Use LLVM Unicode conversion. Improved extra headers parsing.

6 files changed, 997 insertions, 20 deletions

diff --git a/build-mac/mailcore2.xcodeproj/project.pbxproj b/build-mac/mailcore2.xcodeproj/project.pbxproj
index 43e39820..4fadc330 100755
--- a/build-mac/mailcore2.xcodeproj/project.pbxproj
+++ b/build-mac/mailcore2.xcodeproj/project.pbxproj
@@ -155,6 +155,8 @@
 		9EF9AB23175F406D0027FA3B /* MCOIMAPFolderStatusOperation.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = 9EF9AB16175F35C60027FA3B /* MCOIMAPFolderStatusOperation.h */; };
 		9EF9AB24175F409D0027FA3B /* MCIMAPFolderStatusOperation.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = 9EF9AB0E175F30C20027FA3B /* MCIMAPFolderStatusOperation.h */; };
 		9EF9AB25175F40C70027FA3B /* MCIMAPFolderStatusOperation.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = 9EF9AB0E175F30C20027FA3B /* MCIMAPFolderStatusOperation.h */; };
+		BD49963719FEC6DD000945BC /* ConvertUTF.c in Sources */ = {isa = PBXBuildFile; fileRef = BD49963519FEC6DD000945BC /* ConvertUTF.c */; };
+		BD49963819FEC6DD000945BC /* ConvertUTF.c in Sources */ = {isa = PBXBuildFile; fileRef = BD49963519FEC6DD000945BC /* ConvertUTF.c */; };
 		BD63713B177DFF080094121B /* MCLibetpan.cc in Sources */ = {isa = PBXBuildFile; fileRef = BD637139177DFF080094121B /* MCLibetpan.cc */; };
 		C07AD5D7FD82F8ACAB576231 /* NSError+MCO.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = C07AD44B013BB42A240B4F04 /* NSError+MCO.h */; };
 		C07AD99B2E2054C684DB8FF6 /* NSError+MCO.mm in Sources */ = {isa = PBXBuildFile; fileRef = C07ADFE43E22B38EFF23ADB5 /* NSError+MCO.mm */; };
@@ -1439,6 +1441,8 @@
 		9EF9AB10175F319A0027FA3B /* MCIMAPFolderStatusOperation.cc */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.cpp.cpp; path = MCIMAPFolderStatusOperation.cc; sourceTree = "<group>"; };
 		9EF9AB16175F35C60027FA3B /* MCOIMAPFolderStatusOperation.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = MCOIMAPFolderStatusOperation.h; sourceTree = "<group>"; };
 		9EF9AB18175F36600027FA3B /* MCOIMAPFolderStatusOperation.mm */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.cpp.objcpp; path = MCOIMAPFolderStatusOperation.mm; sourceTree = "<group>"; };
+		BD49963519FEC6DD000945BC /* ConvertUTF.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = ConvertUTF.c; sourceTree = "<group>"; };
+		BD49963619FEC6DD000945BC /* ConvertUTF.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = ConvertUTF.h; sourceTree = "<group>"; };
 		BD637139177DFF080094121B /* MCLibetpan.cc */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.cpp.cpp; path = MCLibetpan.cc; sourceTree = "<group>"; };
 		BD63713A177DFF080094121B /* MCLibetpan.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = MCLibetpan.h; sourceTree = "<group>"; };
 		C07AD057D3C8FBDC7AC95733 /* MCOIMAPSession.mm */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.cpp.objcpp; path = MCOIMAPSession.mm; sourceTree = "<group>"; };
@@ -2303,6 +2307,8 @@
 		C64EA6A1169E847800778456 /* basetypes */ = {
 			isa = PBXGroup;
 			children = (
+				BD49963519FEC6DD000945BC /* ConvertUTF.c */,
+				BD49963619FEC6DD000945BC /* ConvertUTF.h */,
 				C64EA6A5169E847800778456 /* MCArray.cc */,
 				C64EA6A6169E847800778456 /* MCArray.h */,
 				C64EA6A7169E847800778456 /* MCAssert.c */,
@@ -3091,6 +3097,7 @@
 				C64EA7FC16A2959800778456 /* MCIMAPFetchFoldersOperation.cc in Sources */,
 				C64EA80216A295E400778456 /* MCIMAPRenameFolderOperation.cc in Sources */,
 				C64EA80516A2997E00778456 /* MCIMAPDeleteFolderOperation.cc in Sources */,
+				BD49963719FEC6DD000945BC /* ConvertUTF.c in Sources */,
