valuetostring.cpp

OSB-PIK-OpenVXI-3.0.0源代码 “中国XML论坛 - 专业的XML技术讨论区--XML在语音技术中的应用”

 /****************License************************************************
  *
  * Copyright 2000-2003.  ScanSoft, Inc.    
  *
  * Use of this software is subject to notices and obligations set forth 
  * in the SpeechWorks Public License - Software Version 1.2 which is 
  * included with this software. 
  *
  * ScanSoft is a registered trademark of ScanSoft, Inc., and OpenSpeech, 
  * SpeechWorks and the SpeechWorks logo are registered trademarks or 
  * trademarks of SpeechWorks International, Inc. in the United States 
  * and other countries.
  *
  ***********************************************************************/
 

 #define VXIVALUE_EXPORTS
 #include "VXIvalue.h"                    // Header for this function
 
 #ifndef NO_STL
 
 #include <stdio.h>                       // For sprintf( )
 #include <string.h>                      // For strlen( )
 #include <wchar.h>
 #include <string>                        // For std::basic_string
 #include <vector>
 #include <iostream>
 
 #ifdef WIN32
 #define WIN32_LEAN_AND_MEAN
 #include <windows.h>
 
 static inline char GetDecimalSeparator( ) {
   // Get the locale dependant decimal separator using a Win32 API call
   char sepBuf[32];
   return (GetLocaleInfoA (LOCALE_USER_DEFAULT, LOCALE_SDECIMAL,
 			  sepBuf, 32) > 0 ? sepBuf[0] : '\0');
 }
 #else
 #include <locale.h>
 
 static inline char GetDecimalSeparator( ) {
   // Get the locale dependant decimal separator using a POSIX call
   struct lconv *info = localeconv( );
   return (info ? info->decimal_point[0] : '\0');
 }
 #endif
 
 // Definitions of VXIchar based C++ string types
 typedef std::basic_string<VXIchar> myString;
 #ifdef WIN32
 typedef std::basic_string<VXIbyte> myUTF8String;
 #else
 // Portability: stl on UNIX is not always portable, therefore using
 // std::basic_string<unsigned char> is not always compiled and linked.
 // The simple solution for this issue is to represent unsigned char string 
 // as a vector
 typedef std::vector<VXIbyte> myUTF8String;
 #endif
 
 // Prototypes for local functions
 static void appendValue(myString &result,
 			const VXIValue *value,
 			const VXIchar *fieldName);
 
 // These functions ensure that the string has enough capacity to hold the extra characters.
 // To ensure amortized linear time, we reserve twice as many bytes as required.
 template <class T> static void ensureCapacity(std::basic_string<T>& buffer,
                                               std::basic_string<T>::size_type extraChars)
 {
   std::basic_string<T>::size_type requiredCapacity = buffer.length() + extraChars;
   if (requiredCapacity > buffer.capacity())
   {
     buffer.reserve(2 * requiredCapacity);
   }
 }
 
 #ifndef WIN32
 static void ensureCapacity(myUTF8String& buffer, unsigned int extraChars)
 {
   unsigned int requiredCapacity = buffer.size() + extraChars;
   if (requiredCapacity > buffer.capacity())
   {
     buffer.reserve(2 * requiredCapacity);
   }
 }
 #endif
 
 //
 // VXIchar string to Unicode UTF-8 string conversion
 //
 static bool convertToUTF8(const VXIchar *input,
 			  myUTF8String &output)
 {
   //  firstByteMark
   //      A list of values to mask onto the first byte of an encoded sequence,
   //      indexed by the number of bytes used to create the sequence.
   static const char firstByteMark[7] =
     {  char(0x00), char(0x00), char(0xC0), char(0xE0),
        char(0xF0), char(0xF8), char(0xFC) };
 
   //  Get pointers to our start and end points of the input buffer.
   const VXIchar* srcPtr = input;
   const VXIchar* srcEnd = srcPtr + wcslen(input);
   output.resize (0);
 
   while (srcPtr < srcEnd) {
     VXIchar curVal = *srcPtr++;
 
