pegasus/src/Pegasus/Common/String.cpp - diff

Return to String.cpp CVS log

Up to [Pegasus] / pegasus / src / Pegasus / Common

Diff for /pegasus/src/Pegasus/Common/String.cpp between version 1.18 and 1.111.6.15

version 1.18, 2001/05/20 20:34:01

version 1.111.6.15, 2005/10/14 14:09:29

Line 1

//%/////////////////////////////////////////////////////////////////////////////

//%2005////////////////////////////////////////////////////////////////////////

// Company, L.P.; IBM Corp.; The Open Group; Tivoli Systems.

// IBM Corp.; EMC Corporation, The Open Group.

// IBM Corp.; EMC Corporation; VERITAS Software Corporation; The Open Group.

// EMC Corporation; VERITAS Software Corporation; The Open Group.

// Permission is hereby granted, free of charge, to any person obtaining a copy

// of this software and associated documentation files (the "Software"), to

// deal in the Software without restriction, including without limitation the

// rights to use, copy, modify, merge, publish, distribute, sublicense, and/or

// sell copies of the Software, and to permit persons to whom the Software is

// furnished to do so, subject to the following conditions:

// Permission is hereby granted, free of charge, to any person obtaining a

// THE ABOVE COPYRIGHT NOTICE AND THIS PERMISSION NOTICE SHALL BE INCLUDED IN

// copy of this software and associated documentation files (the "Software"),

// ALL COPIES OR SUBSTANTIAL PORTIONS OF THE SOFTWARE. THE SOFTWARE IS PROVIDED

// to deal in the Software without restriction, including without limitation

// "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT

// the rights to use, copy, modify, merge, publish, distribute, sublicense,

// LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR

// and/or sell copies of the Software, and to permit persons to whom the

// PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT

// Software is furnished to do so, subject to the following conditions:

// HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN

// ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION

// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR

// WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL

// THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER

// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING

// FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER

// DEALINGS IN THE SOFTWARE.

//==============================================================================

// Author: Mike Brasher (mbrasher@bmc.com)

// Author: Mike Brasher (mbrasher@austin.rr.com)

// Modified By:

// Roger Kumpf, Hewlett-Packard Company (roger_kumpf@hp.com)

// Josephine Eskaline Joyce, IBM (jojustin@in.ibm.com) for Bug#3297

// David Dillard, VERITAS Software Corp. (david.dillard@veritas.com)

// Mike Brasher (mike-brasher@austin.rr.com)

//%/////////////////////////////////////////////////////////////////////////////

#include <cassert>

#include <cctype>

#include "InternalException.h"

#include "String.h"

#include "CommonUTF.h"

#include "Exception.h"

#include "MessageLoader.h"

#include "String.h"

#include "StringRep.h"

#include <iostream>

#ifdef PEGASUS_HAS_ICU

#include <unicode/ustring.h>

#include <unicode/uchar.h>

#endif

PEGASUS_NAMESPACE_BEGIN

const String String::EMPTY;

//==============================================================================

// Compile-time macros (undefined by default).

// PEGASUS_STRING_NO_THROW -- suppresses throwing of exceptions

// PEGASUS_STRING_NO_UTF8 -- don't generate slower UTF8 code.

//==============================================================================

inline Uint32 StrLen(const char* str)

//==============================================================================

// File-scope definitions:

//==============================================================================

// Note: this table is much faster than the system toupper(). Please do not

// change.

const Uint8 _toUpperTable[256] =

{

if (!str)

0x00,0x01,0x02,0x03,0x04,0x05,0x06,0x07,

throw NullPointer();

0x08,0x09,0x0A,0x0B,0x0C,0x0D,0x0E,0x0F,

0x10,0x11,0x12,0x13,0x14,0x15,0x16,0x17,

0x18,0x19,0x1A,0x1B,0x1C,0x1D,0x1E,0x1F,

0x20,0x21,0x22,0x23,0x24,0x25,0x26,0x27,

0x28,0x29,0x2A,0x2B,0x2C,0x2D,0x2E,0x2F,

0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,

0x38,0x39,0x3A,0x3B,0x3C,0x3D,0x3E,0x3F,

0x40,0x41,0x42,0x43,0x44,0x45,0x46,0x47,

0x48,0x49,0x4A,0x4B,0x4C,0x4D,0x4E,0x4F,

0x50,0x51,0x52,0x53,0x54,0x55,0x56,0x57,

0x58,0x59,0x5A,0x5B,0x5C,0x5D,0x5E,0x5F,

0x60,0x41,0x42,0x43,0x44,0x45,0x46,0x47,

0x48,0x49,0x4A,0x4B,0x4C,0x4D,0x4E,0x4F,

0x50,0x51,0x52,0x53,0x54,0x55,0x56,0x57,

0x58,0x59,0x5A,0x7B,0x7C,0x7D,0x7E,0x7F,

0x80,0x81,0x82,0x83,0x84,0x85,0x86,0x87,

0x88,0x89,0x8A,0x8B,0x8C,0x8D,0x8E,0x8F,

0x90,0x91,0x92,0x93,0x94,0x95,0x96,0x97,

0x98,0x99,0x9A,0x9B,0x9C,0x9D,0x9E,0x9F,

0xA0,0xA1,0xA2,0xA3,0xA4,0xA5,0xA6,0xA7,

0xA8,0xA9,0xAA,0xAB,0xAC,0xAD,0xAE,0xAF,

0xB0,0xB1,0xB2,0xB3,0xB4,0xB5,0xB6,0xB7,

0xB8,0xB9,0xBA,0xBB,0xBC,0xBD,0xBE,0xBF,

0xC0,0xC1,0xC2,0xC3,0xC4,0xC5,0xC6,0xC7,

0xC8,0xC9,0xCA,0xCB,0xCC,0xCD,0xCE,0xCF,

0xD0,0xD1,0xD2,0xD3,0xD4,0xD5,0xD6,0xD7,

0xD8,0xD9,0xDA,0xDB,0xDC,0xDD,0xDE,0xDF,

0xE0,0xE1,0xE2,0xE3,0xE4,0xE5,0xE6,0xE7,

0xE8,0xE9,0xEA,0xEB,0xEC,0xED,0xEE,0xEF,

0xF0,0xF1,0xF2,0xF3,0xF4,0xF5,0xF6,0xF7,

0xF8,0xF9,0xFA,0xFB,0xFC,0xFD,0xFE,0xFF,

};

