String.cpp

Go to the documentation of this file.

/*******************************************************************************
* Copyright (C) 2005, Vintela, Inc. All rights reserved.
* Copyright (C) 2006, Novell, Inc. All rights reserved.
* 
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
* 
*     * Redistributions of source code must retain the above copyright notice,
*       this list of conditions and the following disclaimer.
*     * Redistributions in binary form must reproduce the above copyright
*       notice, this list of conditions and the following disclaimer in the
*       documentation and/or other materials provided with the distribution.
*     * Neither the name of 
*       Vintela, Inc., 
*       nor Novell, Inc., 
*       nor the names of its contributors or employees may be used to 
*       endorse or promote products derived from this software without 
*       specific prior written permission.
* 
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*******************************************************************************/


#include "blocxx/BLOCXX_config.h"
#include "blocxx/String.hpp"
#include "blocxx/Char16.hpp"
#include "blocxx/Array.hpp"
#include "blocxx/StringStream.hpp"
#include "blocxx/Format.hpp"
#include "blocxx/BinarySerialization.hpp"
#include "blocxx/Assertion.hpp"
#include "blocxx/AutoPtr.hpp"
#include "blocxx/Bool.hpp"
#include "blocxx/UTF8Utils.hpp"
#include "blocxx/ExceptionIds.hpp"
#include "blocxx/COWIntrusiveCountableBase.hpp"

#include <cstdio>
#include <cstdlib>
#include <cstring>
#include <cctype>
#include <cstdarg>
#include <cerrno>
#if defined(BLOCXX_HAVE_ISTREAM) && defined(BLOCXX_HAVE_OSTREAM)
#include <istream>
#include <ostream>
#else
#include <iostream>
#endif
#include <cmath> // for HUGE_VAL
#include <cfloat> // for DBL_MANT_DIG

#ifdef BLOCXX_WIN32
#define SNPRINTF _snprintf
#else
#define SNPRINTF snprintf
#endif

namespace BLOCXX_NAMESPACE
{

using std::istream;
using std::ostream;
using std::streambuf;

BLOCXX_DEFINE_EXCEPTION_WITH_ID(StringConversion);

static inline int
strncmpi(const char* s1, const char* s2, size_t n)
{
   String ls1(s1, n);
   String ls2(s2, n);
   return ls1.compareToIgnoreCase(ls2);
}

// class invariant: m_buf points to a null-terminated sequence of characters. m_buf is m_len+1 bytes long.
class String::ByteBuf : public COWIntrusiveCountableBase
{
public:
   ByteBuf(const char* s) :
      m_len(::strlen(s)), m_buf(new char[m_len+1])
   {
      strcpy(m_buf, s);
   }

   ByteBuf(const ByteBuf& arg)
      : COWIntrusiveCountableBase(arg)
      , m_len(arg.m_len)
      , m_buf(new char[m_len+1])
   {
      strcpy(m_buf, arg.m_buf);
   }
   
   ByteBuf(AutoPtrVec<char>& s, size_t len)
      : m_len(len), m_buf(s.release())
   {
   }
   
   ~ByteBuf() { delete [] m_buf; }
   
   ByteBuf& operator= (const ByteBuf& arg)
   {
      char* buf = new char[arg.m_len+1];
      strcpy(buf, arg.m_buf);
      delete [] m_buf;
      m_buf = buf;
      m_len = arg.m_len;
      return *this;
   }
   