 				C63D315C17C9155C00A4D993 /* MCIMAPIdentity.cc in Sources */,
 				84D73742199BF963005124E5 /* MCNNTPFetchAllArticlesOperation.cc in Sources */,
 				C64EA80816A2999A00778456 /* MCIMAPCreateFolderOperation.cc in Sources */,
@@ -3331,6 +3338,7 @@
 				943F1A9E17D96C5500F0C798 /* MCIMAPConnectOperation.cc in Sources */,
 				C6BA2BC51705F4E6003F0E9E /* MCIMAPFetchFoldersOperation.cc in Sources */,
 				C6BA2BC61705F4E6003F0E9E /* MCIMAPRenameFolderOperation.cc in Sources */,
+				BD49963819FEC6DD000945BC /* ConvertUTF.c in Sources */,
 				C6BA2BC71705F4E6003F0E9E /* MCIMAPDeleteFolderOperation.cc in Sources */,
 				C63D315D17C9155C00A4D993 /* MCIMAPIdentity.cc in Sources */,
 				84D73743199BF963005124E5 /* MCNNTPFetchAllArticlesOperation.cc in Sources */,
diff --git a/src/core/abstract/MCMessageHeader.cc b/src/core/abstract/MCMessageHeader.cc
index 08b0f045..e222c785 100644
--- a/src/core/abstract/MCMessageHeader.cc
+++ b/src/core/abstract/MCMessageHeader.cc
@@ -516,7 +516,7 @@ void MessageHeader::importIMFFields(struct mailimf_fields * fields)
             String * fieldValueStr;
             
             fieldValue = field->fld_data.fld_optional_field->fld_value;
-            fieldValueStr = String::stringWithUTF8Characters(fieldValue);
+            fieldValueStr = String::stringByDecodingMIMEHeaderValue(fieldValue);
             setExtraHeader(fieldNameStr, fieldValueStr);
         }
     }
diff --git a/src/core/basetypes/ConvertUTF.c b/src/core/basetypes/ConvertUTF.c
new file mode 100644
index 00000000..f5be6e79
--- /dev/null
+++ b/src/core/basetypes/ConvertUTF.c
@@ -0,0 +1,708 @@
+/*===--- ConvertUTF.c - Universal Character Names conversions ---------------===
+ *
+ *                     The LLVM Compiler Infrastructure
+ *
+ * This file is distributed under the University of Illinois Open Source
+ * License. See LICENSE.TXT for details.
+ *
+ *===------------------------------------------------------------------------=*/
+/*
+ * Copyright 2001-2004 Unicode, Inc.
+ * 
+ * Disclaimer
+ * 
+ * This source code is provided as is by Unicode, Inc. No claims are
+ * made as to fitness for any particular purpose. No warranties of any
+ * kind are expressed or implied. The recipient agrees to determine
+ * applicability of information provided. If this file has been
+ * purchased on magnetic or optical media from Unicode, Inc., the
+ * sole remedy for any claim will be exchange of defective media
+ * within 90 days of receipt.
+ * 
+ * Limitations on Rights to Redistribute This Code
+ * 
+ * Unicode, Inc. hereby grants the right to freely use the information
+ * supplied in this file in the creation of products supporting the
+ * Unicode Standard, and to make copies of this file in any form
+ * for internal or external distribution as long as this notice
+ * remains attached.
+ */
+
+/* ---------------------------------------------------------------------
+
+    Conversions between UTF32, UTF-16, and UTF-8. Source code file.
+    Author: Mark E. Davis, 1994.
+    Rev History: Rick McGowan, fixes & updates May 2001.
+    Sept 2001: fixed const & error conditions per
+        mods suggested by S. Parent & A. Lillich.
+    June 2002: Tim Dodd added detection and handling of incomplete
+        source sequences, enhanced error detection, added casts
+        to eliminate compiler warnings.
+    July 2003: slight mods to back out aggressive FFFE detection.
+    Jan 2004: updated switches in from-UTF8 conversions.
+    Oct 2004: updated to use UNI_MAX_LEGAL_UTF32 in UTF-32 conversions.
+
+    See the header file "ConvertUTF.h" for complete documentation.