// Note: this table is much faster than the system tulower(). Please do not

// change.

const Uint8 _toLowerTable[256] =

{

0x00,0x01,0x02,0x03,0x04,0x05,0x06,0x07,

0x08,0x09,0x0A,0x0B,0x0C,0x0D,0x0E,0x0F,

0x10,0x11,0x12,0x13,0x14,0x15,0x16,0x17,

0x18,0x19,0x1A,0x1B,0x1C,0x1D,0x1E,0x1F,

0x20,0x21,0x22,0x23,0x24,0x25,0x26,0x27,

0x28,0x29,0x2A,0x2B,0x2C,0x2D,0x2E,0x2F,

0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,

0x38,0x39,0x3A,0x3B,0x3C,0x3D,0x3E,0x3F,

0x40,0x61,0x62,0x63,0x64,0x65,0x66,0x67,

0x68,0x69,0x6A,0x6B,0x6C,0x6D,0x6E,0x6F,

0x70,0x71,0x72,0x73,0x74,0x75,0x76,0x77,

0x78,0x79,0x7A,0x5B,0x5C,0x5D,0x5E,0x5F,

0x60,0x61,0x62,0x63,0x64,0x65,0x66,0x67,

0x68,0x69,0x6A,0x6B,0x6C,0x6D,0x6E,0x6F,

0x70,0x71,0x72,0x73,0x74,0x75,0x76,0x77,

0x78,0x79,0x7A,0x7B,0x7C,0x7D,0x7E,0x7F,

0x80,0x81,0x82,0x83,0x84,0x85,0x86,0x87,

0x88,0x89,0x8A,0x8B,0x8C,0x8D,0x8E,0x8F,

0x90,0x91,0x92,0x93,0x94,0x95,0x96,0x97,

0x98,0x99,0x9A,0x9B,0x9C,0x9D,0x9E,0x9F,

0xA0,0xA1,0xA2,0xA3,0xA4,0xA5,0xA6,0xA7,

0xA8,0xA9,0xAA,0xAB,0xAC,0xAD,0xAE,0xAF,

0xB0,0xB1,0xB2,0xB3,0xB4,0xB5,0xB6,0xB7,

0xB8,0xB9,0xBA,0xBB,0xBC,0xBD,0xBE,0xBF,

0xC0,0xC1,0xC2,0xC3,0xC4,0xC5,0xC6,0xC7,

0xC8,0xC9,0xCA,0xCB,0xCC,0xCD,0xCE,0xCF,

0xD0,0xD1,0xD2,0xD3,0xD4,0xD5,0xD6,0xD7,

0xD8,0xD9,0xDA,0xDB,0xDC,0xDD,0xDE,0xDF,

0xE0,0xE1,0xE2,0xE3,0xE4,0xE5,0xE6,0xE7,

0xE8,0xE9,0xEA,0xEB,0xEC,0xED,0xEE,0xEF,

0xF0,0xF1,0xF2,0xF3,0xF4,0xF5,0xF6,0xF7,

0xF8,0xF9,0xFA,0xFB,0xFC,0xFD,0xFE,0xFF,

};

// Converts 16-bit characters to upper case. This routine is faster than the

// system toupper(). Please do not change.

inline Uint16 _toUpper(Uint16 x)

{

return (x & 0xFF00) ? x : _toUpperTable[x];

}

// Converts 16-bit characters to lower case. This routine is faster than the

// system toupper(). Please do not change.

inline Uint16 _toLower(Uint16 x)

{

return (x & 0xFF00) ? x : _toLowerTable[x];

}

// Rounds x up to the nearest power of two (or just returns 8 if x < 8).

static Uint32 _roundUpToPow2(Uint32 x)

{

#ifndef PEGASUS_STRING_NO_THROW

if (x > 0x0FFFFFFF)

throw PEGASUS_STD(bad_alloc)();

#endif

if (x < 8)

return 8;

x--;

x |= (x >> 1);

x |= (x >> 2);

x |= (x >> 4);

x |= (x >> 8);

x |= (x >> 16);

x++;

return x;

}

template<class P, class Q>

static void _copy(P* p, const Q* q, size_t n)

{

// The following employs loop unrolling for efficiency. Please do not

// eliminate.

while (n >= 8)

{

p[0] = q[0];

p[1] = q[1];

p[2] = q[2];

p[3] = q[3];

p[4] = q[4];

p[5] = q[5];

p[6] = q[6];

p[7] = q[7];

p += 8;

q += 8;

n -= 8;

}

while (n >= 4)

{

p[0] = q[0];

p[1] = q[1];

p[2] = q[2];

p[3] = q[3];

p += 4;

q += 4;

n -= 4;

}

return strlen(str);

while (n--)

*p++ = *q++;

}

inline Uint32 StrLen(const Char16* str)

static Uint16* _find(const Uint16* s, size_t n, Uint16 c)

{

if (!str)

// The following employs loop unrolling for efficiency. Please do not

throw NullPointer();

// eliminate.

Uint32 n = 0;

while (n >= 4)

{

if (s[0] == c)

return (Uint16*)s;

if (s[1] == c)

return (Uint16*)&s[1];

if (s[2] == c)

return (Uint16*)&s[2];

if (s[3] == c)

return (Uint16*)&s[3];

while (*str++)

n -= 4;

n++;

s += 4;

}

return n;

if (n)

{

if (*s == c)

return (Uint16*)s;

s++;

n--;

}

String::String()

if (n)

{

_rep.append('\0');

if (*s == c)

return (Uint16*)s;

s++;

n--;

}

String::String(const String& x) : _rep(x._rep)

if (n && *s == c)

return (Uint16*)s;

// Not found!

return 0;

}

static int _compare(const Uint16* s1, const Uint16* s2)

{

while (*s1 && *s2)

{

int r = *s1++ - *s2++;

if (r)

return r;

}

if (*s2)

return -1;

else if (*s1)

return 1;

return 0;

}

String::String(const String& x, Uint32 n)

static int _compareNoUTF8(const Uint16* s1, const char* s2)

{

Uint16 c1;

Uint16 c2;