   size_t length() const { return m_len; }
   char* data() const { return m_buf; }
   ByteBuf* clone() const { return new ByteBuf(*this); }
private:
   size_t m_len;
   char* m_buf;
};
#if defined(BLOCXX_AIX)
const size_t String::npos = ~0;
#endif
String::String() :
   m_buf(0)
{
}
String::String(Bool parm) :
   m_buf(parm.toString().m_buf)
{
}
String::String(const Char16& parm) :
   m_buf(parm.toString().m_buf)
{
}
String::String(const std::string& str) :
   m_buf(str.empty() ? 0 : new ByteBuf(str.c_str()))
{
}
#if defined(BLOCXX_WIN32)
#define snprintf _snprintf // stupid windoze...
#endif
String::String(Int32 val) :
   m_buf(NULL)
{
   char tmpbuf[32];
   int len = snprintf(tmpbuf, sizeof(tmpbuf), "%d", val);
   AutoPtrVec<char> bfr(new char[len+1]);
   ::snprintf(bfr.get(), len+1, "%d", val);
   m_buf = new ByteBuf(bfr, len);
}
String::String(UInt32 val) :
   m_buf(NULL)
{
   char tmpbuf[32];
   int len = ::snprintf(tmpbuf, sizeof(tmpbuf), "%u", val);
   AutoPtrVec<char> bfr(new char[len+1]);
   ::snprintf(bfr.get(), len+1, "%u", val);
   m_buf = new ByteBuf(bfr, len);
}
#if defined(BLOCXX_INT32_IS_INT) && defined(BLOCXX_INT64_IS_LONG_LONG)
String::String(long val) :
   m_buf(NULL)
{
   char tmpbuf[32];
   int len = snprintf(tmpbuf, sizeof(tmpbuf), "%ld", val);
   AutoPtrVec<char> bfr(new char[len+1]);
   ::snprintf(bfr.get(), len+1, "%ld", val);
   m_buf = new ByteBuf(bfr, len);
}
String::String(unsigned long val) :
   m_buf(NULL)
{
   char tmpbuf[32];
   int len = ::snprintf(tmpbuf, sizeof(tmpbuf), "%lu", val);
   AutoPtrVec<char> bfr(new char[len+1]);
   ::snprintf(bfr.get(), len+1, "%lu", val);
   m_buf = new ByteBuf(bfr, len);
}
#endif
#if defined(BLOCXX_WIN32)
#undef snprintf
#endif
String::String(Int64 val) :
   m_buf(NULL)
{
   OStringStream ss(33);
   ss << val;
   m_buf = new ByteBuf(ss.c_str());
}
String::String(UInt64 val) :
   m_buf(NULL)
{
#if defined(BLOCXX_INT64_IS_LONG)
   char tmpbuf[32];
   ::snprintf(tmpbuf, sizeof(tmpbuf), "%lu", val);
   m_buf = new ByteBuf(tmpbuf);
#elif defined(BLOCXX_INT64_IS_LONG_LONG)
   // unfortunately not all C libraries support long long with snprintf().
   // but the C++ iostream library handles it.
   OStringStream ss;
   ss << val;
   m_buf = new ByteBuf(ss.c_str());
#endif
}
// decimal digits = ceiling((bits)*ln(2)/ln(10))
String::String(Real32 val) :
   m_buf(NULL)
{
   char tmpbuf[128];
#if FLT_RADIX == 2
#if defined(BLOCXX_REAL32_IS_FLOAT)
   ::SNPRINTF(tmpbuf, sizeof(tmpbuf), "%.*g", FLT_MANT_DIG * 3 / 10 + 1, static_cast<double>(val));
#elif defined(BLOCXX_REAL32_IS_DOUBLE)
   ::SNPRINTF(tmpbuf, sizeof(tmpbuf), "%.*g", DBL_MANT_DIG * 3 / 10 + 1, val);
#endif
#else
#error "The formula for computing the number of digits of precision for a floating point needs to be implmented. It's ceiling(bits * log(FLT_RADIX) / log(10))"
#endif
   m_buf = new ByteBuf(tmpbuf);
}
String::String(Real64 val) :
   m_buf(NULL)
{
   char tmpbuf[128];
#if FLT_RADIX == 2
#if defined(BLOCXX_REAL64_IS_DOUBLE)
   ::SNPRINTF(tmpbuf, sizeof(tmpbuf), "%.*g", DBL_MANT_DIG * 3 / 10 + 1, val);
#elif defined(BLOCXX_REAL64_IS_LONG_DOUBLE)
   ::SNPRINTF(tmpbuf, sizeof(tmpbuf), "%.*Lg", LDBL_MANT_DIG * 3 / 10 + 1, val);
#endif
#else
#error "The formula for computing the number of digits of precision for a floating point needs to be implmented. It's ceiling(bits * log(FLT_RADIX) / log(10))"
#endif
   m_buf = new ByteBuf(tmpbuf);
}
String::String(const char* str) :
   m_buf(NULL)
{
   m_buf = (NULL == str) ? 0 : new ByteBuf(str);
}
String::String(ETakeOwnershipFlag, char* allocatedMemory, size_t len) :
   m_buf(NULL)
{
   BLOCXX_ASSERT(allocatedMemory != 0);
   AutoPtrVec<char> p(allocatedMemory);
   m_buf = new ByteBuf(p, len);
}
String::String(const char* str, size_t len) :
   m_buf(NULL)
{
   if (NULL == str)
   {
      m_buf = 0;
   }
   else
   {
      AutoPtrVec<char> bfr(new char[len+1]);
      ::memcpy(bfr.get(), str, len);
      bfr[len] = '\0';
      m_buf = new ByteBuf(bfr, len);
   }
}
String::String(const String& arg) :
   m_buf(arg.m_buf)
{
}
String::String(const Char16Array& ra) :
   m_buf(NULL)
{
   size_t sz = ra.size();
   if (sz > 0)
   {
      StringBuffer buf(sz * 2);
      for (size_t i = 0; i < sz; i++)
      {
         buf += ra[i].toString();
      }
      m_buf = buf.releaseString().m_buf;
   }
   else
   {
      m_buf = 0;
   }
}
String::String(char c) :
   m_buf(NULL)
{
   if (c != '\0')
   {
      char bfr[2];
      bfr[0] = c;
      bfr[1] = '\0';
      m_buf = new ByteBuf(bfr);
   }
   else
   {
      m_buf = 0;
   }
}
String::~String() 
{
}
void
String::swap(String& x)
{
   m_buf.swap(x.m_buf);
}
char*
String::allocateCString() const
{
   size_t len = length() + 1;
   char* str = static_cast<char*>(malloc(len));
   ::strcpy(str, c_str());
   return str;
}
size_t
String::length() const
{
   return (m_buf) ? m_buf->length() : 0;
}
size_t
String::UTF8Length() const
{
   return UTF8Utils::charCount(c_str());
}
#ifdef BLOCXX_WIN32
#define vsnprintf _vsnprintf // stupid windoze
#endif
int
String::format(const char* fmt, ...)
{
   int n, size = 64;
   AutoPtrVec<char> p(new char[size]);
   
