Clean up CPDF_SyntaxParser a little bit

core/fpdfapi/parser/cpdf_syntax_parser.cpp

Index: core/fpdfapi/parser/cpdf_syntax_parser.cpp
diff --git a/core/fpdfapi/parser/cpdf_syntax_parser.cpp b/core/fpdfapi/parser/cpdf_syntax_parser.cpp
index 3b1df94503372795840e022d651e763b3fb7625f..afe386b19ed08c3f7795a9cde5f0f1c979b77b14 100644
--- a/core/fpdfapi/parser/cpdf_syntax_parser.cpp
+++ b/core/fpdfapi/parser/cpdf_syntax_parser.cpp
@@ -31,6 +31,14 @@ struct SearchTagRecord {
   FX_STRSIZE m_Offset;
 };
 
+enum SyntaxParser_ReadStatus {

[dsinclair, 2016/11/01 19:14:42]

Make this an enum class then you can drop the SyntaxParser_Read from the name
prefixes below as they have to be fully qualified.

Rename to just ReadStatus? Then the usage would be ReadStatus::Normal,
ReadStatus::Backlash which reads nicely.

[npm, 2016/11/01 20:09:10]

Done.

+  SyntaxParser_ReadNormal,
+  SyntaxParser_ReadBackslash,
+  SyntaxParser_ReadOctal,
+  SyntaxParser_ReadFinishOctal,
+  SyntaxParser_ReadCarriageReturn
+};
+
 }  // namespace
 
 // static
@@ -57,30 +65,36 @@ FX_BOOL CPDF_SyntaxParser::GetCharAt(FX_FILESIZE pos, uint8_t& ch) {
   return GetNextChar(ch);
 }
 
+FX_BOOL CPDF_SyntaxParser::ReadChar(FX_FILESIZE read_pos, uint32_t read_size) {

[dsinclair, 2016/11/01 19:14:42]

Should this be ReadChars? It will potentially read more then a single character,
yea?

[npm, 2016/11/01 20:09:10]

But it still represents a single char, as it is called in from GetNextChar and
GetCharAtBackward

+  if (static_cast<FX_FILESIZE>(read_pos + read_size) > m_FileLen) {
+    if (m_FileLen < static_cast<FX_FILESIZE>(read_size)) {
+      read_pos = 0;
+      read_size = static_cast<uint32_t>(m_FileLen);
+    } else {
+      read_pos = m_FileLen - read_size;
+    }
+  }
+  if (!m_pFileAccess->ReadBlock(m_pFileBuf, read_pos, read_size))
+    return FALSE;
+
+  m_BufOffset = read_pos;
+  return TRUE;
+}
+
 FX_BOOL CPDF_SyntaxParser::GetNextChar(uint8_t& ch) {
   FX_FILESIZE pos = m_Pos + m_HeaderOffset;
   if (pos >= m_FileLen)
     return FALSE;
 
-  if (m_BufOffset >= pos || (FX_FILESIZE)(m_BufOffset + m_BufSize) <= pos) {
+  if (m_BufOffset >= pos ||
+      static_cast<FX_FILESIZE>(m_BufOffset + m_BufSize) <= pos) {
     FX_FILESIZE read_pos = pos;
     uint32_t read_size = m_BufSize;
-    if ((FX_FILESIZE)read_size > m_FileLen)
-      read_size = (uint32_t)m_FileLen;
-
-    if ((FX_FILESIZE)(read_pos + read_size) > m_FileLen) {
-      if (m_FileLen < (FX_FILESIZE)read_size) {
-        read_pos = 0;
-        read_size = (uint32_t)m_FileLen;
-      } else {
-        read_pos = m_FileLen - read_size;
-      }
-    }
-
-    if (!m_pFileAccess->ReadBlock(m_pFileBuf, read_pos, read_size))
+    if (static_cast<FX_FILESIZE>(read_size) > m_FileLen)
+      read_size = static_cast<uint32_t>(m_FileLen);
+    if (!ReadChar(read_pos, read_size))
       return FALSE;
 