+
+------------------------------------------------------------------------ */
+
+
+#include "ConvertUTF.h"
+#ifdef CVTUTF_DEBUG
+#include <stdio.h>
+#endif
+#include <assert.h>
+
+static const int halfShift  = 10; /* used for shifting by 10 bits */
+
+static const UTF32 halfBase = 0x0010000UL;
+static const UTF32 halfMask = 0x3FFUL;
+
+#define UNI_SUR_HIGH_START  (UTF32)0xD800
+#define UNI_SUR_HIGH_END    (UTF32)0xDBFF
+#define UNI_SUR_LOW_START   (UTF32)0xDC00
+#define UNI_SUR_LOW_END     (UTF32)0xDFFF
+#define false      0
+#define true        1
+
+/* --------------------------------------------------------------------- */
+
+/*
+ * Index into the table below with the first byte of a UTF-8 sequence to
+ * get the number of trailing bytes that are supposed to follow it.
+ * Note that *legal* UTF-8 values can't have 4 or 5-bytes. The table is
+ * left as-is for anyone who may want to do such conversion, which was
+ * allowed in earlier algorithms.
+ */
+static const char trailingBytesForUTF8[256] = {
+    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
+    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
+    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
+    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
+    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
+    0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
+    1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
+    2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, 3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5
+};
+
+/*
+ * Magic values subtracted from a buffer value during UTF8 conversion.
+ * This table contains as many values as there might be trailing bytes
+ * in a UTF-8 sequence.
+ */
+static const UTF32 offsetsFromUTF8[6] = { 0x00000000UL, 0x00003080UL, 0x000E2080UL, 
+                     0x03C82080UL, 0xFA082080UL, 0x82082080UL };
+
+/*
+ * Once the bits are split out into bytes of UTF-8, this is a mask OR-ed
+ * into the first byte, depending on how many bytes follow.  There are
+ * as many entries in this table as there are UTF-8 sequence types.
+ * (I.e., one byte sequence, two byte... etc.). Remember that sequencs
+ * for *legal* UTF-8 will be 4 or fewer bytes total.
+ */
+static const UTF8 firstByteMark[7] = { 0x00, 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC };
+
+/* --------------------------------------------------------------------- */
+
+/* The interface converts a whole buffer to avoid function-call overhead.
+ * Constants have been gathered. Loops & conditionals have been removed as
+ * much as possible for efficiency, in favor of drop-through switches.
+ * (See "Note A" at the bottom of the file for equivalent code.)
+ * If your compiler supports it, the "isLegalUTF8" call can be turned
+ * into an inline function.
+ */
+
+
+/* --------------------------------------------------------------------- */
+
+ConversionResult ConvertUTF32toUTF16 (
+        const UTF32** sourceStart, const UTF32* sourceEnd, 
+        UTF16** targetStart, UTF16* targetEnd, ConversionFlags flags) {
+    ConversionResult result = conversionOK;
+    const UTF32* source = *sourceStart;
+    UTF16* target = *targetStart;
+    while (source < sourceEnd) {
+        UTF32 ch;
+        if (target >= targetEnd) {
+            result = targetExhausted; break;
+        }
+        ch = *source++;
+        if (ch <= UNI_MAX_BMP) { /* Target is a character <= 0xFFFF */
+            /* UTF-16 surrogate values are illegal in UTF-32; 0xffff or 0xfffe are both reserved values */
+            if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_LOW_END) {
+                if (flags == strictConversion) {
+                    --source; /* return to the illegal value itself */
+                    result = sourceIllegal;
+                    break;
+                } else {
+                    *target++ = UNI_REPLACEMENT_CHAR;
+                }
+            } else {
+                *target++ = (UTF16)ch; /* normal case */
+            }
+        } else if (ch > UNI_MAX_LEGAL_UTF32) {
+            if (flags == strictConversion) {
+                result = sourceIllegal;
+            } else {
+                *target++ = UNI_REPLACEMENT_CHAR;
+            }
+        } else {
+            /* target is a character in range 0xFFFF - 0x10FFFF. */
+            if (target + 1 >= targetEnd) {
+                --source; /* Back up source pointer! */