   va_list ap;
   
   // Try to print in the allocated space
   while (true)
   {
      va_start(ap, fmt);
      n = vsnprintf(p.get(), size, fmt, ap);
      va_end(ap);                // If that worked, return the string.
      if (n > -1 && n < size)
      {
         m_buf = new ByteBuf(p, n);
         return static_cast<int>(length());
      }
      if (n > -1)    // glibc 2.1
         size = n+1; // precisely what is needed
      else           // glibc 2.0
         size *= 2;  // twice the old size
      p = new char[size];
   }
   // Not reachable.
   return 0;
}
#ifdef BLOCXX_WIN32
#undef vsnprintf // stupid windoze
#endif
char
String::charAt(size_t ndx) const
{
   return (m_buf) ? m_buf->data()[ndx] : '\0';
}
int
String::compareTo(const char* arg) const
{
   const char* lhs = "";
   if (m_buf)
   {
      lhs = m_buf->data();
   }
   return ::strcmp(lhs, arg);
}
int
String::compareTo(const String& arg) const
{
   return compareTo(arg.c_str());
}
int
String::compareToIgnoreCase(const char* arg) const
{
   const char* lhs = "";
   if (m_buf)
   {
      lhs = m_buf->data();
   }
   return UTF8Utils::compareToIgnoreCase(lhs, arg);
}
int
String::compareToIgnoreCase(const String& arg) const
{
   return compareToIgnoreCase(arg.c_str());
}

String&
String::concat(const char* arg)
{
   if (arg && *arg)
   {
      size_t len = length() + ::strlen(arg);
      AutoPtrVec<char> bfr(new char[len+1]);
      bfr[0] = 0;
      if (m_buf)
      {
         ::strcpy(bfr.get(), m_buf->data());
      }
      ::strcat(bfr.get(), arg);
      m_buf = new ByteBuf(bfr, len);
   }
   return *this;
}

String&
String::concat(char arg)
{
   size_t newlen = length() + 1;
   AutoPtrVec<char> bfr(new char[newlen+1]);
   bfr[0] = 0;
   if (m_buf)
   {
      ::strcpy(bfr.get(), m_buf->data());
   }
   *(bfr.get()+length()) = arg;
   *(bfr.get()+newlen) = 0;
   m_buf = new ByteBuf(bfr, newlen);
   return *this;
}

bool
String::endsWith(char arg) const
{
   return (m_buf
         && m_buf->length()
         && m_buf->data()[m_buf->length()-1] == arg);
}

bool
String::endsWith(const char* arg, EIgnoreCaseFlag ignoreCase) const
{
   if (!arg || !*arg)
   {
      return (length() == 0);
   }

   if (!m_buf)
   {
      return false;
   }

   int ndx = static_cast<int>(length() - ::strlen(arg));
   if (ndx < 0)
   {
      return false;
   }

