CF-1153.18.tar.gz

[apple/cf.git] / CFStringEncodings.c

/*
 * Copyright (c) 2015 Apple Inc. All rights reserved.
 *
 * @APPLE_LICENSE_HEADER_START@
 *
 * This file contains Original Code and/or Modifications of Original Code
 * as defined in and that are subject to the Apple Public Source License
 * Version 2.0 (the 'License'). You may not use this file except in
 * compliance with the License. Please obtain a copy of the License at
 * http://www.opensource.apple.com/apsl/ and read it before using this
 * file.
 *
 * The Original Code and all software distributed under the License are
 * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
 * Please see the License for the specific language governing rights and
 * limitations under the License.
 *
 * @APPLE_LICENSE_HEADER_END@
 */

/*	CFStringEncodings.c
	Copyright (c) 1999-2014, Apple Inc. All rights reserved.
	Responsibility: Aki Inoue
*/

#include "CFInternal.h"
#include <CoreFoundation/CFString.h>
#include <CoreFoundation/CFByteOrder.h>
#include <CoreFoundation/CFPriv.h>
#include <string.h>
#include <CoreFoundation/CFStringEncodingConverterExt.h>
#include <CoreFoundation/CFUniChar.h>
#include <CoreFoundation/CFUnicodeDecomposition.h>
#if (TARGET_OS_MAC && !(TARGET_OS_EMBEDDED || TARGET_OS_IPHONE)) || (TARGET_OS_EMBEDDED || TARGET_OS_IPHONE)
#include <stdlib.h>
#include <fcntl.h>
#include <pwd.h>
#include <sys/param.h>
#include <unistd.h>
#include <string.h>
#include <stdio.h>
#include <xlocale.h>
#include <CoreFoundation/CFStringDefaultEncoding.h>
#endif

static bool __CFWantsToUseASCIICompatibleConversion = false;
CF_INLINE UInt32 __CFGetASCIICompatibleFlag(void) { return __CFWantsToUseASCIICompatibleConversion; }

void _CFStringEncodingSetForceASCIICompatibility(Boolean flag) {
    __CFWantsToUseASCIICompatibleConversion = (flag ? (UInt32)true : (UInt32)false);
}

Boolean (*__CFCharToUniCharFunc)(UInt32 flags, uint8_t ch, UniChar *unicodeChar) = NULL;

// To avoid early initialization issues, we just initialize this here
// This should not be const as it is changed
CF_PRIVATE UniChar __CFCharToUniCharTable[256] = {
  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,  10,  11,  12,  13,  14,  15,
 16,  17,  18,  19,  20,  21,  22,  23,  24,  25,  26,  27,  28,  29,  30,  31,
 32,  33,  34,  35,  36,  37,  38,  39,  40,  41,  42,  43,  44,  45,  46,  47,
 48,  49,  50,  51,  52,  53,  54,  55,  56,  57,  58,  59,  60,  61,  62,  63,
 64,  65,  66,  67,  68,  69,  70,  71,  72,  73,  74,  75,  76,  77,  78,  79,
 80,  81,  82,  83,  84,  85,  86,  87,  88,  89,  90,  91,  92,  93,  94,  95,
 96,  97,  98,  99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111,
112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127,
128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143,
144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159,
160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175,
176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191,
192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207,
208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223,
224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239,
240, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 254, 255
};    

CF_PRIVATE void __CFSetCharToUniCharFunc(Boolean (*func)(UInt32 flags, UInt8 ch, UniChar *unicodeChar)) {
    if (__CFCharToUniCharFunc != func) {
        int ch;
        __CFCharToUniCharFunc = func;
        if (func) {
            for (ch = 128; ch < 256; ch++) {
                UniChar uch;
                __CFCharToUniCharTable[ch] = (__CFCharToUniCharFunc(0, ch, &uch) ? uch : 0xFFFD);
            }
        } else {	// If we have no __CFCharToUniCharFunc, assume 128..255 return the value as-is
            for (ch = 128; ch < 256; ch++) __CFCharToUniCharTable[ch] = ch;
        }
    }
}

CF_PRIVATE void __CFStrConvertBytesToUnicode(const uint8_t *bytes, UniChar *buffer, CFIndex numChars) {
    CFIndex idx;
    for (idx = 0; idx < numChars; idx++) buffer[idx] = __CFCharToUniCharTable[bytes[idx]];
}


/* The minimum length the output buffers should be in the above functions
*/
#define kCFCharConversionBufferLength 512


#define MAX_LOCAL_CHARS		(sizeof(buffer->localBuffer) / sizeof(uint8_t))
#define MAX_LOCAL_UNICHARS	(sizeof(buffer->localBuffer) / sizeof(UniChar))

/* Convert a byte stream to ASCII (7-bit!) or Unicode, with a CFVarWidthCharBuffer struct on the stack. false return indicates an error occured during the conversion. The caller needs to free the returned buffer in either ascii or unicode (indicated by isASCII), if shouldFreeChars is true. 
9/18/98 __CFStringDecodeByteStream now avoids to allocate buffer if buffer->chars is not NULL
Added useClientsMemoryPtr; if not-NULL, and the provided memory can be used as is, this is set to true
__CFStringDecodeByteStream2() is kept around for any internal clients who might be using it; it should be deprecated
!!! converterFlags is only used for the UTF8 converter at this point
*/
Boolean __CFStringDecodeByteStream2(const uint8_t *bytes, UInt32 len, CFStringEncoding encoding, Boolean alwaysUnicode, CFVarWidthCharBuffer *buffer, Boolean *useClientsMemoryPtr) {
    return __CFStringDecodeByteStream3(bytes, len, encoding, alwaysUnicode, buffer, useClientsMemoryPtr, 0);
}

enum {
    __NSNonLossyErrorMode = -1,
    __NSNonLossyASCIIMode = 0,
    __NSNonLossyBackslashMode = 1,
    __NSNonLossyHexInitialMode = __NSNonLossyBackslashMode + 1,
    __NSNonLossyHexFinalMode = __NSNonLossyHexInitialMode + 4,
    __NSNonLossyOctalInitialMode = __NSNonLossyHexFinalMode + 1,
    __NSNonLossyOctalFinalMode = __NSNonLossyHexFinalMode + 3
};