+                result = targetExhausted; break;
+            }
+            ch -= halfBase;
+            *target++ = (UTF16)((ch >> halfShift) + UNI_SUR_HIGH_START);
+            *target++ = (UTF16)((ch & halfMask) + UNI_SUR_LOW_START);
+        }
+    }
+    *sourceStart = source;
+    *targetStart = target;
+    return result;
+}
+
+/* --------------------------------------------------------------------- */
+
+ConversionResult ConvertUTF16toUTF32 (
+        const UTF16** sourceStart, const UTF16* sourceEnd, 
+        UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags) {
+    ConversionResult result = conversionOK;
+    const UTF16* source = *sourceStart;
+    UTF32* target = *targetStart;
+    UTF32 ch, ch2;
+    while (source < sourceEnd) {
+        const UTF16* oldSource = source; /*  In case we have to back up because of target overflow. */
+        ch = *source++;
+        /* If we have a surrogate pair, convert to UTF32 first. */
+        if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_HIGH_END) {
+            /* If the 16 bits following the high surrogate are in the source buffer... */
+            if (source < sourceEnd) {
+                ch2 = *source;
+                /* If it's a low surrogate, convert to UTF32. */
+                if (ch2 >= UNI_SUR_LOW_START && ch2 <= UNI_SUR_LOW_END) {
+                    ch = ((ch - UNI_SUR_HIGH_START) << halfShift)
+                        + (ch2 - UNI_SUR_LOW_START) + halfBase;
+                    ++source;
+                } else if (flags == strictConversion) { /* it's an unpaired high surrogate */
+                    --source; /* return to the illegal value itself */
+                    result = sourceIllegal;
+                    break;
+                }
+            } else { /* We don't have the 16 bits following the high surrogate. */
+                --source; /* return to the high surrogate */
+                result = sourceExhausted;
+                break;
+            }
+        } else if (flags == strictConversion) {
+            /* UTF-16 surrogate values are illegal in UTF-32 */
+            if (ch >= UNI_SUR_LOW_START && ch <= UNI_SUR_LOW_END) {
+                --source; /* return to the illegal value itself */
+                result = sourceIllegal;
+                break;
+            }
+        }
+        if (target >= targetEnd) {
+            source = oldSource; /* Back up source pointer! */
+            result = targetExhausted; break;
+        }
+        *target++ = ch;
+    }
+    *sourceStart = source;
+    *targetStart = target;
+#ifdef CVTUTF_DEBUG
+if (result == sourceIllegal) {
+    fprintf(stderr, "ConvertUTF16toUTF32 illegal seq 0x%04x,%04x\n", ch, ch2);
+    fflush(stderr);
+}
+#endif
+    return result;
+}
+ConversionResult ConvertUTF16toUTF8 (
+        const UTF16** sourceStart, const UTF16* sourceEnd, 
+        UTF8** targetStart, UTF8* targetEnd, ConversionFlags flags) {
+    ConversionResult result = conversionOK;
+    const UTF16* source = *sourceStart;
+    UTF8* target = *targetStart;
+    while (source < sourceEnd) {
+        UTF32 ch;
+        unsigned short bytesToWrite = 0;
+        const UTF32 byteMask = 0xBF;
+        const UTF32 byteMark = 0x80; 
+        const UTF16* oldSource = source; /* In case we have to back up because of target overflow. */
+        ch = *source++;
+        /* If we have a surrogate pair, convert to UTF32 first. */
+        if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_HIGH_END) {
+            /* If the 16 bits following the high surrogate are in the source buffer... */
+            if (source < sourceEnd) {
+                UTF32 ch2 = *source;
+                /* If it's a low surrogate, convert to UTF32. */
+                if (ch2 >= UNI_SUR_LOW_START && ch2 <= UNI_SUR_LOW_END) {
+                    ch = ((ch - UNI_SUR_HIGH_START) << halfShift)
+                        + (ch2 - UNI_SUR_LOW_START) + halfBase;
+                    ++source;
+                } else if (flags == strictConversion) { /* it's an unpaired high surrogate */
+                    --source; /* return to the illegal value itself */
+                    result = sourceIllegal;
+                    break;
+                }
+            } else { /* We don't have the 16 bits following the high surrogate. */
+                --source; /* return to the high surrogate */