   return (ignoreCase)
      ? (UTF8Utils::compareToIgnoreCase(m_buf->data()+ndx, arg) == 0)
      : (::strcmp(m_buf->data()+ndx, arg) == 0);
}
bool
String::equals(const char* arg) const
{
   return(compareTo(arg) == 0);
}
bool
String::equals(const String& arg) const
{
   return equals(arg.c_str());
}
bool
String::equalsIgnoreCase(const char* arg) const
{
   return(compareToIgnoreCase(arg) == 0);
}
bool
String::equalsIgnoreCase(const String& arg) const
{
   return equalsIgnoreCase(arg.c_str());
}
UInt32
String::hashCode() const
{
   UInt32 hash = 0;
   size_t len = length();
   for (size_t i = 0; i < len; i++)
   {
      // Don't need to check if m_buf is null, because if it is, len == 0,
      // and this loop won't be executed.
      const char temp = m_buf->data()[i];
      hash = (hash << 4) + (temp * 13);
      UInt32 g = hash & 0xf0000000;
      if (g)
      {
         hash ^= (g >> 24);
         hash ^= g;
      }
   }
   return hash;
}
size_t
String::indexOf(char ch, size_t fromIndex) const
{
   //if (fromIndex < 0)
   //{
   // fromIndex = 0;
   //}
   size_t cc = npos;
   if (fromIndex < length())
   {
      // Don't need to check m_buf for NULL, because if length() == 0,
      // this code won't be executed.
      const char* p = String::strchr(m_buf->data()+fromIndex, ch);
      if (p)
      {
         cc = p - m_buf->data();
      }
   }
   return cc;
}
size_t
String::indexOf(const char* arg, size_t fromIndex) const
{
   size_t cc = npos;
   if (fromIndex < length())
   {
      // Don't need to check m_buf for NULL, because if length() == 0,
      // this code won't be executed
      char* p(0);
      if (arg && *arg)
      {
         p = ::strstr(m_buf->data()+fromIndex, arg);
      }
      else
      {
         p = m_buf->data()+fromIndex;
      }

      if (p != NULL)
      {
         cc = static_cast<size_t>(p - m_buf->data());
      }
   }
   return cc;
}
size_t
String::lastIndexOf(char ch, size_t fromIndex) const
{
   if (fromIndex == npos)
   {
      if ((fromIndex = length()-1) == npos)
      {
         return npos;
      }
   }
   size_t cc = npos;
   if (fromIndex < length())
   {
      for (size_t i = fromIndex; i != npos; i--)
      {
         // Don't need to check m_buf for NULL, because if length() == 0,
         // this code won't be executed.
         if (m_buf->data()[i] == ch)
         {
            cc = i;
            break;
         }
      }
   }
   return cc;
}
size_t
String::lastIndexOf(const char* arg, size_t fromIndex) const
{
   if (fromIndex == npos || fromIndex >= length())
   {
      if (static_cast<int>(fromIndex = length()-1) < 0)
      {
         return npos;
      }
   }

   int arglen = (arg) ? ::strlen(arg) : 0;
   if (static_cast<int>(fromIndex -= arglen - 1) < 0)
   {
      return npos;
   }
   if (!arg)
   {
      return length() - 1;
   }
   while (fromIndex != npos)
   {
      // Don't need to check m_buf for NULL, because if length() == 0,
      // this code won't be executed.
      if (::strncmp(m_buf->data()+fromIndex, arg, arglen) == 0)
      {
         break;
      }
      fromIndex--;
   }
   return fromIndex;
}
bool
String::startsWith(char arg) const
{
   return (m_buf
         && m_buf->length()
         && m_buf->data()[0] == arg);
}

bool
String::startsWith(const char* arg, EIgnoreCaseFlag ignoreCase) const
{
   bool cc = false;
   if (!arg || !*arg) // treat NULL as identical to the empty string
   {
      return true; // the empty string is a prefix of any string
   }