{

_rep.append('\0');

c1 = *s1++;

append(x.getData(), n);

c2 = *s2++;

if (c1 == 0)

return c1 - c2;

}

while (c1 == c2);

String::String(const Char16* x) : _rep(x, StrLen(x) + 1)

return c1 - c2;

}

static int _compare(const Uint16* s1, const Uint16* s2, size_t n)

{

// This should only be called when s1 and s2 have the same length.

while (n-- && (*s1++ - *s2++) == 0)

;

return s1[-1] - s2[-1];

}

String::String(const Char16* x, Uint32 n)

static inline void _copy(Uint16* s1, const Uint16* s2, size_t n)

{

assign(x, n);

memcpy(s1, s2, n * sizeof(Uint16));

}

String::String(const char* str)

void StringThrowOutOfBounds()

{

Uint32 n = ::strlen(str) + 1;

throw IndexOutOfBoundsException();

reserve(n);

}

while (n--)

inline void _checkNullPointer(const void* ptr)

_rep.append(*str++);

{

#ifdef PEGASUS_STRING_NO_THROW

if (!ptr)

throw NullPointer();

#endif

}

String::String(const char* str, Uint32 n_)

static void _StringThrowBadUTF8(Uint32 index)

{

Uint32 n = _min(strlen(str), n_);

MessageLoaderParms parms(

reserve(n + 1);

"Common.String.BAD_UTF8",

"The byte sequence starting at index $0 "

"is not valid UTF-8 encoding.",

index);

throw Exception(parms);

}

while (n--)

static size_t _copyFromUTF8(

_rep.append(*str++);

Uint16* dest,

const char* src,

size_t n,

size_t& utf8_error_index)

{

Uint16* p = dest;

const Uint8* q = (const Uint8*)src;

_rep.append('\0');

// Process leading 7-bit ASCII characters (to avoid UTF8 overhead later).

// Use loop-unrolling.

while (n >=8 && ((q[0]|q[1]|q[2]|q[3]|q[4]|q[5]|q[6]|q[7]) & 0x80) == 0)

{

p[0] = q[0];

p[1] = q[1];

p[2] = q[2];

p[3] = q[3];

p[4] = q[4];

p[5] = q[5];

p[6] = q[6];

p[7] = q[7];

p += 8;

q += 8;

n -= 8;

}

String& String::assign(const Char16* x)

while (n >=4 && ((q[0]|q[1]|q[2]|q[3]) & 0x80) == 0)

{

_rep.clear();

p[0] = q[0];

_rep.append(x, StrLen(x) + 1);

p[1] = q[1];

return *this;

p[2] = q[2];

p[3] = q[3];

p += 4;

q += 4;

n -= 4;

}

String& String::assign(const Char16* str, Uint32 n)

switch (n)

{

_rep.clear();

case 0:

Uint32 m = _min(StrLen(str), n);

return p - dest;

_rep.append(str, m);

case 1:

_rep.append('\0');

if (q[0] < 128)

return *this;

{

p[0] = q[0];

return p + 1 - dest;

}

break;

case 2:

if (((q[0]|q[1]) & 0x80) == 0)

{

p[0] = q[0];

p[1] = q[1];

return p + 2 - dest;

}

break;

case 3:

if (((q[0]|q[1]|q[2]) & 0x80) == 0)

{

p[0] = q[0];

p[1] = q[1];

p[2] = q[2];

return p + 3 - dest;

}

break;

}

String& String::assign(const char* x)

// Process remaining characters.

while (n)

{

_rep.clear();

// Optimize for 7-bit ASCII case.

Uint32 n = strlen(x);

_rep.reserve(n + 1);

while (n--)

if (*q < 128)

_rep.append(*x++);

{

*p++ = *q++;

n--;

}

else

{

Uint8 c = UTF_8_COUNT_TRAIL_BYTES(*q) + 1;

_rep.append('\0');

if (c > n || !isValid_U8(q, c) ||

UTF8toUTF16(&q, q + c, &p, p + n) != 0)

{

utf8_error_index = q - (const Uint8*)src;

return size_t(-1);

}

return *this;

n -= c;

}

return p - dest;

}

String& String::assign(const char* x, Uint32 n_)

// Note: dest must be at least three times src (plus an extra byte for

// terminator).

static inline size_t _copyToUTF8(char* dest, const Uint16* src, size_t n)

{

_rep.clear();

// The following employs loop unrolling for efficiency. Please do not

// eliminate.

Uint32 n = _min(strlen(x), n_);

const Uint16* q = src;

_rep.reserve(n + 1);

Uint8* p = (Uint8*)dest;

while (n--)

while (n >= 4 && q[0] < 128 && q[1] < 128 && q[2] < 128 && q[3] < 128)

_rep.append(*x++);

{

p[0] = q[0];

p[1] = q[1];

p[2] = q[2];

p[3] = q[3];

p += 4;

q += 4;

n -= 4;

}

switch (n)

{

case 0:

return p - (Uint8*)dest;

case 1:

if (q[0] < 128)

{

p[0] = q[0];

return p + 1 - (Uint8*)dest;

}

break;

case 2:

if (q[0] < 128 && q[1] < 128)

{

p[0] = q[0];

p[1] = q[1];

return p + 2 - (Uint8*)dest;

}

break;

case 3:

if (q[0] < 128 && q[1] < 128 && q[2] < 128)

{

p[0] = q[0];

p[1] = q[1];

p[2] = q[2];

return p + 3 - (Uint8*)dest;

}

break;

}

_rep.append('\0');

// If this line was reached, there must be characters greater than 128.

return *this;

UTF16toUTF8(&q, q + n, &p, p + 3 * n);

return p - (Uint8*)dest;

}

char* String::allocateCString(Uint32 extraBytes, Boolean noThrow) const

static inline size_t _convert(

Uint16* p, const char* q, size_t n, size_t& utf8_error_index)

{

Uint32 n = size() + 1;

#ifdef PEGASUS_STRING_NO_UTF8

char* str = new char[n + extraBytes];

_copy(p, q, n);

char* p = str;

return n;

const Char16* q = getData();

#else

return _copyFromUTF8(p, q, n, utf8_error_index);

#endif

}

for (Uint32 i = 0; i < n; i++)

//==============================================================================

// class CString

//==============================================================================

CString::CString(const CString& cstr) : _rep(0)

