From 821e59c5d2034066843dd6040a06fe637a6369df Mon Sep 17 00:00:00 2001
From: Henrique Nakashima <hnakashima@chromium.org>
Date: Wed, 9 Aug 2017 17:02:51 -0400
Subject: Remove CFX_ByteTextBuf from CFX_UTF8Encoder

Also making CFX_ByteString(ostringstream) const.

Bug: pdfium:731
Change-Id: Iaee610248a5634e8731eaf05944c32d52c4e8015
Reviewed-on: https://pdfium-review.googlesource.com/10151
Commit-Queue: Henrique Nakashima <hnakashima@chromium.org>
Reviewed-by: Tom Sepez <tsepez@chromium.org>
---
 core/fxcrt/fx_basic.h       | 13 +++++++++----
 core/fxcrt/fx_basic_utf.cpp | 38 +++++++++++++++++++++++++++-----------
 2 files changed, 36 insertions(+), 15 deletions(-)

diff --git a/core/fxcrt/fx_basic.h b/core/fxcrt/fx_basic.h
index e856a3e0ba..c6ffc0d535 100644
--- a/core/fxcrt/fx_basic.h
+++ b/core/fxcrt/fx_basic.h
@@ -9,6 +9,7 @@
 
 #include <algorithm>
 #include <memory>
+#include <vector>
 
 #include "core/fxcrt/cfx_retain_ptr.h"
 #include "core/fxcrt/fx_memory.h"
@@ -128,14 +129,18 @@ class CFX_UTF8Decoder {
 
 class CFX_UTF8Encoder {
  public:
-  CFX_UTF8Encoder() {}
+  CFX_UTF8Encoder();
+  ~CFX_UTF8Encoder();
 
   void Input(wchar_t unicode);
-  void AppendStr(const CFX_ByteStringC& str) { m_Buffer << str; }
-  CFX_ByteStringC GetResult() const { return m_Buffer.AsStringC(); }
+  void AppendStr(const CFX_ByteStringC& str);
+
+  // The data returned by GetResult() is invalidated when this is modified by
+  // appending any data.
+  CFX_ByteStringC GetResult() const;
 
  private:
-  CFX_ByteTextBuf m_Buffer;
+  std::vector<uint8_t> m_Buffer;
 };
 
 template <class DataType, int FixedSize>
diff --git a/core/fxcrt/fx_basic_utf.cpp b/core/fxcrt/fx_basic_utf.cpp
index 4dbfa37284..fcfd5934e4 100644
--- a/core/fxcrt/fx_basic_utf.cpp
+++ b/core/fxcrt/fx_basic_utf.cpp
@@ -10,9 +10,11 @@ void CFX_UTF8Decoder::Clear() {
   m_Buffer.Clear();
   m_PendingBytes = 0;
 }
+
 void CFX_UTF8Decoder::AppendChar(uint32_t ch) {
   m_Buffer.AppendChar((wchar_t)ch);
 }
+
 void CFX_UTF8Decoder::Input(uint8_t byte) {
   if (byte < 0x80) {
     m_PendingBytes = 0;
@@ -43,37 +45,51 @@ void CFX_UTF8Decoder::Input(uint8_t byte) {
     m_PendingChar = (byte & 0x01) << 30;
   }
 }
-void CFX_UTF8Encoder::Input(wchar_t unicode) {
-  if ((uint32_t)unicode < 0x80) {
-    m_Buffer.AppendChar(unicode);
+
+CFX_UTF8Encoder::CFX_UTF8Encoder() {}
+
+CFX_UTF8Encoder::~CFX_UTF8Encoder() {}
+
+void CFX_UTF8Encoder::Input(wchar_t unicodeAsWchar) {
+  uint32_t unicode = static_cast<uint32_t>(unicodeAsWchar);
+  if (unicode < 0x80) {
+    m_Buffer.push_back(unicode);
   } else {
-    if ((uint32_t)unicode >= 0x80000000) {
+    if (unicode >= 0x80000000) {
       return;
     }
     int nbytes = 0;
-    if ((uint32_t)unicode < 0x800) {
+    if (unicode < 0x800) {
       nbytes = 2;
-    } else if ((uint32_t)unicode < 0x10000) {
+    } else if (unicode < 0x10000) {
       nbytes = 3;
-    } else if ((uint32_t)unicode < 0x200000) {
+    } else if (unicode < 0x200000) {
       nbytes = 4;
-    } else if ((uint32_t)unicode < 0x4000000) {
+    } else if (unicode < 0x4000000) {
       nbytes = 5;
     } else {
       nbytes = 6;
     }
     static uint8_t prefix[] = {0xc0, 0xe0, 0xf0, 0xf8, 0xfc};
     int order = 1 << ((nbytes - 1) * 6);
-    int code = unicode;
-    m_Buffer.AppendChar(prefix[nbytes - 2] | (code / order));
+    int code = unicodeAsWchar;
+    m_Buffer.push_back(prefix[nbytes - 2] | (code / order));
     for (int i = 0; i < nbytes - 1; i++) {
       code = code % order;
       order >>= 6;
-      m_Buffer.AppendChar(0x80 | (code / order));
+      m_Buffer.push_back(0x80 | (code / order));
     }
   }
 }
 
+void CFX_UTF8Encoder::AppendStr(const CFX_ByteStringC& str) {
+  m_Buffer.insert(m_Buffer.end(), str.begin(), str.end());
+}
+
+CFX_ByteStringC CFX_UTF8Encoder::GetResult() const {
+  return CFX_ByteStringC(m_Buffer.data(), m_Buffer.size());
+}
+
 CFX_ByteString FX_UTF8Encode(const CFX_WideStringC& wsStr) {
   FX_STRSIZE len = wsStr.GetLength();
   const wchar_t* pStr = wsStr.unterminated_c_str();
-- 
cgit v1.2.3