   size_t arglen = ::strlen(arg);
   if (arglen <= length())
   {
      // Don't need to check m_buf for NULL, because if length() == 0,
      // this code won't be executed.
      if (ignoreCase == E_CASE_INSENSITIVE)
      {
         cc = (strncmpi(m_buf->data(), arg, arglen) == 0);
      }
      else
      {
         cc = (::strncmp(m_buf->data(), arg, arglen) == 0);
      }
   }
   return cc;
}
String
String::substring(size_t beginIndex, size_t len) const
{
   String nil;
   size_t count = len;
   size_t l = length();
   if (0 == l)
   {
      return nil;
   }
   if (beginIndex >= l)
   {
      return nil;
   }
   else if (0 == len)
   {
      return nil;
   }
   else if (len == npos)
   {
      count = l - beginIndex;
   }
   if (count + beginIndex > l)
   {
      count = l - beginIndex;
   }
   // Don't need to check m_buf for NULL, because if length() == 0,
   // this code won't be executed.
   return String(static_cast<const char*>(m_buf->data()+beginIndex), count);
}
bool
String::isSpaces() const
{
   if (!m_buf)
   {
      return true;
   }
   char* p = m_buf->data();
   while (isspace(*p) && *p != '\0')
   {
      p++;
   }
   return (*p == '\0');
}
String&
String::ltrim()
{
   if (!m_buf)
   {
      return *this;
   }
   char* s1 = m_buf->data();
   while (isspace(*s1) && *s1 != '\0')
   {
      s1++;
   }
   if (s1 == m_buf->data())
   {
      return *this;
   }
   *this = String(s1);
   return *this;
}
String&
String::rtrim()
{
   if (length() == 0)
   {
      return *this;
   }
   char* s1 = m_buf->data() + (length()-1);
   while (isspace(*s1) && s1 >= m_buf->data())
   {
      s1--;
   }
   if (s1 == (m_buf->data() + (length()-1)))
   {
      return *this;
   }
   if (s1 < m_buf->data())
   {
      *this = String();
      return *this;
   }
   size_t len = (s1 - m_buf->data()) + 1;
   *this = String(m_buf->data(), len);
   return *this;
}
String&
String::trim()
{
   if (length() == 0)
   {
      return *this;
   }
   char* s1 = m_buf->data();
   while (isspace(*s1) && *s1 != '\0')
   {
      s1++;
   }
   if (*s1 == '\0')
   {
      // String is all spaces
      *this = String();
      return *this;
   }
   const char* p2 = String::strchr(s1, '\0');
   const char* s2 = p2 - 1;
   while (isspace(*s2))
   {
      s2--;
   }
   if (s1 == m_buf->data() && s2 == p2)
   {
      // String has no leading or trailing spaces
      return *this;
   }
   size_t len = (s2 - s1) + 1;
   *this = String(s1, len);
   return *this;
}
String&
String::erase()
{
   m_buf = 0;
   return *this;
}
String&
String::erase(size_t idx, size_t len)
{
   if ( idx >= length() )
   {
      return *this;
   }
   if (len == npos)
   {
      *this = substring(0, idx);
   }
   else
   {
      *this = substring(0, idx) + substring(idx + len);
   }
   return *this;
}
String&
String::toLowerCase()
{
   if (m_buf)
   {
      if (!UTF8Utils::toLowerCaseInPlace(m_buf->data()))
      {
         *this = UTF8Utils::toLowerCase(m_buf->data());
      }
   }
   return *this;
}