     // Watchout for surrogates, if found truncate
     if ((curVal >= 0xD800) && (curVal <= 0xDBFF)) {
       break;
     }
 
     // Figure out how many bytes we need
     unsigned int encodedBytes;
     if (curVal < 0x80)                encodedBytes = 1;
     else if (curVal < 0x800)          encodedBytes = 2;
     else if (curVal < 0x10000)        encodedBytes = 3;
     else if (curVal < 0x200000)       encodedBytes = 4;
     else if (curVal < 0x4000000)      encodedBytes = 5;
     else if (curVal <= 0x7FFFFFFF)    encodedBytes = 6;
     else {
       // THIS SHOULD NOT HAPPEN!
       return false;
     }
 
     //  And spit out the bytes. We spit them out in reverse order
     //  here, so bump up the output pointer and work down as we go.
     char buffer[7] = { 0, 0, 0, 0, 0, 0, 0 };
 
     char * outPtr = buffer + encodedBytes;
     switch(encodedBytes) {
     case 6 : *--outPtr = char((curVal | 0x80) & 0xBF);
       curVal >>= 6;
     case 5 : *--outPtr = char((curVal | 0x80) & 0xBF);
       curVal >>= 6;
     case 4 : *--outPtr = char((curVal | 0x80) & 0xBF);
       curVal >>= 6;
     case 3 : *--outPtr = char((curVal | 0x80) & 0xBF);
       curVal >>= 6;
     case 2 : *--outPtr = char((curVal | 0x80) & 0xBF);
       curVal >>= 6;
     case 1 : *--outPtr = char(curVal | firstByteMark[encodedBytes]);
     }
 
     ensureCapacity(output, encodedBytes);
     for (int i = 0; buffer[i] != 0; i++)
     {
 #ifdef WIN32
       output += buffer[i];
 #else
       output.push_back( buffer[i] );
 #endif      
     }
   }
   return true;
 }
 
 
 //
 // Append escaped binary data
 //
 static void appendEscapedData(myString &result,
 			      const VXIbyte *data,
 			      VXIulong size)
 {
   static const unsigned char isAcceptable[96] =
     /*0x0 0x1 0x2 0x3 0x4 0x5 0x6 0x7 0x8 0x9 0xA 0xB 0xC 0xD 0xE 0xF */
     {
       /* 2x  !"#$%&'()*+,-./   */
       0x0,0x0,0x0,0x0,0x0,0x0,0x0,0x0,0x0,0x0,0xF,0xE,0x0,0xF,0xF,0xC,
       /* 3x 0123456789:;<=>?   */
       0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0x8,0x0,0x0,0x0,0x0,0x0,
       /* 4x @ABCDEFGHIJKLMNO   */
       0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,
       /* 5X PQRSTUVWXYZ[\]^_   */
       0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0x0,0x0,0x0,0x0,0xF,
       /* 6x `abcdefghijklmno   */
       0x0,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,
       /* 7X pqrstuvwxyz{\}~DEL */
       0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0xF,0x0,0x0,0x0,0x0,0x0
     };
 
   static const VXIchar *hexChar = L"0123456789ABCDEF";
   static const int mask = 0x1;
 
   if ( ! data)
     return;
 
   VXIulong i;
   const VXIbyte *p;
   for (i = 0, p = data; i < size; i++, p++)
   {
     VXIbyte a = *p;
     if (a < 32 || a >= 128 || !(isAcceptable[a-32] & mask))
     {
       ensureCapacity(result, 3);
       result += '%';
       result += hexChar[a >> 4];
       result += hexChar[a & 15];
     }
     else
     {
       ensureCapacity(result, 1);
       result += *p;
     }
   }
 }
 