+                result = sourceExhausted;
+                break;
+            }
+        } else if (flags == strictConversion) {
+            /* UTF-16 surrogate values are illegal in UTF-32 */
+            if (ch >= UNI_SUR_LOW_START && ch <= UNI_SUR_LOW_END) {
+                --source; /* return to the illegal value itself */
+                result = sourceIllegal;
+                break;
+            }
+        }
+        /* Figure out how many bytes the result will require */
+        if (ch < (UTF32)0x80) {      bytesToWrite = 1;
+        } else if (ch < (UTF32)0x800) {     bytesToWrite = 2;
+        } else if (ch < (UTF32)0x10000) {   bytesToWrite = 3;
+        } else if (ch < (UTF32)0x110000) {  bytesToWrite = 4;
+        } else {                            bytesToWrite = 3;
+                                            ch = UNI_REPLACEMENT_CHAR;
+        }
+
+        target += bytesToWrite;
+        if (target > targetEnd) {
+            source = oldSource; /* Back up source pointer! */
+            target -= bytesToWrite; result = targetExhausted; break;
+        }
+        switch (bytesToWrite) { /* note: everything falls through. */
+            case 4: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
+            case 3: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
+            case 2: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
+            case 1: *--target =  (UTF8)(ch | firstByteMark[bytesToWrite]);
+        }
+        target += bytesToWrite;
+    }
+    *sourceStart = source;
+    *targetStart = target;
+    return result;
+}
+
+/* --------------------------------------------------------------------- */
+
+ConversionResult ConvertUTF32toUTF8 (
+        const UTF32** sourceStart, const UTF32* sourceEnd, 
+        UTF8** targetStart, UTF8* targetEnd, ConversionFlags flags) {
+    ConversionResult result = conversionOK;
+    const UTF32* source = *sourceStart;
+    UTF8* target = *targetStart;
+    while (source < sourceEnd) {
+        UTF32 ch;
+        unsigned short bytesToWrite = 0;
+        const UTF32 byteMask = 0xBF;
+        const UTF32 byteMark = 0x80; 
+        ch = *source++;
+        if (flags == strictConversion ) {
+            /* UTF-16 surrogate values are illegal in UTF-32 */
+            if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_LOW_END) {
+                --source; /* return to the illegal value itself */
+                result = sourceIllegal;
+                break;
+            }
+        }
+        /*
+         * Figure out how many bytes the result will require. Turn any
+         * illegally large UTF32 things (> Plane 17) into replacement chars.
+         */
+        if (ch < (UTF32)0x80) {      bytesToWrite = 1;
+        } else if (ch < (UTF32)0x800) {     bytesToWrite = 2;
+        } else if (ch < (UTF32)0x10000) {   bytesToWrite = 3;
+        } else if (ch <= UNI_MAX_LEGAL_UTF32) {  bytesToWrite = 4;
+        } else {                            bytesToWrite = 3;
+                                            ch = UNI_REPLACEMENT_CHAR;
+                                            result = sourceIllegal;
+        }
+        
+        target += bytesToWrite;
+        if (target > targetEnd) {
+            --source; /* Back up source pointer! */
+            target -= bytesToWrite; result = targetExhausted; break;
+        }
+        switch (bytesToWrite) { /* note: everything falls through. */
+            case 4: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
+            case 3: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
+            case 2: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
+            case 1: *--target = (UTF8) (ch | firstByteMark[bytesToWrite]);
+        }
+        target += bytesToWrite;
+    }
+    *sourceStart = source;
+    *targetStart = target;
+    return result;
+}
+
+/* --------------------------------------------------------------------- */
+
+/*
+ * Utility routine to tell whether a sequence of bytes is legal UTF-8.
+ * This must be called with the length pre-determined by the first byte.
+ * If not calling this from ConvertUTF8to*, then the length can be set by:
+ *  length = trailingBytesForUTF8[*source]+1;
+ * and the sequence is illegal right away if there aren't that many bytes
+ * available.
+ * If presented with a length > 4, this returns false.  The Unicode
+ * definition of UTF-8 goes up to 4-byte sequences.