String&
String::toUpperCase()
{
   if (m_buf)
   {
      if (!UTF8Utils::toUpperCaseInPlace(m_buf->data()))
      {
         *this = UTF8Utils::toUpperCase(m_buf->data());
      }
   }
   return *this;
}
void
String::readObject(streambuf & istrm)
{
   UInt32 len;
   BinarySerialization::readLen(istrm, len);
   AutoPtrVec<char> bfr(new char[len+1]);
   BinarySerialization::read(istrm, bfr.get(), len);
   bfr[len] = '\0';
   m_buf = new ByteBuf(bfr, len);
}
void
String::writeObject(streambuf & ostrm) const
{
   UInt32 len = static_cast<UInt32>(length());
   BinarySerialization::writeLen(ostrm, len);
   if (len)
   {
      BinarySerialization::write(ostrm, m_buf->data(), len);
   }
}
String&
String::operator= (const String& arg)
{
   m_buf = arg.m_buf;
   return *this;
}
const char*
String::c_str() const
{
   if (m_buf)
   {
      return m_buf->data();
   }
   else
   {
      return "";
   }
}
static const char cnullChar = '\0';
const char&
String::operator[] (size_t ndx) const
{
#ifdef BLOCXX_DEBUG
   BLOCXX_ASSERT(ndx <= length());
#endif
   if (!m_buf) // then length() == 0
   {
      // Only needed in case ndx == 0, but doesn't hurt if ndx > 0
      const_cast<buf_t &>(m_buf) = new ByteBuf("");
   }
   return *(ndx <= m_buf->length() ? m_buf->data() + ndx : &cnullChar);
}
static char nullChar = '\0';
char&
String::operator[] (size_t ndx)
{
#ifdef BLOCXX_DEBUG
   BLOCXX_ASSERT(ndx <= length());
#endif
   if (!m_buf) // then length() == 0
   {
      // Only needed in case ndx == 0, but doesn't hurt if ndx > 0
      m_buf = new ByteBuf("");
   }
   return (ndx <= m_buf->length() ? m_buf->data()[ndx] : nullChar);
}
String
String::toString() const
{
   return *this;
}
static inline void
throwStringConversion(const String::buf_t& m_buf, const char* type)
{
   BLOCXX_THROW(StringConversionException, Format("Unable to convert \"%1\" into %2", m_buf->data(), type).c_str());
}
static inline void
throwStringConversion(const char* str, const char* type)
{
   BLOCXX_THROW(StringConversionException, Format("Unable to convert \"%1\" into %2", str, type).c_str());
}
Char16
String::toChar16() const
{
   if (UTF8Length() != 1)
   {
      throwStringConversion(c_str(), "Char16");
   }
   return Char16(*this);
}
template <typename T, typename FP>
static inline
T convertToRealType(const String::buf_t& m_buf, const char* type, FP fp)
{
   if (m_buf && m_buf->length() > 0)
   {
      char* endptr(0);
      errno = 0;     // errno is thread local
      T rv = fp(m_buf->data(), &endptr);
      if (*endptr != '\0' || errno == ERANGE || rv == HUGE_VAL || rv == -HUGE_VAL)
      {
         throwStringConversion(m_buf, type);
      }
      return rv;
   }
   else
   {
      throwStringConversion("", type);
   }
   return T(); // to make compiler happy
}
Real32
String::toReal32() const
{
#if defined(BLOCXX_REAL32_IS_FLOAT) && defined(BLOCXX_HAVE_STRTOF)
   return convertToRealType<Real32>(m_buf, "Real32", &strtof);
#elif defined(BLOCXX_REAL32_IS_DOUBLE) || (defined(BLOCXX_REAL32_IS_FLOAT) && !defined(BLOCXX_HAVE_STRTOF))
   return convertToRealType<Real32>(m_buf, "Real32", &strtod);
#endif
}
Real64
String::toReal64() const
{
#if defined(BLOCXX_REAL64_IS_DOUBLE)
   return convertToRealType<Real64>(m_buf, "Real64", &strtod);
#elif defined(BLOCXX_REAL64_IS_LONG_DOUBLE)
   return convertToRealType<Real64>(m_buf, "Real64", &strtold);
#endif
}
bool
String::toBool() const
{
   if (equalsIgnoreCase("true"))
   {
      return true;
   }
   else if (equalsIgnoreCase("false"))
   {
      return false;
   }
   else
   {
      throwStringConversion(c_str(), "bool");
   }
   return false; // to make compiler happy
}
template <typename T, typename FP, typename FPRT>
static inline
T doConvertToIntType(const String::buf_t& m_buf, const char* type, FP fp, int base)
{
   // the error detecting code below won't detect an empty string, so
   // we have to check for it explicitly.
   if (m_buf && m_buf->length() > 0)
   {
      char* endptr(0);
      errno = 0;     // errno is thread local
      FPRT v = fp(m_buf->data(), &endptr, base);
      T rv = static_cast<T>(v);
      if (*endptr != '\0' || errno == ERANGE || FPRT(rv) != v)
      {
         throwStringConversion(m_buf, type);
      }
      return rv;
   }
   else
   {
      throwStringConversion("", type);
   }
   return T(); // to make compiler happy
}
typedef unsigned long int (*strtoulfp_t)(const char *, char **,int);
typedef long int (*strtolfp_t)(const char *, char **,int);
typedef unsigned long long int (*strtoullfp_t)(const char *, char **,int);
typedef long long int (*strtollfp_t)(const char *, char **,int);
template <typename T>
static inline
T convertToUIntType(const String::buf_t& m_buf, const char* msg, int base)
{
   return doConvertToIntType<T, strtoulfp_t, unsigned long int>(m_buf, msg, &strtoul, base);
}
template <typename T>
static inline
T convertToIntType(const String::buf_t& m_buf, const char* msg, int base)
{
   return doConvertToIntType<T, strtolfp_t, long int>(m_buf, msg, &strtol, base);
}
template <typename T>
static inline
T convertToUInt64Type(const String::buf_t& m_buf, const char* msg, int base)
{
   return doConvertToIntType<T, strtoullfp_t, unsigned long long int>(m_buf, msg, &String::strtoull, base);
}
template <typename T>
static inline
T convertToInt64Type(const String::buf_t& m_buf, const char* msg, int base)
{
   return doConvertToIntType<T, strtollfp_t, long long int>(m_buf, msg, &String::strtoll, base);
}
UInt8
String::toUInt8(int base) const
{
   return convertToUIntType<UInt8>(m_buf, "UInt8", base);
}
Int8
String::toInt8(int base) const
{
   return convertToIntType<Int8>(m_buf, "Int8", base);
}
UInt16
String::toUInt16(int base) const
{
   return convertToUIntType<UInt16>(m_buf, "UInt16", base);
}
Int16
String::toInt16(int base) const
{
   return convertToIntType<Int16>(m_buf, "Int16", base);
}
UInt32
String::toUInt32(int base) const
{
   return convertToUIntType<UInt32>(m_buf, "UInt32", base);
}
Int32
String::toInt32(int base) const
{
   return convertToIntType<Int32>(m_buf, "Int32", base);
}
UInt64
String::toUInt64(int base) const
{
   return convertToUInt64Type<UInt64>(m_buf, "UInt64", base);
}
Int64
String::toInt64(int base) const
{
   return convertToInt64Type<Int64>(m_buf, "Int64", base);
}
unsigned int
String::toUnsignedInt(int base) const
{
   return convertToUIntType<unsigned int>(m_buf, "unsigned int", base);
}
int
String::toInt(int base) const
{
   return convertToIntType<int>(m_buf, "int", base);
}
StringArray
String::tokenize(const char* delims, EReturnDelimitersFlag returnDelimitersAsTokens, EEmptyTokenReturnFlag returnEmptyTokens) const
{
   StringArray ra;
   if (empty())
   {
      return ra;
   }
   if (delims == 0)
   {
      ra.append(*this);
      return ra;
   }
   // Don't need to check m_buf for NULL, because if length() == 0,
   // this code won't be executed.
   char* pstr = m_buf->data();
   AutoPtrVec<char> data(new char[m_buf->length()+1]);
   data[0] = 0;
   int i = 0;
   bool last_was_delim = false;
   while (*pstr)
   {
      if (String::strchr(delims, *pstr))
      {
         if (data[0] != 0)
         {
            ra.append(String(data.get()));
            data[0] = 0;
         }
         if ( (returnEmptyTokens == E_RETURN_EMPTY_TOKENS) && last_was_delim )
         {
            ra.append(String());
         }
         if ( returnDelimitersAsTokens == E_RETURN_DELIMITERS || returnDelimitersAsTokens == E_RETURN_TOKENS )
         {
            ra.append(String(*pstr));
         }        
         i = 0;
         last_was_delim = true;
      }
      else
      {
         last_was_delim = false;
         data[i++] = *pstr;
         data[i] = 0;
      }
      pstr++;
   }
   if (data[0] != 0)
   {
      ra.append(String(data.get()));
   }
   return ra;
}

