[log] Support logging two-byte characters

Bug: chromium:1130673
Change-Id: I78ae388daa1c4c2b594981bdadd201c2dfb39eb0
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2426618
Reviewed-by: Marja Hölttä <marja@chromium.org>
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70116}

src/logging/log-utils.cc

@@ -128,12 +128,19 @@ void Log::MessageBuilder::AppendString(const char* str) {
   AppendString(str, strlen(str));
 }
 
-void Log::MessageBuilder::AppendString(const char* str, size_t length) {
+void Log::MessageBuilder::AppendString(const char* str, size_t length,
+                                       bool is_one_byte) {
   if (str == nullptr) return;
-
-  for (size_t i = 0; i < length; i++) {
-    DCHECK_NE(str[i], '\0');
-    AppendCharacter(str[i]);
+  if (is_one_byte) {
+    for (size_t i = 0; i < length; i++) {
+      DCHECK_IMPLIES(is_one_byte, str[i] != '\0');
+      AppendCharacter(str[i]);
+    }
+  } else {
+    DCHECK_EQ(length % 2, 0);
+    for (size_t i = 0; i + 1 < length; i += 2) {
+      AppendTwoByteCharacter(str[i], str[i + 1]);
+    }
   }
 }
 
@@ -148,6 +155,14 @@ void Log::MessageBuilder::AppendFormatString(const char* format, ...) {
   }
 }
 
+void Log::MessageBuilder::AppendTwoByteCharacter(char c1, char c2) {
+  if (c2 == 0) {
+    AppendCharacter(c1);
+  } else {
+    // Escape non-printable characters.
+    AppendRawFormatString("\\u%02x%02x", c1 & 0xFF, c2 & 0xFF);
+  }
+}
 void Log::MessageBuilder::AppendCharacter(char c) {
   if (c >= 32 && c <= 126) {
     if (c == ',') {

src/logging/log-utils.h

@@ -69,9 +69,10 @@ class Log {
                       base::Optional<int> length_limit = base::nullopt);
     void AppendString(Vector<const char> str);
     void AppendString(const char* str);
-    void AppendString(const char* str, size_t length);
+    void AppendString(const char* str, size_t length, bool is_one_byte = true);
     void PRINTF_FORMAT(2, 3) AppendFormatString(const char* format, ...);
     void AppendCharacter(char c);
+    void AppendTwoByteCharacter(char c1, char c2);
     void AppendSymbolName(Symbol symbol);
 
     // Delegate insertion to the underlying {log_}.

src/logging/log.cc

@@ -1602,7 +1602,7 @@ void Logger::FunctionEvent(const char* reason, int script_id, double time_delta,
 void Logger::FunctionEvent(const char* reason, int script_id, double time_delta,
                            int start_position, int end_position,
                            const char* function_name,
-                           size_t function_name_length) {
+                           size_t function_name_length, bool is_one_byte) {
   if (!FLAG_log_function_events) return;
   std::unique_ptr<Log::MessageBuilder> msg_ptr = log_->NewMessageBuilder();
   if (!msg_ptr) return;
@@ -1610,7 +1610,7 @@ void Logger::FunctionEvent(const char* reason, int script_id, double time_delta,
   AppendFunctionMessage(msg, reason, script_id, time_delta, start_position,
                         end_position, &timer_);
   if (function_name_length > 0) {
-    msg.AppendString(function_name, function_name_length);
+    msg.AppendString(function_name, function_name_length, is_one_byte);
   }
   msg.WriteToLogFile();
 }

src/logging/log.h

@@ -170,7 +170,7 @@ class Logger : public CodeEventListener {
   void FunctionEvent(const char* reason, int script_id, double time_delta_ms,
                      int start_position, int end_position,
                      const char* function_name = nullptr,
-                     size_t function_name_length = 0);
+                     size_t function_name_length = 0, bool is_one_byte = true);
 
   void CompilationCacheEvent(const char* action, const char* cache_type,
                              SharedFunctionInfo sfi);

src/parsing/parser.cc

@@ -2507,7 +2507,7 @@ FunctionLiteral* Parser::ParseFunctionLiteral(
         event_name, flags().script_id(), ms, scope->start_position(),
         scope->end_position(),
         reinterpret_cast<const char*>(function_name->raw_data()),
-        function_name->byte_length());
+        function_name->byte_length(), function_name->is_one_byte());
   }
   if (V8_UNLIKELY(TracingFlags::is_runtime_stats_enabled()) &&
       did_preparse_successfully) {

src/parsing/preparser.cc

@@ -347,14 +347,16 @@ PreParser::Expression PreParser::ParseFunctionLiteral(
     // reconstructed from the script id and the byte range in the log processor.
     const char* name = "";
     size_t name_byte_length = 0;
+    bool is_one_byte = true;
     const AstRawString* string = function_name.string_;
     if (string != nullptr) {
       name = reinterpret_cast<const char*>(string->raw_data());
       name_byte_length = string->byte_length();
+      is_one_byte = string->is_one_byte();
     }
     logger_->FunctionEvent(
         event_name, flags().script_id(), ms, function_scope->start_position(),
-        function_scope->end_position(), name, name_byte_length);
+        function_scope->end_position(), name, name_byte_length, is_one_byte);
   }
 
   return Expression::Default();

test/mjsunit/tools/log_two_byte.js

@@ -0,0 +1,40 @@
+// Copyright 2020 the V8 project authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+// Flags: --logfile='+' --log --log-code --log-function-events --no-stress-opt
+
+let twoByteName = "twoByteName_🍕"
+let o = {
+  [twoByteName](obj) {
+    return obj.a
+  }
+}
+function testFunctionWithFunnyName(o) {
+  return o.a;
+}
+
+(function testLoopWithFunnyName() {
+  let object = {a:1};
+  let result = 0;
+  for (let i = 0; i < 1000; i++) {
+    result += o[twoByteName](object);
+  }
+  console.log(result);
+  console.log(twoByteName);
+
+})();
+
+var __v_3 = {};
+({})['foobar\u2653'] = null;
+eval('__v_3 = function foobar() { return foobar };');
+__v_3();
+
+const log = d8.log.getAndStop();
+console.log(log)
+// Check that we have a minimally working log file.
+assertTrue(log.length > 0);
+assertTrue(log.indexOf('v8-version') == 0);
+assertTrue(log.indexOf('testFunctionWithFunnyName') >= 10);
+assertTrue(log.indexOf("twoByteName") >= 10);
+assertTrue(log.indexOf('testLoopWithFunnyName') >= 10);