 
 //
 // Append a string. Resulting byte sequence is a URL encoded Unicode
 // UTF-8 string, meaning that the ASCII character set appears as
 // normally done for URL encoded data, and any Latin-1 or Unicode
 // characters will be translated into a URL encoded UTF-8 byte
 // sequence where each character consumes 2 to 18 bytes (2 or more
 // bytes in the UTF-8 representation where each of those bytes is
 // encoded as 1 to 3 bytes depending on whether they need to get
 // escaped for URL encoding).
 //
 static void appendEscapedString(myString &result,
 				const VXIchar *str)
 {
   if (!str)
     return;
 
   // Convert to UTF-8
   myUTF8String utf8;
   utf8.reserve (wcslen(str) * 4); // For efficiency, pre-allocate buffer
   if (! convertToUTF8 (str, utf8))
     return;
     
   // Now URL encode the bytes
 #ifdef WIN32
   appendEscapedData (result, utf8.c_str(), utf8.length());
 #else
   appendEscapedData (result, &utf8[0], utf8.size());  
 #endif
 }
 
 
 static void appendKeyValuePair(myString &result,
 			       const VXIchar *key,
 			       const VXIchar *value)
 {
   if ((key) && (value) && (*key))
   {
     if (result.length() > 0)
     {
       ensureCapacity(result, 1);
       result += L'&';
     }
 
     appendEscapedString(result, key);
     ensureCapacity(result, 1);
     result += L'=';
 
     if (*value)
       appendEscapedString(result, value);
   }
 }
 
 
 static void appendKeyValuePair(myString &result,
 			       const VXIchar *key,
 			       char *value)
 {
   if ((key) && (value) && (*key))
   {
     if (result.length() > 0)
     {
       ensureCapacity(result, 1);
       result += L'&';
     }
 
     appendEscapedString(result, key);
     ensureCapacity(result, 1);
     result += L'=';
     
     if (*value)
       appendEscapedData (result,
 			 reinterpret_cast<const unsigned char *>(value),
 			 strlen(value));
   }
 }
 
 
 static void appendVector(myString &result,
 			 const VXIVector *vxivector,
 			 const VXIchar *fieldName)
 {
   const VXIValue *value = NULL;
 
   VXIunsigned vectorLen = VXIVectorLength(vxivector);
   VXIunsigned i;
 
   for (i = 0; i < vectorLen; i++)
   {
     value = VXIVectorGetElement(vxivector, i);
     if ( value )
     {
       // sprintf( ) is more portable then swprintf( )
       char intStr[64];
       sprintf(intStr, "%d", i);
       myString fieldNameWithIndex(fieldName);
       if (! fieldNameWithIndex.empty())
         fieldNameWithIndex += L'.';
       for (const char *ptr = intStr; *ptr; ptr++)
         fieldNameWithIndex += static_cast<VXIchar>(*ptr);
       appendValue(result, value, fieldNameWithIndex.c_str());
     }
   }
 }
 
 
 static void appendMap(myString &result,
 		      const VXIMap *vximap,
 		      const VXIchar *fieldName)
 {
   const VXIchar  *key = NULL;
   const VXIValue *value = NULL;
 
   VXIMapIterator *mapIterator = VXIMapGetFirstProperty(vximap, &key, &value);
   do
   {
     if ((key) && (value))
     {
       myString fieldNameWithIndex (fieldName);
       if (! fieldNameWithIndex.empty())
         fieldNameWithIndex += L'.';
       fieldNameWithIndex += key;
 
       appendValue(result, value, fieldNameWithIndex.c_str());
     }
   } while (VXIMapGetNextProperty(mapIterator, &key, &value) ==
            VXIvalue_RESULT_SUCCESS);
 
   VXIMapIteratorDestroy(&mapIterator);
 }
 
 
 static void appendContent(myString &result,
 			  const VXIContent *vxicontent,
 			  const VXIchar *fieldName)
 {
   const VXIchar *type;
   const VXIbyte *data;
   VXIulong size;
 