#ifdef BLOCXX_HAVE_STRTOLL
long long int
String::strtoll(const char* nptr, char** endptr, int base)
{
   return ::strtoll(nptr, endptr, base);
}
#else
#ifndef LLONG_MAX
#if BLOCXX_SIZEOF_LONG_LONG_INT == 8
#define LLONG_MAX 9223372036854775807LL
#else
#define LLONG_MAX 2147483647LL
#endif
#define LLONG_MIN (-LLONG_MAX - 1LL)
#endif
long long int
String::strtoll(const char* nptr, char** endptr, int base)
{
   const char *s;
   long long acc, cutoff;
   int c;
   int neg, any, cutlim;
   // Skip white space and pick up leading +/- sign if any.
   // If base is 0, allow 0x for hex and 0 for octal, else
   // assume decimal; if base is already 16, allow 0x.
   s = nptr;
   do
   {
      c = (unsigned char) *s++;
   } while (isspace(c));
   if (c == '-')
   {
      neg = 1;
      c = *s++;
   }
   else
   {
      neg = 0;
      if (c == '+')
      {
         c = *s++;
      }
   }
   if ((base == 0 || base == 16)
      && c == '0'
      && (*s == 'x' || *s == 'X'))
   {
      c = s[1];
      s += 2;
      base = 16;
   }
   if (base == 0)
   {
      base = c == '0' ? 8 : 10;
   }
   // Compute the cutoff value between legal numbers and illegal
   // numbers.  That is the largest legal value, divided by the
   // base.  An input number that is greater than this value, if
   // followed by a legal input character, is too big.  One that
   // is equal to this value may be valid or not; the limit
   // between valid and invalid numbers is then based on the last
   // digit.  For instance, if the range for longs is
   // [-2147483648..2147483647] and the input base is 10,
   // cutoff will be set to 214748364 and cutlim to either
   // 7 (neg==0) or 8 (neg==1), meaning that if we have accumulated
   // a value > 214748364, or equal but the next digit is > 7 (or 8),
   // the number is too big, and we will return a range error.
   //
   // Set any if any `digits' consumed; make it negative to indicate
   // overflow.
   cutoff = neg ? LLONG_MIN : LLONG_MAX;
   cutlim = static_cast<int>(cutoff % base);
   cutoff /= base;
   if (neg)
   {
      if (cutlim > 0)
      {
         cutlim -= base;
         cutoff += 1;
      }
      cutlim = -cutlim;
   }
   for (acc = 0, any = 0;; c = (unsigned char) *s++)
   {
      if (isdigit(c))
      {
         c -= '0';
      }
      else if (isalpha(c))
      {
         c -= isupper(c) ? 'A' - 10 : 'a' - 10;
      }
      else
      {
         break;
      }
      if (c >= base)
      {
         break;
      }
      if (any < 0)
      {
         continue;
      }
      if (neg)
      {
         if (acc < cutoff || acc == cutoff && c > cutlim)
         {
            any = -1;
            acc = LLONG_MIN;
            errno = ERANGE;
         }
         else
         {
            any = 1;
            acc *= base;
            acc -= c;
         }
      }
      else
      {
         if (acc > cutoff || acc == cutoff && c > cutlim)
         {
            any = -1;
            acc = LLONG_MAX;
            errno = ERANGE;
         }
         else
         {
            any = 1;
            acc *= base;
            acc += c;
         }
      }
   }
   if (endptr != 0)
   {
      *endptr = (char *) (any ? s - 1 : nptr);
   }
   return(acc);
}
#endif   // #ifdef BLOCXX_HAVE_STRTOLL