-    m_BufOffset = read_pos;
   }
   ch = m_pFileBuf[pos - m_BufOffset];
   m_Pos++;
@@ -92,27 +106,17 @@ FX_BOOL CPDF_SyntaxParser::GetCharAtBackward(FX_FILESIZE pos, uint8_t& ch) {
   if (pos >= m_FileLen)
     return FALSE;
 
-  if (m_BufOffset >= pos || (FX_FILESIZE)(m_BufOffset + m_BufSize) <= pos) {
+  if (m_BufOffset >= pos ||
+      static_cast<FX_FILESIZE>(m_BufOffset + m_BufSize) <= pos) {
     FX_FILESIZE read_pos;
-    if (pos < (FX_FILESIZE)m_BufSize)
+    if (pos < static_cast<FX_FILESIZE>(m_BufSize))
       read_pos = 0;
     else
       read_pos = pos - m_BufSize + 1;
-
     uint32_t read_size = m_BufSize;
-    if ((FX_FILESIZE)(read_pos + read_size) > m_FileLen) {
-      if (m_FileLen < (FX_FILESIZE)read_size) {
-        read_pos = 0;
-        read_size = (uint32_t)m_FileLen;
-      } else {
-        read_pos = m_FileLen - read_size;
-      }
-    }
-
-    if (!m_pFileAccess->ReadBlock(m_pFileBuf, read_pos, read_size))
+    if (!ReadChar(read_pos, read_size))
       return FALSE;
 
-    m_BufOffset = read_pos;
   }
   ch = m_pFileBuf[pos - m_BufOffset];
   return TRUE;
@@ -215,30 +219,27 @@ CFX_ByteString CPDF_SyntaxParser::ReadString() {
 
   CFX_ByteTextBuf buf;
   int32_t parlevel = 0;
-  int32_t status = 0;
+  int32_t status = SyntaxParser_ReadNormal;
   int32_t iEscCode = 0;
   while (1) {
     switch (status) {
-      case 0:
+      case SyntaxParser_ReadNormal:
         if (ch == ')') {
-          if (parlevel == 0) {
+          if (parlevel == 0)
             return buf.MakeString();
-          }
           parlevel--;
-          buf.AppendChar(')');
         } else if (ch == '(') {
           parlevel++;
-          buf.AppendChar('(');
-        } else if (ch == '\\') {
-          status = 1;
-        } else {
-          buf.AppendChar(ch);
         }
+        if (ch == '\\')
+          status = SyntaxParser_ReadBackslash;
+        else
+          buf.AppendChar(ch);
         break;
-      case 1:
+      case SyntaxParser_ReadBackslash:
         if (ch >= '0' && ch <= '7') {
           iEscCode = FXSYS_toDecimalDigit(static_cast<FX_WCHAR>(ch));
-          status = 2;
+          status = SyntaxParser_ReadOctal;
           break;
         }
 
@@ -253,38 +254,38 @@ CFX_ByteString CPDF_SyntaxParser::ReadString() {
         } else if (ch == 'f') {
           buf.AppendChar('\f');
         } else if (ch == '\r') {
-          status = 4;
+          status = SyntaxParser_ReadCarriageReturn;
           break;
         } else if (ch != '\n') {
           buf.AppendChar(ch);
         }
-        status = 0;
+        status = SyntaxParser_ReadNormal;
         break;
-      case 2:
+      case SyntaxParser_ReadOctal:
         if (ch >= '0' && ch <= '7') {
           iEscCode =
               iEscCode * 8 + FXSYS_toDecimalDigit(static_cast<FX_WCHAR>(ch));
-          status = 3;
+          status = SyntaxParser_ReadFinishOctal;
         } else {
           buf.AppendChar(iEscCode);
-          status = 0;
+          status = SyntaxParser_ReadNormal;
           continue;
         }
         break;
-      case 3:
+      case SyntaxParser_ReadFinishOctal:
         if (ch >= '0' && ch <= '7') {
           iEscCode =
               iEscCode * 8 + FXSYS_toDecimalDigit(static_cast<FX_WCHAR>(ch));
           buf.AppendChar(iEscCode);
-          status = 0;
+          status = SyntaxParser_ReadNormal;

[dsinclair, 2016/11/01 19:14:42]

Move this before the if() as it happens in both branches?

[npm, 2016/11/01 20:09:10]

Done.