   VXIContentValue(vxicontent, &type, &data, &size);
 
   if (result.length() > 0)
     result += L'&';
 
   appendEscapedString(result, fieldName);
   ensureCapacity(result, 1);
   result += L'=';
   appendEscapedData(result, data, size);
 }
 
 
 static void appendValue(myString &result,
 			const VXIValue *value,
 			const VXIchar *fieldName)
 {
   char valueStr[128];
 
   switch (VXIValueGetType(value)) {
   case VALUE_MAP:
     // nested object
     // append the object name to the field name prefix
     appendMap(result, (const VXIMap *) value, fieldName);
     break;
   case VALUE_VECTOR:
     // nested vector
     // append the vector name to the field name prefix
     appendVector(result, (const VXIVector *) value, fieldName);
     break;
   case VALUE_CONTENT:
     appendContent(result, (const VXIContent *) value, fieldName);
     break;
   case VALUE_BOOLEAN:
     if (VXIBooleanValue((const VXIBoolean *) value) == TRUE)
       appendKeyValuePair(result, fieldName, L"true");
     else
       appendKeyValuePair(result, fieldName, L"false");
     break;
   case VALUE_INTEGER:
     // sprintf( ) is more portable then swprintf( )
     sprintf(valueStr, "%d", VXIIntegerValue((const VXIInteger *) value));
     appendKeyValuePair(result, fieldName, valueStr);
     break;
   case VALUE_FLOAT:
     {
       // sprintf( ) is more portable then swprintf( ), but have to
       // watch for locale dependance where the integer and fractional
       // parts may be separated by a comma in some locales
       sprintf(valueStr, "%#.6g",
 	      (double) VXIFloatValue((const VXIFloat *) value));
       char sep = GetDecimalSeparator( );
       // printf("decimalSeparator = '%c'\n", (sep ? sep : 'e'));
       if ((sep) && (sep != '.')) {
 	char *ptr = strchr(valueStr, sep);
 	if (ptr)
 	  *ptr = '.';
       }
       appendKeyValuePair(result, fieldName, valueStr);
     }
     break;
   case VALUE_STRING:
     appendKeyValuePair(result, fieldName,
 		       VXIStringCStr((const VXIString *) value));
     break;
   case VALUE_PTR:
     {
       // Can't rely on the C library to give consistant enough results.
       // sprintf( ) is more portable then swprintf( )
       sprintf(valueStr, "%p", VXIPtrValue((const VXIPtr *) value));
       myString finalValue;
       if ( strncmp(valueStr, "0x", 2) != 0 )
 	finalValue += L"0x";
       for (const char *ptr = valueStr; *ptr; ptr++) {
 	switch (*ptr) {
 	case 'a': finalValue += L'A'; break;
 	case 'b': finalValue += L'B'; break;
 	case 'c': finalValue += L'C'; break;
 	case 'd': finalValue += L'D'; break;
 	case 'e': finalValue += L'E'; break;
 	case 'f': finalValue += L'F'; break;
 	default:  finalValue += static_cast<VXIchar>(*ptr);
 	}
       }
       appendKeyValuePair(result, fieldName, finalValue.c_str());
     }
     break;
   default:
     appendKeyValuePair(result, fieldName, L"errorInvalidType");
     break;
   }
 }
 
 
 #endif /* #ifndef NO_STL */
 
 
 /**
  * Generic Value to string conversion
  *
  * This converts any VXIValue type to a string.
  *
  * @param   v       Value to convert to a string
  * @param   name    Name to use for labeling the VXIValue data
  * @param   format  Format to use for the string, see above
  * @return          VXIString, NULL on error
  */
 VXIVALUE_API VXIString *VXIValueToString(const VXIValue      *v,
 					 const VXIchar       *name,
 					 VXIValueStringFormat format)
 {
   if (( ! v ) || ( ! name ))
     return NULL;
 
 #ifdef NO_STL
   return NULL;
 #else
   myString str;
   str.reserve(1024); // For efficiency, preallocate 1024 char buffer
   appendValue (str, v, name);
   return (str.length( ) > 0 ? VXIStringCreate (str.c_str()) : NULL);
 #endif
 }