{

if (cstr._rep)

{

size_t n = strlen(cstr._rep) + 1;

_rep = (char*)operator new(n);

memcpy(_rep, cstr._rep, n);

}

CString& CString::operator=(const CString& cstr)

{

if (&cstr != this)

{

Uint16 c = *q++;

if (_rep)

*p++ = char(c);

{

operator delete(_rep);

_rep = 0;

}

if ((c & 0xff00) && !noThrow)

if (cstr._rep)

throw TruncatedCharacter();

{

size_t n = strlen(cstr._rep) + 1;

_rep = (char*)operator new(n);

memcpy(_rep, cstr._rep, n);

}

return str;

return *this;

}

void String::appendToCString(

//==============================================================================

char* str,

Uint32 length,

// class StringRep

Boolean noThrow) const

//==============================================================================

StringRep StringRep::_emptyRep;

inline StringRep* StringRep::alloc(size_t cap)

{

if (!str)

#ifndef PEGASUS_STRING_NO_THROW

throw NullPointer();

// Any string bigger than this is seriously suspect.

if (cap > 0x0FFFFFFF)

throw PEGASUS_STD(bad_alloc)();

Uint32 n = _min(size(), length);

#endif

char* p = str + strlen(str);

StringRep* rep = (StringRep*)::operator new(

const Char16* q = getData();

sizeof(StringRep) + cap * sizeof(Uint16));

rep->cap = cap;

new(&rep->refs) AtomicInt(1);

for (Uint32 i = 0; i < n; i++)

return rep;

}

static inline void _reserve(StringRep*& rep, Uint32 cap)

{

if (cap > rep->cap || rep->refs.value() != 1)

{

Uint16 c = *q++;

size_t n = _roundUpToPow2(cap);

*p++ = char(c);

StringRep* newRep = StringRep::alloc(n);

newRep->size = rep->size;

_copy(newRep->data, rep->data, rep->size + 1);

StringRep::unref(rep);

rep = newRep;

}

if ((c & 0xff00) && !noThrow)

StringRep* StringRep::create(const Uint16* data, size_t size)

throw TruncatedCharacter();

{

StringRep* rep = StringRep::alloc(size);

rep->size = size;

_copy(rep->data, data, size);

rep->data[size] = '\0';

return rep;

}

*p = '\0';

StringRep* StringRep::copyOnWrite(StringRep* rep)

{

// Return a new copy of rep. Release rep.

StringRep* newRep = StringRep::alloc(rep->size);

newRep->size = rep->size;

_copy(newRep->data, rep->data, rep->size);

newRep->data[newRep->size] = '\0';

StringRep::unref(rep);

return newRep;

}

Char16& String::operator[](Uint32 i)

StringRep* StringRep::create(const char* data, size_t size)

{

if (i > size())

StringRep* rep = StringRep::alloc(size);

ThrowOutOfBounds();

size_t utf8_error_index;

rep->size = _convert((Uint16*)rep->data, data, size, utf8_error_index);

return _rep[i];

#ifndef PEGASUS_STRING_NO_THROW

if (rep->size == size_t(-1))

{

StringRep::free(rep);

_StringThrowBadUTF8(utf8_error_index);

}

#endif

const Char16 String::operator[](Uint32 i) const

rep->data[rep->size] = '\0';

return rep;

}

Uint32 StringRep::length(const Uint16* str)

{

if (i > size())

// Note: We could unroll this but it is rarely called.

ThrowOutOfBounds();

const Uint16* end = (Uint16*)str;

while (*end++)

;

return _rep[i];

return end - str - 1;

}

String& String::append(const Char16* str, Uint32 n)

//==============================================================================

// class String

//==============================================================================

const String String::EMPTY;

String::String(const String& str, Uint32 n)

{

Uint32 m = _min(StrLen(str), n);

_checkBounds(n, str._rep->size);

_rep.reserve(_rep.size() + m);

_rep = StringRep::create(str._rep->data, n);

_rep.remove(_rep.size() - 1);

_rep.append(str, m);

_rep.append('\0');

return *this;

}

void String::remove(Uint32 pos, Uint32 size)

String::String(const Char16* str)

{

if (size == Uint32(-1))

_checkNullPointer(str);

size = this->size() - pos;

_rep = StringRep::create((Uint16*)str, StringRep::length((Uint16*)str));

}

if (pos + size > this->size())

String::String(const Char16* str, Uint32 n)

ThrowOutOfBounds();

{

_checkNullPointer(str);

_rep = StringRep::create((Uint16*)str, n);

}

if (size)

String::String(const char* str)

_rep.remove(pos, size);

{

_checkNullPointer(str);

// Set this just in case create() throws an exception.

_rep = &StringRep::_emptyRep;

_rep = StringRep::create(str, strlen(str));

}

int String::compare(const Char16* s1, const Char16* s2, Uint32 n)

String::String(const char* str, Uint32 n)

{

while (n--)

_checkNullPointer(str);

// Set this just in case create() throws an exception.

_rep = &StringRep::_emptyRep;

_rep = StringRep::create(str, n);

}

String::String(const String& s1, const String& s2)

{

int r = *s1++ - *s2++;

size_t n1 = s1._rep->size;

size_t n2 = s2._rep->size;

size_t n = n1 + n2;

_rep = StringRep::alloc(n);

_copy(_rep->data, s1._rep->data, n1);

_copy(_rep->data + n1, s2._rep->data, n2);

_rep->size = n;

_rep->data[n] = '\0';

}

if (r)

String::String(const String& s1, const char* s2)

return r;

{

_checkNullPointer(s2);

size_t n1 = s1._rep->size;

size_t n2 = strlen(s2);

_rep = StringRep::alloc(n1 + n2);

_copy(_rep->data, s1._rep->data, n1);

size_t utf8_error_index;

size_t tmp = _convert((Uint16*)_rep->data + n1, s2, n2, utf8_error_index);

#ifndef PEGASUS_STRING_NO_THROW

if (tmp == size_t(-1))

{

StringRep::free(_rep);

_rep = &StringRep::_emptyRep;

_StringThrowBadUTF8(utf8_error_index);

}

#endif

return 0;

_rep->size = n1 + tmp;

_rep->data[_rep->size] = '\0';

}

int String::compareNoCase(const char* s1, const char* s2, Uint32 n)