         } else {
           buf.AppendChar(iEscCode);
-          status = 0;
+          status = SyntaxParser_ReadNormal;
           continue;
         }
         break;
-      case 4:
-        status = 0;
+      case SyntaxParser_ReadCarriageReturn:
+        status = SyntaxParser_ReadNormal;
         if (ch != '\n')
           continue;
         break;
@@ -648,7 +649,9 @@ CPDF_Stream* CPDF_SyntaxParser::ReadStream(CPDF_Dictionary* pDict,
   const CFX_ByteStringC kEndObjStr("endobj");
 
   CPDF_CryptoHandler* pCryptoHandler =
-      objnum == (uint32_t)m_MetadataObjnum ? nullptr : m_pCryptoHandler.get();
+      objnum == static_cast<uint32_t>(m_MetadataObjnum)
+          ? nullptr
+          : m_pCryptoHandler.get();
   if (!pCryptoHandler) {
     FX_BOOL bSearchForKeyword = TRUE;
     if (len >= 0) {
@@ -794,7 +797,9 @@ void CPDF_SyntaxParser::InitParser(IFX_SeekableReadStream* pFileAccess,
   m_BufOffset = 0;
   pFileAccess->ReadBlock(
       m_pFileBuf, 0,
-      (size_t)((FX_FILESIZE)m_BufSize > m_FileLen ? m_FileLen : m_BufSize));
+      static_cast<size_t>(static_cast<FX_FILESIZE>(m_BufSize) > m_FileLen
+                              ? m_FileLen
+                              : m_BufSize));
 }
 
 uint32_t CPDF_SyntaxParser::GetDirectNum() {
@@ -905,62 +910,6 @@ FX_BOOL CPDF_SyntaxParser::SearchWord(const CFX_ByteStringC& tag,
   return FALSE;
 }
 
-int32_t CPDF_SyntaxParser::SearchMultiWord(const CFX_ByteStringC& tags,
-                                           FX_BOOL bWholeWord,
-                                           FX_FILESIZE limit) {
-  int32_t ntags = 1;
-  for (int i = 0; i < tags.GetLength(); ++i) {
-    if (tags[i] == 0)
-      ++ntags;
-  }
-
-  // Ensure that the input byte string happens to be nul-terminated. This
-  // need not be the case, but the loop below uses this guarantee to put
-  // the last pattern into the vector.
-  ASSERT(tags[tags.GetLength()] == 0);
-  std::vector<SearchTagRecord> patterns(ntags);
-  uint32_t start = 0;
-  uint32_t itag = 0;
-  uint32_t max_len = 0;
-  for (int i = 0; i <= tags.GetLength(); ++i) {
-    if (tags[i] == 0) {
-      uint32_t len = i - start;
-      max_len = std::max(len, max_len);
-      patterns[itag].m_bsTag = tags.Mid(start, len);
-      patterns[itag].m_Offset = 0;
-      start = i + 1;
-      ++itag;
-    }
-  }
-
-  const FX_FILESIZE pos_limit = m_Pos + limit;
-  for (FX_FILESIZE pos = m_Pos; !limit || pos < pos_limit; ++pos) {
-    uint8_t byte;
-    if (!GetCharAt(pos, byte))
-      break;
-
-    for (int i = 0; i < ntags; ++i) {
-      SearchTagRecord& pat = patterns[i];
-      if (pat.m_bsTag[pat.m_Offset] != byte) {
-        pat.m_Offset = (pat.m_bsTag[0] == byte) ? 1 : 0;
-        continue;
-      }
-
-      ++pat.m_Offset;
-      if (pat.m_Offset != pat.m_bsTag.GetLength())
-        continue;
-
-      if (!bWholeWord || IsWholeWord(pos - pat.m_bsTag.GetLength(), limit,
-                                     pat.m_bsTag, FALSE)) {
-        return i;
-      }
-
-      pat.m_Offset = (pat.m_bsTag[0] == byte) ? 1 : 0;
-    }
-  }
-  return -1;
-}
-
 FX_FILESIZE CPDF_SyntaxParser::FindTag(const CFX_ByteStringC& tag,
                                        FX_FILESIZE limit) {
   int32_t taglen = tag.GetLength();