Boolean __CFStringDecodeByteStream3(const uint8_t *bytes, CFIndex len, CFStringEncoding encoding, Boolean alwaysUnicode, CFVarWidthCharBuffer *buffer, Boolean *useClientsMemoryPtr, UInt32 converterFlags) {
    CFIndex idx;
    const uint8_t *chars = (const uint8_t *)bytes;
    const uint8_t *end = chars + len;
    Boolean result = TRUE;

    if (useClientsMemoryPtr) *useClientsMemoryPtr = false;

    buffer->isASCII = !alwaysUnicode;
    buffer->shouldFreeChars = false;
    buffer->numChars = 0;

    if (0 == len) return true;

    buffer->allocator = (buffer->allocator ? buffer->allocator : __CFGetDefaultAllocator());

    if ((encoding == kCFStringEncodingUTF16) || (encoding == kCFStringEncodingUTF16BE) || (encoding == kCFStringEncodingUTF16LE)) { // UTF-16
        const UTF16Char *src = (const UTF16Char *)bytes;
        const UTF16Char *limit = src + (len / sizeof(UTF16Char)); // <rdar://problem/7854378> avoiding odd len issue
        bool swap = false;

        if (kCFStringEncodingUTF16 == encoding) {
            UTF16Char bom = ((*src == 0xFFFE) || (*src == 0xFEFF) ? *(src++) : 0);

#if __CF_BIG_ENDIAN__
            if (bom == 0xFFFE) swap = true;
#else
            if (bom != 0xFEFF) swap = true;
#endif
            if (bom) useClientsMemoryPtr = NULL;
        } else {
#if __CF_BIG_ENDIAN__
            if (kCFStringEncodingUTF16LE == encoding) swap = true;
#else
            if (kCFStringEncodingUTF16BE == encoding) swap = true;
#endif
        }

        buffer->numChars = limit - src;

        if (useClientsMemoryPtr && !swap) { // If the caller is ready to deal with no-copy situation, and the situation is possible, indicate it...
            *useClientsMemoryPtr = true;
            buffer->chars.unicode = (UniChar *)src;
            buffer->isASCII = false;
        } else {
            if (buffer->isASCII) {	// Let's see if we can reduce the Unicode down to ASCII...
                const UTF16Char *characters = src;
                UTF16Char mask = (swap ? 0x80FF : 0xFF80);
    
                while (characters < limit) {
                    if (*(characters++) & mask) {
                        buffer->isASCII = false;
                        break;
                    }
                }
            }
    
            if (buffer->isASCII) {
                uint8_t *dst;
                if (NULL == buffer->chars.ascii) { // we never reallocate when buffer is supplied
                    if (buffer->numChars > MAX_LOCAL_CHARS) {
                        buffer->chars.ascii = (UInt8 *)CFAllocatorAllocate(buffer->allocator, (buffer->numChars * sizeof(uint8_t)), 0);
			if (!buffer->chars.ascii) goto memoryErrorExit;
                        buffer->shouldFreeChars = true;
                    } else {
                        buffer->chars.ascii = (uint8_t *)buffer->localBuffer;
                    }
                }
                dst = buffer->chars.ascii;
    
                if (swap) {
                    while (src < limit) *(dst++) = (*(src++) >> 8);
                } else {
                    while (src < limit) *(dst++) = (uint8_t)*(src++);
                }
            } else {
                UTF16Char *dst;

                if (NULL == buffer->chars.unicode) { // we never reallocate when buffer is supplied
                    if (buffer->numChars > MAX_LOCAL_UNICHARS) {
                        buffer->chars.unicode = (UniChar *)CFAllocatorAllocate(buffer->allocator, (buffer->numChars * sizeof(UTF16Char)), 0);
			if (!buffer->chars.unicode) goto memoryErrorExit;
                        buffer->shouldFreeChars = true;
                    } else {
                        buffer->chars.unicode = (UTF16Char *)buffer->localBuffer;
                    }
                }
                dst = buffer->chars.unicode;

                if (swap) {
                    while (src < limit) *(dst++) = CFSwapInt16(*(src++));
                } else {
                    memmove(dst, src, buffer->numChars * sizeof(UTF16Char));
                }
            }
        }
    } else if ((encoding == kCFStringEncodingUTF32) || (encoding == kCFStringEncodingUTF32BE) || (encoding == kCFStringEncodingUTF32LE)) {
        const UTF32Char *src = (const UTF32Char *)bytes;
        const UTF32Char *limit =  src + (len / sizeof(UTF32Char)); // <rdar://problem/7854378> avoiding odd len issue
        bool swap = false;
        static bool strictUTF32 = (bool)-1;

        if ((bool)-1 == strictUTF32) strictUTF32 = (1 != 0);

        if (kCFStringEncodingUTF32 == encoding) {
            UTF32Char bom = ((*src == 0xFFFE0000) || (*src == 0x0000FEFF) ? *(src++) : 0);

#if __CF_BIG_ENDIAN__
            if (bom == 0xFFFE0000) swap = true;
#else
            if (bom != 0x0000FEFF) swap = true;
#endif
        } else {
#if __CF_BIG_ENDIAN__
            if (kCFStringEncodingUTF32LE == encoding) swap = true;
#else
            if (kCFStringEncodingUTF32BE == encoding) swap = true;
#endif
        }

        buffer->numChars = limit - src;

        {
            // Let's see if we have non-ASCII or non-BMP
            const UTF32Char *characters = src;
            UTF32Char asciiMask = (swap ? 0x80FFFFFF : 0xFFFFFF80);
            UTF32Char bmpMask = (swap ? 0x0000FFFF : 0xFFFF0000);
    
            while (characters < limit) {
                if (*characters & asciiMask) {
                    buffer->isASCII = false;
                    if (*characters & bmpMask) {
                        if (strictUTF32 && ((swap ? (UTF32Char)CFSwapInt32(*characters) : *characters) > 0x10FFFF)) return false; // outside of Unicode Scaler Value. Haven't allocated buffer, yet.
                        ++(buffer->numChars);
                    }
                }
                ++characters;
            }
        }
    
        if (buffer->isASCII) {
            uint8_t *dst;
            if (NULL == buffer->chars.ascii) { // we never reallocate when buffer is supplied
                if (buffer->numChars > MAX_LOCAL_CHARS) {
                    buffer->chars.ascii = (UInt8 *)CFAllocatorAllocate(buffer->allocator, (buffer->numChars * sizeof(uint8_t)), 0);
		    if (!buffer->chars.ascii) goto memoryErrorExit;
                    buffer->shouldFreeChars = true;
                } else {
                    buffer->chars.ascii = (uint8_t *)buffer->localBuffer;
                }
            }
            dst = buffer->chars.ascii;