String::String(const char* s1, const String& s2)

{

while (n--)

_checkNullPointer(s1);

size_t n1 = strlen(s1);

size_t n2 = s2._rep->size;

_rep = StringRep::alloc(n1 + n2);

size_t utf8_error_index;

size_t tmp = _convert((Uint16*)_rep->data, s1, n1, utf8_error_index);

#ifndef PEGASUS_STRING_NO_THROW

if (tmp == size_t(-1))

{

int r = tolower(*s1++) - tolower(*s2++);

StringRep::free(_rep);

_rep = &StringRep::_emptyRep;

_StringThrowBadUTF8(utf8_error_index);

}

#endif

if (r)

_rep->size = n2 + tmp;

return r;

_copy(_rep->data + n1, s2._rep->data, n2);

_rep->data[_rep->size] = '\0';

}

return 0;

String& String::assign(const String& str)

{

if (_rep != str._rep)

{

StringRep::unref(_rep);

StringRep::ref(_rep = str._rep);

}

Boolean String::equal(const String& x, const String& y)

return *this;

}

String& String::assign(const Char16* str, Uint32 n)

{

if (x.size() != y.size())

_checkNullPointer(str);

return false;

if (n > _rep->cap || _rep->refs.value() != 1)

{

StringRep::unref(_rep);

_rep = StringRep::alloc(n);

}

_rep->size = n;

_copy(_rep->data, (Uint16*)str, n);

_rep->data[n] = '\0';

return *this;

}

String& String::assign(const char* str, Uint32 n)

{

_checkNullPointer(str);

return String::compare(x.getData(), y.getData(), x.size()) == 0;

if (n > _rep->cap || _rep->refs.value() != 1)

{

StringRep::unref(_rep);

_rep = StringRep::alloc(n);

}

Boolean String::equal(const String& x, const Char16* y)

size_t utf8_error_index;

_rep->size = _convert(_rep->data, str, n, utf8_error_index);

#ifndef PEGASUS_STRING_NO_THROW

if (_rep->size == size_t(-1))

{

if (x.size() != StrLen(y))

StringRep::free(_rep);

return false;

_rep = &StringRep::_emptyRep;

_StringThrowBadUTF8(utf8_error_index);

}

#endif

_rep->data[_rep->size] = 0;

return String::compare(x.getData(), y, x.size()) == 0;

return *this;

}

Boolean String::equal(const Char16* x, const String& y)

void String::clear()

{

if (_rep->size)

{

if (_rep->refs.value() == 1)

{

return equal(y, x);

_rep->size = 0;

_rep->data[0] = '\0';

}

else

{

StringRep::unref(_rep);

_rep = &StringRep::_emptyRep;

}

Boolean String::equal(const String& x, const char* y)

void String::reserveCapacity(Uint32 cap)

{

return equal(x, String(y));

_reserve(_rep, cap);

}

Boolean String::equal(const char* x, const String& y)

CString String::getCString() const

{

return equal(String(x), y);

// A UTF8 string can have three times as many characters as its UTF16

// counterpart, so we allocate extra memory for the worst case. In the

// best case, we may need only one third of the memory allocated. But

// downsizing the string afterwards is expensive and unecessary since

// CString objects are usually short-lived (disappearing after only a few

// instructions). CString objects are typically created on the stack as

// means to obtain a char* pointer.

#ifdef PEGASUS_STRING_NO_UTF8

char* str = (char*)operator new(_rep->size + 1);

_copy(str, _rep->data, _rep->size);

str[_rep->size] = '\0';

return CString(str);

#else

Uint32 n = 3 * _rep->size;

char* str = (char*)operator new(n + 1);

size_t size = _copyToUTF8(str, _rep->data, _rep->size);

str[size] = '\0';

return CString(str);

#endif

}

Boolean String::equalNoCase(const String& x, const String& y)

String& String::append(const Char16* str, Uint32 n)

{

if (x.size() != y.size())

_checkNullPointer(str);

return false;

const Char16* p = x.getData();

size_t oldSize = _rep->size;

const Char16* q = y.getData();

size_t newSize = oldSize + n;

_reserve(_rep, newSize);

_copy(_rep->data + oldSize, (Uint16*)str, n);

_rep->size = newSize;

_rep->data[newSize] = '\0';

Uint32 n = x.size();

return *this;

}

while (n--)

String& String::append(const String& str)

{

if (*p <= 127 && *q <= 127)

return append((Char16*)str._rep->data, str._rep->size);

}

String& String::append(const char* str, Uint32 size)

{

if (tolower(*p++) != tolower(*q++))

_checkNullPointer(str);

return false;

size_t oldSize = _rep->size;

size_t cap = oldSize + size;

_reserve(_rep, cap);

size_t utf8_error_index;

size_t tmp = _convert(

(Uint16*)_rep->data + oldSize, str, size, utf8_error_index);

#ifndef PEGASUS_STRING_NO_THROW

if (tmp == size_t(-1))

{

StringRep::free(_rep);

_rep = &StringRep::_emptyRep;

_StringThrowBadUTF8(utf8_error_index);

}

else if (*p++ != *q++)

#endif

return false;

_rep->size += tmp;

_rep->data[_rep->size] = '\0';

return *this;

}

return true;

void String::remove(Uint32 index, Uint32 n)

{

if (n == PEG_NOT_FOUND)

n = _rep->size - index;

_checkBounds(index + n, _rep->size);

if (_rep->refs.value() != 1)

_rep = StringRep::copyOnWrite(_rep);

assert(index + n <= _rep->size);

size_t rem = _rep->size - (index + n);

Uint16* data = _rep->data;

if (rem)

memmove(data + index, data + index + n, rem * sizeof(Uint16));

_rep->size -= n;

data[_rep->size] = '\0';

}

String String::subString(Uint32 pos, Uint32 length) const

String String::subString(Uint32 index, Uint32 n) const

{

if (pos < size())

// Note: this implementation is very permissive but used for

// backwards compatibility.

if (index < _rep->size)

{

if (length == Uint32(-1))

if (n == PEG_NOT_FOUND || n > _rep->size - index)

length = size() - pos;

n = _rep->size - index;

return String(getData() + pos, length);

return String((Char16*)_rep->data + index, n);

}

else

return String();