#ifdef BLOCXX_HAVE_STRTOULL
unsigned long long int
String::strtoull(const char* nptr, char** endptr, int base)
{
   return ::strtoull(nptr, endptr, base);
}
#else
#ifndef ULLONG_MAX
#if BLOCXX_SIZEOF_LONG_LONG_INT == 8
#define ULLONG_MAX 18446744073709551615ULL
#else
#define ULLONG_MAX 4294967295ULL
#endif
#endif
unsigned long long int
String::strtoull(const char* nptr, char** endptr, int base)
{
   const char *s;
   unsigned long long acc, cutoff, cutlim;
   unsigned int c;
   int neg, any;
   s = nptr;
   do
   {
      c = (unsigned char) *s++;
   } while (isspace(c));
   if (c == '-')
   {
      neg = 1;
      c = *s++;
   }
   else
   {
      neg = 0;
      if (c == '+')
      {
         c = *s++;
      }
   }
   if ((base == 0 || base == 16)
      && c == '0'
      && (*s == 'x' || *s == 'X'))
   {
      c = s[1];
      s += 2;
      base = 16;
   }
   if (base == 0)
   {
      base = c == '0' ? 8 : 10;
   }
   cutoff = ULLONG_MAX / (unsigned long long)base;
   cutlim = ULLONG_MAX % (unsigned long long)base;
   for (acc = 0, any = 0;; c = (unsigned char) *s++)
   {
      if (isdigit(c))
      {
         c -= '0';
      }
      else if (isalpha(c))
      {
         c -= isupper(c) ? 'A' - 10 : 'a' - 10;
      }
      else
      {
         break;
      }
      if (c >= (unsigned int)base)
      {
         break;
      }
      if (any < 0)
      {
         continue;
      }
      if (acc > cutoff || acc == cutoff && c > cutlim)
      {
         any = -1;
         acc = ULLONG_MAX;
         errno = ERANGE;
      }
      else
      {
         any = 1;
         acc *= (unsigned long)base;
         acc += c;
      }
   }
   if (neg && any > 0)
   {
#ifdef BLOCXX_WIN32
#pragma warning (push)
#pragma warning (disable: 4146)
#endif

      acc = -acc;

#ifdef BLOCXX_WIN32
#pragma warning (pop)
#endif

   }
   if (endptr != 0)
   {
      *endptr = (char *) (any ? s - 1 : nptr);
   }
   return(acc);
}
#endif   // #ifdef BLOCXX_HAVE_STRTOULL
String
operator + (const String& s1, const String& s2)
{
   String rstr(s1);
   rstr += s2;
   return rstr;
}
String
operator + (const char* p, const String& s)
{
   String rstr(p);
   rstr += s;
   return rstr;
}
String
operator + (const String& s, const char* p)
{
   String rstr(s);
   rstr += p;
   return rstr;
}
String
operator + (char c, const String& s)
{
   String rstr(c);
   rstr += s;
   return rstr;
}
String
operator + (const String& s, char c)
{
   String rstr(s);
   rstr += String(c);
   return rstr;
}
ostream&
operator<< (ostream& ostr, const String& arg)
{
   ostr.write(arg.c_str(), arg.length());
   return ostr;
}
// static
String
String::getLine(istream& is)
{
   StringBuffer rv(80);
   rv.getLine(is);
   return rv.releaseString();
}
// STATIC
const char*
String::strchr(const char* theStr, int c)
{
   const char* tmpChar = theStr;
   for (; *tmpChar && *tmpChar != c; tmpChar++)
   {
      // empty
   }
   return ((*tmpChar) == c ? tmpChar : 0);
}

} // end namespace BLOCXX_NAMESPACE