            if (swap) {
                while (src < limit) *(dst++) = (*(src++) >> 24);
            } else {
                while (src < limit) *(dst++) = *(src++);
            }
        } else {
            if (NULL == buffer->chars.unicode) { // we never reallocate when buffer is supplied
                if (buffer->numChars > MAX_LOCAL_UNICHARS) {
                    buffer->chars.unicode = (UniChar *)CFAllocatorAllocate(buffer->allocator, (buffer->numChars * sizeof(UTF16Char)), 0);
		    if (!buffer->chars.unicode) goto memoryErrorExit;
                    buffer->shouldFreeChars = true;
                } else {
                    buffer->chars.unicode = (UTF16Char *)buffer->localBuffer;
                }
            }
            result = (CFUniCharFromUTF32(src, limit - src, buffer->chars.unicode, (strictUTF32 ? false : true), __CF_BIG_ENDIAN__ ? !swap : swap) ? TRUE : FALSE);
        }
    } else if (kCFStringEncodingUTF8 == encoding) {    
        if ((len >= 3) && (chars[0] == 0xef) && (chars[1] == 0xbb) && (chars[2] == 0xbf)) {	// If UTF8 BOM, skip
            chars += 3;
            len -= 3;
            if (0 == len) return true;
        }
        if (buffer->isASCII) {
            for (idx = 0; idx < len; idx++) {
                if (128 <= chars[idx]) {
                    buffer->isASCII = false;
                    break;
                }
            }
        }
        if (buffer->isASCII) {
            buffer->numChars = len;
            buffer->shouldFreeChars = !buffer->chars.ascii && (len <= MAX_LOCAL_CHARS) ? false : true;
            buffer->chars.ascii = (buffer->chars.ascii ? buffer->chars.ascii : (len <= MAX_LOCAL_CHARS) ? (uint8_t *)buffer->localBuffer : (UInt8 *)CFAllocatorAllocate(buffer->allocator, len * sizeof(uint8_t), 0));
	    if (!buffer->chars.ascii) goto memoryErrorExit;
            memmove(buffer->chars.ascii, chars, len * sizeof(uint8_t));
        } else {
            CFIndex numDone;
            static CFStringEncodingToUnicodeProc __CFFromUTF8 = NULL;
            
            if (!__CFFromUTF8) {
                const CFStringEncodingConverter *converter = CFStringEncodingGetConverter(kCFStringEncodingUTF8);
                __CFFromUTF8 = (CFStringEncodingToUnicodeProc)converter->toUnicode;
            }
            
            buffer->shouldFreeChars = !buffer->chars.unicode && (len <= MAX_LOCAL_UNICHARS) ? false : true;
            buffer->chars.unicode = (buffer->chars.unicode ? buffer->chars.unicode : (len <= MAX_LOCAL_UNICHARS) ? (UniChar *)buffer->localBuffer : (UniChar *)CFAllocatorAllocate(buffer->allocator, len * sizeof(UniChar), 0));
	    if (!buffer->chars.unicode) goto memoryErrorExit;
            buffer->numChars = 0;
            while (chars < end) {
                numDone = 0;
                chars += __CFFromUTF8(converterFlags, chars, end - chars, &(buffer->chars.unicode[buffer->numChars]), len - buffer->numChars, &numDone);
                
                if (0 == numDone) {
                    result = FALSE;
                    break;
                }
                buffer->numChars += numDone;
            }
        }
    } else if (kCFStringEncodingNonLossyASCII == encoding) {
        UTF16Char currentValue = 0;
        uint8_t character;
        int8_t mode = __NSNonLossyASCIIMode;
        
        buffer->isASCII = false;
        buffer->shouldFreeChars = !buffer->chars.unicode && (len <= MAX_LOCAL_UNICHARS) ? false : true;
        buffer->chars.unicode = (buffer->chars.unicode ? buffer->chars.unicode : (len <= MAX_LOCAL_UNICHARS) ? (UniChar *)buffer->localBuffer : (UniChar *)CFAllocatorAllocate(buffer->allocator, len * sizeof(UniChar), 0));
	if (!buffer->chars.unicode) goto memoryErrorExit;
        buffer->numChars = 0;
        
        while (chars < end) {
            character = (*chars++);
            
            switch (mode) {
                case __NSNonLossyASCIIMode:
                    if (character == '\\') {
                        mode = __NSNonLossyBackslashMode;
                    } else if (character < 0x80) {
                        currentValue = character;
                    } else {
                        mode = __NSNonLossyErrorMode;
                    }
                    break;
                    
                    case __NSNonLossyBackslashMode:
                    if ((character == 'U') || (character == 'u')) {
                        mode = __NSNonLossyHexInitialMode;
                        currentValue = 0;
                    } else if ((character >= '0') && (character <= '9')) {
                        mode = __NSNonLossyOctalInitialMode;
                        currentValue = character - '0';
                    } else if (character == '\\') {
                        mode = __NSNonLossyASCIIMode;
                        currentValue = character;
                    } else {
                        mode = __NSNonLossyErrorMode;
                    }
                    break;
                    
                    default:
                    if (mode < __NSNonLossyHexFinalMode) {
                        if ((character >= '0') && (character <= '9')) {
                            currentValue = (currentValue << 4) | (character - '0');
                            if (++mode == __NSNonLossyHexFinalMode) mode = __NSNonLossyASCIIMode;
                        } else {
                            if (character >= 'a') character -= ('a' - 'A');
                            if ((character >= 'A') && (character <= 'F')) {
                                currentValue = (currentValue << 4) | ((character - 'A') + 10);
                                if (++mode == __NSNonLossyHexFinalMode) mode = __NSNonLossyASCIIMode;
                            } else {
                                mode = __NSNonLossyErrorMode;
                            }
                        }
                    } else {
                        if ((character >= '0') && (character <= '9')) {
                            currentValue = (currentValue << 3) | (character - '0');
                            if (++mode == __NSNonLossyOctalFinalMode) mode = __NSNonLossyASCIIMode;
                        } else {
                            mode = __NSNonLossyErrorMode;
                        }
                    }
                    break;
            }
            
            if (mode == __NSNonLossyASCIIMode) {
                buffer->chars.unicode[buffer->numChars++] = currentValue;
            } else if (mode == __NSNonLossyErrorMode) {
                break;
            }
        }
        result = ((mode == __NSNonLossyASCIIMode) ? YES : NO);
    } else {
        const CFStringEncodingConverter *converter = CFStringEncodingGetConverter(encoding);
        
        if (!converter) return false;
        