}

Uint32 String::find(Char16 c) const

{

const Char16* first = getData();

Uint16* p = (Uint16*)_find(_rep->data, _rep->size, c);

for (const Char16* p = first; *p; p++)

if (p)

{

return p - _rep->data;

if (*p == c)

return p - first;

}

return Uint32(-1);

return PEG_NOT_FOUND;

}

Uint32 String::find(const String& s) const

Uint32 String::find(Uint32 index, Char16 c) const

{

const Char16* pSubStr = s.getData();

_checkBounds(index, _rep->size);

const Char16* pStr = getData();

Uint32 subStrLen = s.size();

if (index >= _rep->size)

Uint32 strLen = size();

return PEG_NOT_FOUND;

Uint16* p = (Uint16*)_find(_rep->data + index, _rep->size - index, c);

if (p)

return p - _rep->data;

return PEG_NOT_FOUND;

}

// loop to find first char match

Uint32 StringFindAux(

Uint32 loc = 0;

const StringRep* _rep, const Char16* s, Uint32 n)

for( ; loc <= (strLen-subStrLen); loc++)

{

if (*pStr++ == *pSubStr) // match first char

_checkNullPointer(s);

const Uint16* data = _rep->data;

size_t rem = _rep->size;

while (n <= rem)

{

// point to substr 2nd char

Uint16* p = (Uint16*)_find(data, rem, s[0]);

const Char16* p = pSubStr + 1;

// Test remaining chars for equal

if (!p)

Uint32 i = 1;

break;

for (; i < subStrLen; i++)

if (*pStr++ != *p++ )

if (memcmp(p, s, n * sizeof(Uint16)) == 0)

{pStr--; break;} // break from loop

return p - _rep->data;

if (i == subStrLen)

return loc;

p++;

}

rem -= p - data;

data = p;

}

return Uint32(-1);

return PEG_NOT_FOUND;

}

// ATTN:KS 5 apr 2000 Need to add the Char16* version.

Uint32 String::find(const char* s) const

{

_checkNullPointer(s);

// Note: could optimize away creation of temporary, but this is rarely

// called.

return find(String(s));

}

Uint32 String::reverseFind(Char16 c) const

{

const Char16* first = getData();

Uint16 x = c;

const Char16* last = getData() + size();

Uint16* p = _rep->data;

Uint16* q = _rep->data + _rep->size;

while (last != first)

while (q != p)

{

if (*--last == c)

if (*--q == x)

return last - first;

return q - p;

}

return Uint32(-1);

return PEG_NOT_FOUND;

}

void String::toLower()

{

for (Char16* p = &_rep[0]; *p; p++)

#ifdef PEGASUS_HAS_ICU

if (InitializeICU::initICUSuccessful())

{

if (_rep->refs.value() != 1)

_rep = StringRep::copyOnWrite(_rep);

// This will do a locale-insensitive, but context-sensitive convert.

// Since context-sensitive casing looks at adjacent chars, this

// prevents optimizations where the us-ascii is converted before

// calling ICU.

// The string may shrink or expand after the convert.

//// First calculate size of resulting string. u_strToLower() returns

//// only the size when zero is passed as the destination size argument.

UErrorCode err = U_ZERO_ERROR;

int32_t newSize = u_strToLower(

NULL, 0, (UChar*)_rep->data, _rep->size, NULL, &err);

err = U_ZERO_ERROR;

//// Reserve enough space for the result.

if ((Uint32)newSize > _rep->cap)

_reserve(_rep, newSize);

//// Perform the conversion (overlapping buffers are allowed).

u_strToLower((UChar*)_rep->data, newSize,

(UChar*)_rep->data, _rep->size, NULL, &err);

_rep->size = newSize;

return;

}

#endif /* PEGASUS_HAS_ICU */

if (_rep->refs.value() != 1)

_rep = StringRep::copyOnWrite(_rep);

Uint16* p = _rep->data;

size_t n = _rep->size;

for (; n--; p++)

{

if (*p <= 127)

if (!(*p & 0xFF00))

*p = tolower(*p);

*p = _toLower(*p);

}

void String::translate(Char16 fromChar, Char16 toChar)

void String::toUpper()

{

for (Char16* p = &_rep[0]; *p; p++)

#ifdef PEGASUS_HAS_ICU

if (InitializeICU::initICUSuccessful())

{

if (*p == fromChar)

if (_rep->refs.value() != 1)

*p = toChar;

_rep = StringRep::copyOnWrite(_rep);

// This will do a locale-insensitive, but context-sensitive convert.

// Since context-sensitive casing looks at adjacent chars, this

// prevents optimizations where the us-ascii is converted before

// calling ICU.

// The string may shrink or expand after the convert.

//// First calculate size of resulting string. u_strToUpper() returns

//// only the size when zero is passed as the destination size argument.

UErrorCode err = U_ZERO_ERROR;

int32_t newSize = u_strToUpper(

NULL, 0, (UChar*)_rep->data, _rep->size, NULL, &err);

err = U_ZERO_ERROR;

//// Reserve enough space for the result.

if ((Uint32)newSize > _rep->cap)

_reserve(_rep, newSize);

//// Perform the conversion (overlapping buffers are allowed).

u_strToUpper((UChar*)_rep->data, newSize,

(UChar*)_rep->data, _rep->size, NULL, &err);

_rep->size = newSize;

return;

}

#endif /* PEGASUS_HAS_ICU */

if (_rep->refs.value() != 1)

_rep = StringRep::copyOnWrite(_rep);

Uint16* p = _rep->data;

size_t n = _rep->size;

for (; n--; p++)

*p = _toUpper(*p);

}

int String::compare(const String& s1, const String& s2, Uint32 n)

{

assert(n <= s1._rep->size);

assert(n <= s2._rep->size);

// Ignoring error in which n is greater than s1.size() or s2.size()

return _compare(s1._rep->data, s2._rep->data, n);

}

int String::compare(const String& s1, const String& s2)

{

return _compare(s1._rep->data, s2._rep->data);

}

int String::compare(const String& s1, const char* s2)

{

_checkNullPointer(s2);

#ifdef PEGASUS_STRING_NO_UTF8

return _compareNoUTF8(s1._rep->data, s2);

#else

// ATTN: optimize this!

return String::compare(s1, String(s2));

