traits>::ParseTemplateLiteral(ExpressionT tag, int start, bool* ok) {
   } while (next == Token::TEMPLATE_SPAN);
 
   DCHECK_EQ(next, Token::TEMPLATE_TAIL);
+  CheckTemplateOctalLiteral(pos, peek_position(), CHECK_OK);
   // Once we've reached a TEMPLATE_TAIL, we can close the TemplateLiteral.
   return Traits::CloseTemplateLiteral(&ts, start, tag);
 }

src/scanner.cc

@@ -733,21 +733,7 @@ bool Scanner::ScanEscape() {
       if (c < 0) return false;
       break;
     }
-    case '0':
+    case '0':  // Fall through.
-      if (in_template_literal) {
-        // \ 0 DecimalDigit is never allowed in templates.
-        if (IsDecimalDigit(c0_)) {
-          Advance<capture_raw>();  // Advance to include the problematic char.
-          return false;
-        }
-
-        // The TV of TemplateCharacter :: \ EscapeSequence is the CV of
-        //     EscapeSequence.
-        // The CV of EscapeSequence :: 0 is the code unit value 0.
-        c = 0;
-        break;
-      }
-    // Fall through.
     case '1':  // fall through
     case '2':  // fall through
     case '3':  // fall through
@@ -755,14 +741,8 @@ bool Scanner::ScanEscape() {
     case '5':  // fall through
     case '6':  // fall through
     case '7':
-      if (!in_template_literal) {
+      c = ScanOctalEscape<capture_raw>(c, 2);
-        c = ScanOctalEscape<capture_raw>(c, 2);
+      break;
-        break;
-      }
-    // Fall through
-    case '8':
-    case '9':
-      if (in_template_literal) return false;
   }
 
   // According to ECMA-262, section 7.8.4, characters not covered by the

test/mjsunit/harmony/templates.js

@@ -476,9 +476,11 @@ var obj = {
 (function testLegacyOctal() {
   assertEquals('\u0000', `\0`);
   assertEquals('\u0000a', `\0a`);
-  for (var i = 0; i < 10; i++) {
+  for (var i = 0; i < 8; i++) {
     var code = "`\\0" + i + "`";
     assertThrows(code, SyntaxError);
+    code = "(function(){})" + code;
+    assertThrows(code, SyntaxError);
   }
 
   assertEquals('\\0', String.raw`\0`);
@@ -488,8 +490,18 @@ var obj = {
 (function testSyntaxErrorsNonEscapeCharacter() {
   assertThrows("`\\x`", SyntaxError);
   assertThrows("`\\u`", SyntaxError);
-  for (var i = 1; i < 10; i++) {
+  for (var i = 1; i < 8; i++) {
     var code = "`\\" + i + "`";
     assertThrows(code, SyntaxError);
+    code = "(function(){})" + code;
+    assertThrows(code, SyntaxError);
   }
 })();
+
+
+(function testValidNumericEscapes() {
+  assertEquals("8", `\8`);
+  assertEquals("9", `\9`);
+  assertEquals("\u00008", `\08`);
+  assertEquals("\u00009", `\09`);
+})();