        Boolean isASCIISuperset = __CFStringEncodingIsSupersetOfASCII(encoding);
        
        if (!isASCIISuperset) buffer->isASCII = false;
        
        if (buffer->isASCII) {
            for (idx = 0; idx < len; idx++) {
                if (128 <= chars[idx]) {
                    buffer->isASCII = false;
                    break;
                }
            }
        }
        
        if (converter->encodingClass == kCFStringEncodingConverterCheapEightBit) {
            if (buffer->isASCII) {
                buffer->numChars = len;
                buffer->shouldFreeChars = !buffer->chars.ascii && (len <= MAX_LOCAL_CHARS) ? false : true;
                buffer->chars.ascii = (buffer->chars.ascii ? buffer->chars.ascii : (len <= MAX_LOCAL_CHARS) ? (uint8_t *)buffer->localBuffer : (UInt8 *)CFAllocatorAllocate(buffer->allocator, len * sizeof(uint8_t), 0));
		if (!buffer->chars.ascii) goto memoryErrorExit;
                memmove(buffer->chars.ascii, chars, len * sizeof(uint8_t));
            } else {
                buffer->shouldFreeChars = !buffer->chars.unicode && (len <= MAX_LOCAL_UNICHARS) ? false : true;
                buffer->chars.unicode = (buffer->chars.unicode ? buffer->chars.unicode : (len <= MAX_LOCAL_UNICHARS) ? (UniChar *)buffer->localBuffer : (UniChar *)CFAllocatorAllocate(buffer->allocator, len * sizeof(UniChar), 0));
		if (!buffer->chars.unicode) goto memoryErrorExit;
                buffer->numChars = len;
                if (kCFStringEncodingASCII == encoding || kCFStringEncodingISOLatin1 == encoding) {
                    for (idx = 0; idx < len; idx++) buffer->chars.unicode[idx] = (UniChar)chars[idx];
                } else {
                    for (idx = 0; idx < len; idx++) {
                        if (chars[idx] < 0x80 && isASCIISuperset) {
                            buffer->chars.unicode[idx] = (UniChar)chars[idx];
                        } else if (!((CFStringEncodingCheapEightBitToUnicodeProc)converter->toUnicode)(0, chars[idx], buffer->chars.unicode + idx)) {
                            result = FALSE;
                            break;
                        }
                    }
                }
            }
        } else {
            if (buffer->isASCII) {
                buffer->numChars = len;
                buffer->shouldFreeChars = !buffer->chars.ascii && (len <= MAX_LOCAL_CHARS) ? false : true;
                buffer->chars.ascii = (buffer->chars.ascii ? buffer->chars.ascii : (len <= MAX_LOCAL_CHARS) ? (uint8_t *)buffer->localBuffer : (UInt8 *)CFAllocatorAllocate(buffer->allocator, len * sizeof(uint8_t), 0));
		if (!buffer->chars.ascii) goto memoryErrorExit;
                memmove(buffer->chars.ascii, chars, len * sizeof(uint8_t));
            } else {
                CFIndex guessedLength = CFStringEncodingCharLengthForBytes(encoding, 0, bytes, len);
                static UInt32 lossyFlag = (UInt32)-1;
                
                buffer->shouldFreeChars = !buffer->chars.unicode && (guessedLength <= MAX_LOCAL_UNICHARS) ? false : true;
                buffer->chars.unicode = (buffer->chars.unicode ? buffer->chars.unicode : (guessedLength <= MAX_LOCAL_UNICHARS) ? (UniChar *)buffer->localBuffer : (UniChar *)CFAllocatorAllocate(buffer->allocator, guessedLength * sizeof(UniChar), 0));
		if (!buffer->chars.unicode) goto memoryErrorExit;
                
                if (lossyFlag == (UInt32)-1) lossyFlag = 0;
                
                if (CFStringEncodingBytesToUnicode(encoding, lossyFlag|__CFGetASCIICompatibleFlag(), bytes, len, NULL, buffer->chars.unicode, (guessedLength > MAX_LOCAL_UNICHARS ? guessedLength : MAX_LOCAL_UNICHARS), &(buffer->numChars))) result = FALSE;
            }
        }
    }

    if (FALSE == result) {
memoryErrorExit:	// Added for <rdar://problem/6581621>, but it's not clear whether an exception would be a better option
	result = FALSE;	// In case we come here from a goto
        if (buffer->shouldFreeChars && buffer->chars.unicode) CFAllocatorDeallocate(buffer->allocator, buffer->chars.unicode);
        buffer->isASCII = !alwaysUnicode;
        buffer->shouldFreeChars = false;
        buffer->chars.ascii = NULL;
        buffer->numChars = 0;
    }
    return result;
}


/* Create a byte stream from a CFString backing. Can convert a string piece at a time
   into a fixed size buffer. Returns number of characters converted. 
   Characters that cannot be converted to the specified encoding are represented
   with the char specified by lossByte; if 0, then lossy conversion is not allowed
   and conversion stops, returning partial results.
   Pass buffer==NULL if you don't care about the converted string (but just the convertability,
   or number of bytes required, indicated by usedBufLen). 
   Does not zero-terminate. If you want to create Pascal or C string, allow one extra byte at start or end. 

   Note: This function is intended to work through CFString functions, so it should work
   with NSStrings as well as CFStrings.
*/
CFIndex __CFStringEncodeByteStream(CFStringRef string, CFIndex rangeLoc, CFIndex rangeLen, Boolean generatingExternalFile, CFStringEncoding encoding, char lossByte, uint8_t *buffer, CFIndex max, CFIndex *usedBufLen) {
    CFIndex totalBytesWritten = 0;	/* Number of written bytes */
    CFIndex numCharsProcessed = 0;	/* Number of processed chars */
    const UniChar *unichars;

    if (encoding == kCFStringEncodingUTF8 && (unichars = CFStringGetCharactersPtr(string))) {
        static CFStringEncodingToBytesProc __CFToUTF8 = NULL;

        if (!__CFToUTF8) {
            const CFStringEncodingConverter *utf8Converter = CFStringEncodingGetConverter(kCFStringEncodingUTF8);
            __CFToUTF8 = (CFStringEncodingToBytesProc)utf8Converter->toBytes;
        }
        numCharsProcessed = __CFToUTF8((generatingExternalFile ? kCFStringEncodingPrependBOM : 0), unichars + rangeLoc, rangeLen, buffer, (buffer ? max : 0), &totalBytesWritten);