#endif

}

int String::compare(const Char16* s1, const Char16* s2)

int String::compareNoCase(const String& str1, const String& str2)

{

#ifdef PEGASUS_HAS_ICU

if (InitializeICU::initICUSuccessful())

{

return u_strcasecmp(

str1._rep->data, str2._rep->data, U_FOLD_CASE_DEFAULT);

}

#endif /* PEGASUS_HAS_ICU */

const Uint16* s1 = str1._rep->data;

const Uint16* s2 = str2._rep->data;

while (*s1 && *s2)

{

int r = *s1++ - *s2++;

int r = _toLower(*s1++) - _toLower(*s2++);

if (r)

return r;

Line 421

Line 1156

return 0;

}

PEGASUS_STD(ostream)& operator<<(PEGASUS_STD(ostream)& os, const String& x)

Boolean StringEqualNoCase(const String& s1, const String& s2)

{

for (Uint32 i = 0, n = x.size(); i < n; i++)

#ifdef PEGASUS_HAS_ICU

os << x[i];

return os;

return String::compareNoCase(s1, s2) == 0;

}

#else /* PEGASUS_HAS_ICU */

void String::toLower(char* str)

// The following employs loop unrolling for efficiency. Please do not

// eliminate.

Uint16* p = (Uint16*)s1.getChar16Data();

Uint16* q = (Uint16*)s2.getChar16Data();

Uint32 n = s2.size();

while (n >= 8)

{

if (((p[0] - q[0]) && (_toUpper(p[0]) - _toUpper(q[0]))) ||

((p[1] - q[1]) && (_toUpper(p[1]) - _toUpper(q[1]))) ||

((p[2] - q[2]) && (_toUpper(p[2]) - _toUpper(q[2]))) ||

((p[3] - q[3]) && (_toUpper(p[3]) - _toUpper(q[3]))) ||

((p[4] - q[4]) && (_toUpper(p[4]) - _toUpper(q[4]))) ||

((p[5] - q[5]) && (_toUpper(p[5]) - _toUpper(q[5]))) ||

((p[6] - q[6]) && (_toUpper(p[6]) - _toUpper(q[6]))) ||

((p[7] - q[7]) && (_toUpper(p[7]) - _toUpper(q[7]))))

{

while (*str)

return false;

tolower(*str++);

}

n -= 8;

p += 8;

q += 8;

}

String ToLower(const String& str)

while (n >= 4)

{

if (((p[0] - q[0]) && (_toUpper(p[0]) - _toUpper(q[0]))) ||

((p[1] - q[1]) && (_toUpper(p[1]) - _toUpper(q[1]))) ||

((p[2] - q[2]) && (_toUpper(p[2]) - _toUpper(q[2]))) ||

((p[3] - q[3]) && (_toUpper(p[3]) - _toUpper(q[3]))))

{

String tmp(str);

return false;

}

for (Uint32 i = 0, n = tmp.size(); i < n; i++)

n -= 4;

p += 4;

q += 4;

}

while (n--)

{

Char16 c = tmp[i];

if (((p[0] - q[0]) && (_toUpper(p[0]) - _toUpper(q[0]))))

return false;

if (c <= 127)

p++;

tmp[i] = tolower(c);

q++;

}

return tmp;

return true;

#endif /* PEGASUS_HAS_ICU */

}

int CompareNoCase(const char* s1, const char* s2)

Boolean String::equalNoCase(const String& s1, const char* s2)

{

while (*s1 && *s2)

_checkNullPointer(s2);

#if defined(PEGASUS_HAS_ICU)

return String::equalNoCase(s1, String(s2));

#elif defined(PEGASUS_STRING_NO_UTF8)

const Uint16* p1 = (Uint16*)s1._rep->data;

const char* p2 = s2;

size_t n = s1._rep->size;

while (n--)

{

int r = tolower(*s1++) - tolower(*s2++);

if (!*p2)

return false;

if (r)

if (_toUpper(*p1++) != _toUpperTable[int(*p2++)])

return r;

return false;

}

if (*s2)

if (*p2)

return -1;

return false;

else if (*s1)

return 1;

return 0;

return true;

#else /* PEGASUS_HAS_ICU */

// ATTN: optimize this!

return String::equalNoCase(s1, String(s2));

#endif /* PEGASUS_HAS_ICU */

}

Boolean GetLine(PEGASUS_STD(istream)& is, String& line)

Boolean String::equal(const String& s1, const String& s2)

{

line.clear();

return s1._rep->size == s2._rep->size && memcmp(s1._rep->data,

s2._rep->data, s1._rep->size * sizeof(Uint16)) == 0;

Boolean gotChar = false;

}

char c;

while (is.get(c))

Boolean String::equal(const String& s1, const char* s2)

{

gotChar = true;

#ifdef PEGASUS_STRING_NO_UTF8

if (c == '\n')

_checkNullPointer(s2);

break;

const Uint16* p = (Uint16*)s1._rep->data;

const char* q = s2;

line.append(c);

while (*p && *q)

{

if (*p++ != Uint16(*q++))

return false;

}

return gotChar;

return !(*p || *q);

#else /* PEGASUS_STRING_NO_UTF8 */

return String::equal(s1, String(s2));

#endif /* PEGASUS_STRING_NO_UTF8 */

}

String::~String()

PEGASUS_STD(ostream)& operator<<(PEGASUS_STD(ostream)& os, const String& str)

{

#if defined(PEGASUS_OS_OS400)

CString cstr = str.getCString();

const char* utf8str = cstr;

os << utf8str;

return os;

#else

#if defined(PEGASUS_HAS_ICU)

if (InitializeICU::initICUSuccessful())

{

char *buf = NULL;

const int size = str.size() * 6;

UnicodeString UniStr(

(const UChar *)str.getChar16Data(), (int32_t)str.size());

Uint32 bufsize = UniStr.extract(0,size,buf);

buf = new char[bufsize+1];

UniStr.extract(0,bufsize,buf);

os << buf;

os.flush();

delete [] buf;

return os;

}

String& String::assign(const String& x)

#endif // PEGASUS_HAS_ICU

for (Uint32 i = 0, n = str.size(); i < n; i++)

{

_rep = x._rep;

Uint16 code = str[i];

return *this;

if (code > 0 && !(code & 0xFF00))

os << char(code);

else

{

// Print in hex format:

char buffer[8];

sprintf(buffer, "\\x%04X", code);

os << buffer;

}

return os;

#endif // PEGASUS_OS_OS400

}

String& String::append(const Char16& c)

void StringAppendCharAux(StringRep*& _rep)

{

_rep.insert(_rep.size() - 1, c);

StringRep* tmp;

return *this;

if (_rep->cap)

{

tmp = StringRep::alloc(2 * _rep->cap);

tmp->size = _rep->size;

_copy(tmp->data, _rep->data, _rep->size);

}

else

{

tmp = StringRep::alloc(8);

tmp->size = 0;

}

void String::clear()

StringRep::unref(_rep);

_rep = tmp;

}

PEGASUS_NAMESPACE_END

================================================================================

String optimizations:

1. Added mechanism allowing certain functions to be inlined only when

used by internal Pegasus modules. External modules (i.e., providers)

link to a non-inline version, which allows for binary compatibility.