    } else if (encoding == kCFStringEncodingNonLossyASCII) {
	const char *hex = "0123456789abcdef";
	UniChar ch;
	CFStringInlineBuffer buf;
	CFStringInitInlineBuffer(string, &buf, CFRangeMake(rangeLoc, rangeLen));
	while (numCharsProcessed < rangeLen) {
	    CFIndex reqLength; /* Required number of chars to encode this UniChar */
	    CFIndex cnt;
	    char tmp[6];
	    ch = CFStringGetCharacterFromInlineBuffer(&buf, numCharsProcessed);
	    if ((ch >= ' ' && ch <= '~' && ch != '\\') || (ch == '\n' || ch == '\r' || ch == '\t')) {
		reqLength = 1;
		tmp[0] = (char)ch;
	    } else {
		if (ch == '\\') {
		    tmp[1] = '\\';
		    reqLength = 2;
		} else if (ch < 256) {	/* \nnn; note that this is not NEXTSTEP encoding but a (small) UniChar */
		    tmp[1] = '0' + (ch >> 6);
		    tmp[2] = '0' + ((ch >> 3) & 7);
		    tmp[3] = '0' + (ch & 7);
		    reqLength = 4;
		} else {	/* \Unnnn */
		    tmp[1] = 'u'; // Changed to small+u in order to be aligned with Java
		    tmp[2] = hex[(ch >> 12) & 0x0f];
		    tmp[3] = hex[(ch >> 8) & 0x0f];
		    tmp[4] = hex[(ch >> 4) & 0x0f];
		    tmp[5] = hex[ch & 0x0f];
		    reqLength = 6;
		}
		tmp[0] = '\\';
	    }
            if (buffer) {
                if (totalBytesWritten + reqLength > max) break; /* Doesn't fit..
.*/
                for (cnt = 0; cnt < reqLength; cnt++) {
                    buffer[totalBytesWritten + cnt] = tmp[cnt];
                }
            }
	    totalBytesWritten += reqLength;
	    numCharsProcessed++;
	}
    } else if ((encoding == kCFStringEncodingUTF16) || (encoding == kCFStringEncodingUTF16BE) || (encoding == kCFStringEncodingUTF16LE)) {
   	CFIndex extraForBOM = (generatingExternalFile && (encoding == kCFStringEncodingUTF16) ? sizeof(UniChar) : 0);
        numCharsProcessed = rangeLen;
        if (buffer && (numCharsProcessed * (CFIndex)sizeof(UniChar) + extraForBOM > max)) {
            numCharsProcessed = (max > extraForBOM) ? ((max - extraForBOM) / sizeof(UniChar)) : 0;
        }
        totalBytesWritten = (numCharsProcessed * sizeof(UniChar)) + extraForBOM;
	if (buffer) {
	    if (extraForBOM) {	/* Generate BOM */
#if __CF_BIG_ENDIAN__
		*buffer++ = 0xfe; *buffer++ = 0xff;
#else
		*buffer++ = 0xff; *buffer++ = 0xfe;
#endif
	    }
	    CFStringGetCharacters(string, CFRangeMake(rangeLoc, numCharsProcessed), (UniChar *)buffer);
            if ((__CF_BIG_ENDIAN__ ?  kCFStringEncodingUTF16LE : kCFStringEncodingUTF16BE) == encoding) { // Need to swap
                UTF16Char *characters = (UTF16Char *)buffer;
                const UTF16Char *limit = characters + numCharsProcessed;

                while (characters < limit) {
                    *characters = CFSwapInt16(*characters);
                    ++characters;
                }
            }
	}
    } else if ((encoding == kCFStringEncodingUTF32) || (encoding == kCFStringEncodingUTF32BE) || (encoding == kCFStringEncodingUTF32LE)) {
        UTF32Char character;
        CFStringInlineBuffer buf;
        UTF32Char *characters = (UTF32Char *)buffer;

        bool swap = (encoding == (__CF_BIG_ENDIAN__ ? kCFStringEncodingUTF32LE : kCFStringEncodingUTF32BE) ? true : false);
        if (generatingExternalFile && (encoding == kCFStringEncodingUTF32)) {
            totalBytesWritten += sizeof(UTF32Char);
            if (characters) {
                if (totalBytesWritten > max) { // insufficient buffer
                    totalBytesWritten = 0;
                } else {
                    *(characters++) = 0x0000FEFF;
                }
            }
        }

        CFStringInitInlineBuffer(string, &buf, CFRangeMake(rangeLoc, rangeLen));
        while (numCharsProcessed < rangeLen) {
            character = CFStringGetCharacterFromInlineBuffer(&buf, numCharsProcessed);

            if (CFUniCharIsSurrogateHighCharacter(character)) {
                UTF16Char otherCharacter;

                if (((numCharsProcessed + 1) < rangeLen) && CFUniCharIsSurrogateLowCharacter((otherCharacter = CFStringGetCharacterFromInlineBuffer(&buf, numCharsProcessed + 1)))) {
                    character = CFUniCharGetLongCharacterForSurrogatePair(character, otherCharacter);
                } else if (lossByte) {
                    character = lossByte;
                } else {
                    break;
                }
            } else if (CFUniCharIsSurrogateLowCharacter(character)) {
                if (lossByte) {
                    character = lossByte;
                } else {
                    break;
                }
            }

            totalBytesWritten += sizeof(UTF32Char);

            if (characters) {
                if (totalBytesWritten > max) {
                    totalBytesWritten -= sizeof(UTF32Char);
                    break;
                }
                *(characters++) = (swap ? CFSwapInt32(character) : character);
            }

            numCharsProcessed += (character > 0xFFFF ? 2 : 1);
        }
    } else {
        CFIndex numChars;
        UInt32 flags;
        const unsigned char *cString = NULL;
        Boolean isASCIISuperset = __CFStringEncodingIsSupersetOfASCII(encoding);

        if (!CFStringEncodingIsValidEncoding(encoding)) return 0;

        if (!CF_IS_OBJC(CFStringGetTypeID(), string) && isASCIISuperset) { // Checking for NSString to avoid infinite recursion
            const unsigned char *ptr;
            if ((cString = (const unsigned char *)CFStringGetCStringPtr(string, __CFStringGetEightBitStringEncoding()))) {
                ptr = (cString += rangeLoc);
                if (__CFStringGetEightBitStringEncoding() == encoding) {
                    numCharsProcessed = (rangeLen < max || buffer == NULL ? rangeLen : max);
                    if (buffer) memmove(buffer, cString, numCharsProcessed);
                    if (usedBufLen) *usedBufLen = numCharsProcessed;
                    return numCharsProcessed;
                }
		
                CFIndex uninterestingTailLen = buffer ? (rangeLen - MIN(max, rangeLen)) : 0;
                while (*ptr < 0x80 && rangeLen > uninterestingTailLen) {
                    ++ptr;
                    --rangeLen;
                }
                numCharsProcessed = ptr - cString;
                if (buffer) {
                    numCharsProcessed = (numCharsProcessed < max ? numCharsProcessed : max);
                    memmove(buffer, cString, numCharsProcessed);
                    buffer += numCharsProcessed;
		    max -= numCharsProcessed;
                }
                if (!rangeLen || (buffer && (max == 0))) {
                    if (usedBufLen) *usedBufLen = numCharsProcessed;
                    return numCharsProcessed;
                }
                rangeLoc += numCharsProcessed;
                totalBytesWritten += numCharsProcessed;
            }
            if (!cString && (cString = CFStringGetPascalStringPtr(string, __CFStringGetEightBitStringEncoding()))) {
                ptr = (cString += (rangeLoc + 1));
                if (__CFStringGetEightBitStringEncoding() == encoding) {
                    numCharsProcessed = (rangeLen < max || buffer == NULL ? rangeLen : max);
                    if (buffer) memmove(buffer, cString, numCharsProcessed);
                    if (usedBufLen) *usedBufLen = numCharsProcessed;
                    return numCharsProcessed;
                }
                while (*ptr < 0x80 && rangeLen > 0) {
                    ++ptr;
                    --rangeLen;
                }
                numCharsProcessed = ptr - cString;
                if (buffer) {
                    numCharsProcessed = (numCharsProcessed < max ? numCharsProcessed : max);
                    memmove(buffer, cString, numCharsProcessed);
                    buffer += numCharsProcessed;
		    max -= numCharsProcessed;
                }
                if (!rangeLen || (buffer && (max == 0))) {
                    if (usedBufLen) *usedBufLen = numCharsProcessed;
                    return numCharsProcessed;
                }
                rangeLoc += numCharsProcessed;
                totalBytesWritten += numCharsProcessed;
            }
        }

        if (!buffer) max = 0;

        // Special case for Foundation. When lossByte == 0xFF && encoding kCFStringEncodingASCII, we do the default ASCII fallback conversion
        // Aki 11/24/04 __CFGetASCIICompatibleFlag() is called only for non-ASCII superset encodings. Otherwise, it could lead to a deadlock (see 3890536).
        flags = (lossByte ? ((unsigned char)lossByte == 0xFF && encoding == kCFStringEncodingASCII ? kCFStringEncodingAllowLossyConversion : CFStringEncodingLossyByteToMask(lossByte)) : 0) | (generatingExternalFile ? kCFStringEncodingPrependBOM : 0) | (isASCIISuperset ? 0 : __CFGetASCIICompatibleFlag());

        if (!cString && (cString = (const unsigned char *)CFStringGetCharactersPtr(string))) { // Must be Unicode string
            CFStringEncodingUnicodeToBytes(encoding, flags, (const UniChar *)cString + rangeLoc, rangeLen, &numCharsProcessed, buffer, max, &totalBytesWritten);
        } else {
            UniChar charBuf[kCFCharConversionBufferLength];
            CFIndex currentLength;
            CFIndex usedLen;
            CFIndex lastUsedLen = 0, lastNumChars = 0;
            uint32_t result;
            uint32_t streamingMask;
            uint32_t streamID = 0;
#define MAX_DECOMP_LEN (6)

            while (rangeLen > 0) {
                currentLength = (rangeLen > kCFCharConversionBufferLength ? kCFCharConversionBufferLength : rangeLen);
                CFStringGetCharacters(string, CFRangeMake(rangeLoc, currentLength), charBuf);

                // could be in the middle of surrogate pair; back up.
                if ((rangeLen > kCFCharConversionBufferLength) && CFUniCharIsSurrogateHighCharacter(charBuf[kCFCharConversionBufferLength - 1])) --currentLength;

                streamingMask = ((rangeLen > currentLength) ? kCFStringEncodingPartialInput : 0)|CFStringEncodingStreamIDToMask(streamID);

                result = CFStringEncodingUnicodeToBytes(encoding, flags|streamingMask, charBuf, currentLength, &numChars, buffer, max, &usedLen);
                streamID = CFStringEncodingStreamIDFromMask(result);
                result &= ~CFStringEncodingStreamIDMask;

                if (result != kCFStringEncodingConversionSuccess) {
                    if (kCFStringEncodingInvalidInputStream == result) {
                        CFRange composedRange;
                        // Check the tail
                        if ((rangeLen > kCFCharConversionBufferLength) && ((currentLength - numChars) < MAX_DECOMP_LEN)) {
                            composedRange = CFStringGetRangeOfComposedCharactersAtIndex(string, rangeLoc + currentLength);
                            
                            if ((composedRange.length <= MAX_DECOMP_LEN) && (composedRange.location < (rangeLoc + numChars))) {
                                result = CFStringEncodingUnicodeToBytes(encoding, flags|streamingMask, charBuf, composedRange.location - rangeLoc, &numChars, buffer, max, &usedLen);
                                streamID = CFStringEncodingStreamIDFromMask(result);
                                result &= ~CFStringEncodingStreamIDMask;
                            }
                        }
                        
                        // Check the head
                        if ((kCFStringEncodingConversionSuccess != result) && (lastNumChars > 0) && (numChars < MAX_DECOMP_LEN)) {
                            composedRange = CFStringGetRangeOfComposedCharactersAtIndex(string, rangeLoc);
                            
                            if ((composedRange.length <= MAX_DECOMP_LEN) && (composedRange.location < rangeLoc)) {
                                // Try if the composed range can be converted
                                CFStringGetCharacters(string, composedRange, charBuf);
                                
                                if (CFStringEncodingUnicodeToBytes(encoding, flags, charBuf, composedRange.length, &numChars, NULL, 0, &usedLen) == kCFStringEncodingConversionSuccess) { // OK let's try the last run
                                    CFIndex lastRangeLoc = rangeLoc - lastNumChars;
                                    
                                    currentLength = composedRange.location - lastRangeLoc;
                                    CFStringGetCharacters(string, CFRangeMake(lastRangeLoc, currentLength), charBuf);