2. Implemented copy-on-write with atomic increment/decrement. This

yieled a 10% improvement for the 'gc' benchmark and a 11% improvment

for the 'ni1000' benchmark.

3. Employed loop unrolling in several places. For example, see:

static Uint16* _find(const Uint16* s, size_t n, Uint16 c);

4. Used the "empty-rep" optimization (described in whitepaper from the

GCC Developers Summit). This reduced default construction to a simple

pointer assignment.

inline String::String() : _rep(&_emptyRep) { }

5. Implemented Uint16 versions of toupper() and tolower() using tables.

For example:

static const char _upper[] =

{

0,1,2,...255

};

inline Uint16 _toUpper(Uint16 x)

{

_rep.clear();

return (x & 0xFF00) ? x : _upper[x];

_rep.append('\0');

}

void String::reserve(Uint32 capacity)

This outperforms the system implementation by avoiding an anding

operation.

6. Implemented char* version of the following member functions to

eliminate unecessary creation of anonymous string objects

(temporaries).

String(const String& s1, const char* s2);

String(const char* s1, const String& s2);

String& String::operator=(const char* str);

Uint32 String::find(const char* s) const;

bool String::equal(const String& s1, const char* s2);

static int String::compare(const String& s1, const char* s2);

String& String::append(const char* str);

String& String::append(const char* str, Uint32 size);

static bool String::equalNoCase(const String& s1, const char* s2);

String& operator=(const char* str)

String& String::assign(const char* str)

String& String::append(const char* str)

Boolean operator==(const String& s1, const char* s2)

Boolean operator==(const char* s1, const String& s2)

Boolean operator!=(const String& s1, const char* s2)

Boolean operator!=(const char* s1, const String& s2)

Boolean operator<(const String& s1, const char* s2)

Boolean operator<(const char* s1, const String& s2)

Boolean operator>(const String& s1, const char* s2)

Boolean operator>(const char* s1, const String& s2)

Boolean operator<=(const String& s1, const char* s2)

Boolean operator<=(const char* s1, const String& s2)

Boolean operator>=(const String& s1, const char* s2)

Boolean operator>=(const char* s1, const String& s2)

String operator+(const String& s1, const char* s2)

String operator+(const char* s1, const String& s2)

7. Optimized _roundUpToPow2(), used in rounding the capacity to the next

power of two (algorithm from the book "Hacker's Delight").

static Uint32 _roundUpToPow2(Uint32 x)

{

_rep.reserve(capacity + 1);

if (x < 8)

return 8;

x--;

x |= (x >> 1);

x |= (x >> 2);

x |= (x >> 4);

x |= (x >> 8);

x |= (x >> 16);

x++;

return x;

}

const Array<String>& EmptyStringArray()

8. Implemented "concatenating constructors" to eliminate temporaries

created by operator+(). This scheme employs the "return-value

optimization" described by Stan Lippman.

inline String operator+(const String& s1, const String& s2)

{

static Array<String> tmp;

return String(s1, s2, 0);

return tmp;

}

PEGASUS_NAMESPACE_END

9. Experimented to find the optimial initial size for a short string.

Eight seems to offer the best tradeoff between space and time.

10. Inlined all members of the Char16 class.

11. Used Uint16 internally in the String class. This showed no improvememnt

since Char16 was already fully inlined and was essentially reduced to

Uint16 in any case.

12. Implemented conditional logic (#if) allowing error checking logic to

be excluded to better performance. Examples include bounds checking

and null-pointer checking.

13. Used memcpy() and memcmp() where possible. These are implemented using

the rep family of intructions under Intel and are much faster.

14. Used loop unrolling, jump-tables, and short-circuiting to reduce UTF8

copy routine overhead.

15. Added ASCII7 form of the constructor and assign().

String s("hello world", String::ASCII7);

s.assignASCII7("hello world");

This avoids slower UTF8 processing when not needed.

================================================================================

TO-DO:

(+) [DONE] Use PEGASUS_USE_EXPERIMENTAL_INTERFACES

(+) [DONE] Submit BUG-2754 (Windows buffer limit).

(+) [DONE] Eliminate char versions of find() and append().

(+) [DONE] Remove PEGASUS_MAX_PRINTABLE_CHARACTER from Config.h

(+) [DONE] Change _next_pow_2() to _roundUpToPow2().

(+) [DONE] Change '99' to '2' in StringRep constructor (comment as well).

(+) [DONE] Comment StringRep allocation layout.

(+) [DONE] Conceal private inline functions.

(+) [DONE] Shorten inclusion of StringInline.h in String.h.

(+) [DONE] Change USE_INTERNAL_INLINE TO DISABLE_INTERNAL_INLINE or get

rid of altogether.

(+) [DONE] useCamelNotationOnAllFunctionNames.

(+) [DONE] Check for overlow condition in StringRep::alloc().

(+) [DONE] Remove tabs (used vim ":set expandtab" and ":retab").

(+) [DONE] Fix throw-related memory leak.

(+) [DONE] Look at PEP223 for coding security guidelines.

(+) [DONE] Use old AtomicInt for now (split new AtomicInt into another

bug.

(+) [DONE] Removed appendASCII() and the ASCII form of the constructor.

-----------

(+) DOC++ String.h

================================================================================

Legend:

Removed from v.1.18
changed lines
	Added in v.1.111.6.15

No CVS admin address has been configured