                                    result = CFStringEncodingUnicodeToBytes(encoding, flags|streamingMask, charBuf, currentLength, &numChars, (max ? buffer - lastUsedLen : NULL), (max ? max + lastUsedLen : 0), &usedLen);
                                    streamID = CFStringEncodingStreamIDFromMask(result);
                                    result &= ~CFStringEncodingStreamIDMask;

                                    if (result == kCFStringEncodingConversionSuccess) { // OK let's try the last run
                                        // Looks good. back up
                                        totalBytesWritten -= lastUsedLen;
                                        numCharsProcessed -= lastNumChars;
                                        
                                        rangeLoc = lastRangeLoc;
                                        rangeLen += lastNumChars;
                                        
                                        if (max) {
                                            buffer -= lastUsedLen;
                                            max += lastUsedLen;
                                        }
                                    }
                                }
                            }
                        }
                    }
                    
                    if (kCFStringEncodingConversionSuccess != result) { // really failed
                        totalBytesWritten += usedLen;
                        numCharsProcessed += numChars;
                        break;
                    }
                }
                
                totalBytesWritten += usedLen;
                numCharsProcessed += numChars;

                rangeLoc += numChars;
                rangeLen -= numChars;
                if (max) {
                    buffer += usedLen;
                    max -= usedLen;
                    if (max <= 0) break;
                }
                lastUsedLen = usedLen; lastNumChars = numChars;
                flags &= ~kCFStringEncodingPrependBOM;
            }
        }
    }
    if (usedBufLen) *usedBufLen = totalBytesWritten;
    return numCharsProcessed;
}

CFStringRef CFStringCreateWithFileSystemRepresentation(CFAllocatorRef alloc, const char *buffer) {
    return CFStringCreateWithCString(alloc, buffer, CFStringFileSystemEncoding());
}

CFIndex CFStringGetMaximumSizeOfFileSystemRepresentation(CFStringRef string) {
    CFIndex len = CFStringGetLength(string);
    CFStringEncoding enc = CFStringGetFastestEncoding(string);
    switch (enc) {
	case kCFStringEncodingASCII:
	case kCFStringEncodingMacRoman:
            if (len > (LONG_MAX - 1L) / 3L) return kCFNotFound;     // Avoid wrap-around
	    return len * 3L + 1L;
	default:
            if (len > (LONG_MAX - 1L) / 9L) return kCFNotFound;     // Avoid wrap-around
	    return len * 9L + 1L;
    }
} 

Boolean CFStringGetFileSystemRepresentation(CFStringRef string, char *buffer, CFIndex maxBufLen) {
#if DEPLOYMENT_TARGET_MACOSX || DEPLOYMENT_TARGET_EMBEDDED
#define MAX_STACK_BUFFER_LEN	(255)
    const UTF16Char *characters = CFStringGetCharactersPtr(string);
    const char *origBuffer = buffer;
    const char *bufferLimit = buffer + maxBufLen;
    CFIndex length = CFStringGetLength(string);
    CFIndex usedBufLen;

    if (maxBufLen < length) return false; // Since we're using UTF-8, the byte length is never shorter than the char length. Also, it filters out 0 == maxBufLen

    if (NULL == characters) {
        UTF16Char charactersBuffer[MAX_STACK_BUFFER_LEN];
        CFRange range = CFRangeMake(0, 0);
        const char *bytes = CFStringGetCStringPtr(string, __CFStringGetEightBitStringEncoding());

        if (NULL != bytes) {
            const char *originalBytes = bytes;
            const char *bytesLimit = bytes + length;

            while ((bytes < bytesLimit) && (buffer < bufferLimit) && (0 == (*bytes & 0x80))) *(buffer++) = *(bytes++);

            range.location = bytes - originalBytes;
        }
        while ((range.location < length) && (buffer < bufferLimit)) {
            range.length = length - range.location;
            if (range.length > MAX_STACK_BUFFER_LEN) range.length = MAX_STACK_BUFFER_LEN;

            CFStringGetCharacters(string, range, charactersBuffer);
            if ((range.length == MAX_STACK_BUFFER_LEN) && CFUniCharIsSurrogateHighCharacter(charactersBuffer[MAX_STACK_BUFFER_LEN - 1])) --range.length; // Backup for a high surrogate

            if (!CFUniCharDecompose(charactersBuffer, range.length, NULL, (void *)buffer, bufferLimit - buffer, &usedBufLen, true, kCFUniCharUTF8Format, true)) return false;

            buffer += usedBufLen;
            range.location += range.length;
        }
    } else {
        if (!CFUniCharDecompose(characters, length, NULL, (void *)buffer, maxBufLen, &usedBufLen, true, kCFUniCharUTF8Format, true)) return false;
        buffer += usedBufLen;
    }

    if (buffer < bufferLimit) { // Since the filename has its own limit, this is ok for now
        *buffer = '\0';
	if (_CFExecutableLinkedOnOrAfter(CFSystemVersionLion)) {
	    while (origBuffer < buffer) if (*origBuffer++ == 0) {	// There's a zero in there. Now see if the rest are all zeroes.
		while (origBuffer < buffer) if (*origBuffer++ != 0) return false;	// Embedded NULLs should cause failure: <rdar://problem/5863219>
            }
	}
        return true;
    } else {
        return false;
    }
#else
    return CFStringGetCString(string, buffer, maxBufLen, CFStringFileSystemEncoding());
#endif
}

Boolean _CFStringGetFileSystemRepresentation(CFStringRef string, uint8_t *buffer, CFIndex maxBufLen) {
    return CFStringGetFileSystemRepresentation(string, (char *)buffer, maxBufLen);
}


#if (TARGET_OS_MAC && !(TARGET_OS_EMBEDDED || TARGET_OS_IPHONE)) || (TARGET_OS_EMBEDDED || TARGET_OS_IPHONE)

/* This function is used to obtain users' default script/region code.
   The function first looks at environment variable __kCFUserEncodingEnvVariableName, then, reads the configuration file in user's home directory.
*/
void _CFStringGetUserDefaultEncoding(UInt32 *oScriptValue, UInt32 *oRegionValue) {
    char *stringValue;
    char buffer[__kCFMaxDefaultEncodingFileLength];
    int uid = getuid();

    if ((stringValue = (char *)__CFgetenv(__kCFUserEncodingEnvVariableName)) != NULL) {
        if ((uid == strtol_l(stringValue, &stringValue, 0, NULL)) && (':' == *stringValue)) {
            ++stringValue;
        } else {
            stringValue = NULL;
        }
    }

    if ((stringValue == NULL) && ((uid > 0) || __CFgetenv("HOME"))) {
        char passwdExtraBuf[1000 + MAXPATHLEN];  // Extra memory buffer for getpwuid_r(); no clue as to how large this should be...
        struct passwd passwdBuf, *passwdp = NULL;

        switch (getpwuid_r((uid_t)uid, &passwdBuf, passwdExtraBuf, sizeof(passwdExtraBuf), &passwdp)) {
            case 0:         // Success
                break;  
            case ERANGE:    // Somehow we didn't give it enough memory; let the system handle the storage this time; but beware 5778609
                passwdp = getpwuid((uid_t)uid); 
                break;
            default:
                passwdp = NULL;
        }
        if (passwdp) {
            char filename[MAXPATHLEN + 1];

	    const char *path = NULL;
	    if (!issetugid()) {
		path = __CFgetenv("CFFIXED_USER_HOME");
	    }
	    if (!path) {
		path = passwdp->pw_dir;
	    }

            strlcpy(filename, path, sizeof(filename));
            strlcat(filename, __kCFUserEncodingFileName, sizeof(filename));

	    int no_hang_fd = __CFProphylacticAutofsAccess ? open("/dev/autofs_nowait", 0) : -1;
            int fd = open(filename, O_RDONLY, 0);
            if (fd == -1) {
                // Cannot open the file. Let's fallback to smRoman/verUS
                snprintf(filename, sizeof(filename), "0x%X:0:0", uid);
                setenv(__kCFUserEncodingEnvVariableName, filename, 1);
            } else {
                ssize_t readSize;
                readSize = read(fd, buffer, __kCFMaxDefaultEncodingFileLength - 1);
                buffer[(readSize < 0 ? 0 : readSize)] = '\0';
                close(fd);
                stringValue = buffer;

                // Well, we already have a buffer, let's reuse it
                snprintf(filename, sizeof(filename), "0x%X:%s", uid, buffer);
                setenv(__kCFUserEncodingEnvVariableName, filename, 1);
            }
	    if (-1 != no_hang_fd) close(no_hang_fd);
        }
    }

    if (stringValue) {
        *oScriptValue = strtol_l(stringValue, &stringValue, 0, NULL);
        // We force using MacRoman for Arabic/Hebrew users <rdar://problem/17633551> When changing language to Arabic and Hebrew, set the default user encoding to MacRoman, not MacArabic/MacHebrew
        if ((*oScriptValue == kCFStringEncodingMacArabic) || (*oScriptValue == kCFStringEncodingMacHebrew)) *oScriptValue = kCFStringEncodingMacRoman;
        if (*stringValue == ':') {
            if (oRegionValue) *oRegionValue = strtol_l(++stringValue, NULL, 0, NULL);
            return;
        }
    }

    // Falling back
    *oScriptValue = 0; // smRoman
    if (oRegionValue) *oRegionValue = 0; // verUS
}

void _CFStringGetInstallationEncodingAndRegion(uint32_t *encoding, uint32_t *region) {
    char buffer[__kCFMaxDefaultEncodingFileLength];
    char *stringValue = NULL;

    *encoding = 0;
    *region = 0;

    struct passwd *passwdp = getpwuid((uid_t)0);
    if (passwdp) {
	const char *path = passwdp->pw_dir;

        char filename[MAXPATHLEN + 1];
        strlcpy(filename, path, sizeof(filename));
        strlcat(filename, __kCFUserEncodingFileName, sizeof(filename));
        
	int no_hang_fd = __CFProphylacticAutofsAccess ? open("/dev/autofs_nowait", 0) : -1;
	int fd = open(filename, O_RDONLY, 0);
	if (0 <= fd) {
            ssize_t size = read(fd, buffer, __kCFMaxDefaultEncodingFileLength - 1);
            buffer[(size < 0 ? 0 : size)] = '\0';
            close(fd);
            stringValue = buffer;
        }
	if (-1 != no_hang_fd) close(no_hang_fd);
    }
    
    if (stringValue) {
        *encoding = strtol_l(stringValue, &stringValue, 0, NULL);
        // We force using MacRoman for Arabic/Hebrew users <rdar://problem/17633551> When changing language to Arabic and Hebrew, set the default user encoding to MacRoman, not MacArabic/MacHebrew
        if ((*encoding == kCFStringEncodingMacArabic) || (*encoding == kCFStringEncodingMacHebrew)) *encoding = kCFStringEncodingMacRoman;
        if (*stringValue == ':') *region = strtol_l(++stringValue, NULL, 0, NULL);
    }
}

Boolean _CFStringSaveUserDefaultEncoding(UInt32 iScriptValue, UInt32 iRegionValue) {
    Boolean success = false;
    struct passwd *passwdp = getpwuid(getuid());
    if (passwdp) {
	const char *path = passwdp->pw_dir;
	if (!issetugid()) {
	    const char *value = __CFgetenv("CFFIXED_USER_HOME");
	    if (value) path = value; // override
	}

        char filename[MAXPATHLEN + 1];
        strlcpy(filename, path, sizeof(filename));
        strlcat(filename, __kCFUserEncodingFileName, sizeof(filename));

	int no_hang_fd = __CFProphylacticAutofsAccess ? open("/dev/autofs_nowait", 0) : -1;
        (void)unlink(filename);
	int fd = open(filename, O_WRONLY|O_CREAT, 0400);
	if (0 <= fd) {
            char buffer[__kCFMaxDefaultEncodingFileLength];
            // We force using MacRoman for Arabic/Hebrew users <rdar://problem/17633551> When changing language to Arabic and Hebrew, set the default user encoding to MacRoman, not MacArabic/MacHebrew
            if ((iScriptValue == kCFStringEncodingMacArabic) || (iScriptValue == kCFStringEncodingMacHebrew)) iScriptValue = kCFStringEncodingMacRoman;
            size_t size = snprintf(buffer, __kCFMaxDefaultEncodingFileLength, "0x%X:0x%X", (unsigned int)iScriptValue, (unsigned int)iRegionValue);
	    if (size <= __kCFMaxDefaultEncodingFileLength) {
                int ret = write(fd, buffer, size);
	        if (size <= ret) success = true;
	    }
	    int save_err = errno;
            close(fd);
	    errno = save_err;
        }
	int save_err = errno;
	if (-1 != no_hang_fd) close(no_hang_fd);
	errno = save_err;
    }
    return success;
}

#endif