+ */
+
+static Boolean isLegalUTF8(const UTF8 *source, int length) {
+    UTF8 a;
+    const UTF8 *srcptr = source+length;
+    switch (length) {
+    default: return false;
+        /* Everything else falls through when "true"... */
+    case 4: if ((a = (*--srcptr)) < 0x80 || a > 0xBF) return false;
+    case 3: if ((a = (*--srcptr)) < 0x80 || a > 0xBF) return false;
+    case 2: if ((a = (*--srcptr)) < 0x80 || a > 0xBF) return false;
+
+        switch (*source) {
+            /* no fall-through in this inner switch */
+            case 0xE0: if (a < 0xA0) return false; break;
+            case 0xED: if (a > 0x9F) return false; break;
+            case 0xF0: if (a < 0x90) return false; break;
+            case 0xF4: if (a > 0x8F) return false; break;
+            default:   if (a < 0x80) return false;
+        }
+
+    case 1: if (*source >= 0x80 && *source < 0xC2) return false;
+    }
+    if (*source > 0xF4) return false;
+    return true;
+}
+
+/* --------------------------------------------------------------------- */
+
+/*
+ * Exported function to return whether a UTF-8 sequence is legal or not.
+ * This is not used here; it's just exported.
+ */
+Boolean isLegalUTF8Sequence(const UTF8 *source, const UTF8 *sourceEnd) {
+    int length = trailingBytesForUTF8[*source]+1;
+    if (length > sourceEnd - source) {
+        return false;
+    }
+    return isLegalUTF8(source, length);
+}
+
+/* --------------------------------------------------------------------- */
+
+static unsigned
+findMaximalSubpartOfIllFormedUTF8Sequence(const UTF8 *source,
+                                          const UTF8 *sourceEnd) {
+  UTF8 b1, b2, b3;
+
+  assert(!isLegalUTF8Sequence(source, sourceEnd));
+
+  /*
+   * Unicode 6.3.0, D93b:
+   *
+   *   Maximal subpart of an ill-formed subsequence: The longest code unit
+   *   subsequence starting at an unconvertible offset that is either:
+   *   a. the initial subsequence of a well-formed code unit sequence, or
+   *   b. a subsequence of length one.
+   */
+
+  if (source == sourceEnd)
+    return 0;
+
+  /*
+   * Perform case analysis.  See Unicode 6.3.0, Table 3-7. Well-Formed UTF-8
+   * Byte Sequences.
+   */
+
+  b1 = *source;
+  ++source;
+  if (b1 >= 0xC2 && b1 <= 0xDF) {
+    /*
+     * First byte is valid, but we know that this code unit sequence is
+     * invalid, so the maximal subpart has to end after the first byte.
+     */
+    return 1;
+  }
+
+  if (source == sourceEnd)
+    return 1;
+
+  b2 = *source;
+  ++source;
+
+  if (b1 == 0xE0) {
+    return (b2 >= 0xA0 && b2 <= 0xBF) ? 2 : 1;
+  }
+  if (b1 >= 0xE1 && b1 <= 0xEC) {
+    return (b2 >= 0x80 && b2 <= 0xBF) ? 2 : 1;
+  }
+  if (b1 == 0xED) {
+    return (b2 >= 0x80 && b2 <= 0x9F) ? 2 : 1;
+  }
+  if (b1 >= 0xEE && b1 <= 0xEF) {
+    return (b2 >= 0x80 && b2 <= 0xBF) ? 2 : 1;
+  }
+  if (b1 == 0xF0) {
+    if (b2 >= 0x90 && b2 <= 0xBF) {
+      if (source == sourceEnd)
+        return 2;
+
+      b3 = *source;
+      return (b3 >= 0x80 && b3 <= 0xBF) ? 3 : 2;
+    }
+    return 1;
+  }
+  if (b1 >= 0xF1 && b1 <= 0xF3) {
+    if (b2 >= 0x80 && b2 <= 0xBF) {
+      if (source == sourceEnd)
+        return 2;
+
+      b3 = *source;
+      return (b3 >= 0x80 && b3 <= 0xBF) ? 3 : 2;
+    }
+    return 1;
+  }
+  if (b1 == 0xF4) {
+    if (b2 >= 0x80 && b2 <= 0x8F) {
+      if (source == sourceEnd)
+        return 2;
+
+      b3 = *source;
+      return (b3 >= 0x80 && b3 <= 0xBF) ? 3 : 2;
+    }
+    return 1;
+  }
+
+  assert((b1 >= 0x80 && b1 <= 0xC1) || b1 >= 0xF5);
+  /*
+   * There are no valid sequences that start with these bytes.  Maximal subpart
+   * is defined to have length 1 in these cases.
+   */
+  return 1;
+}
+
+/* --------------------------------------------------------------------- */
+
+/*
+ * Exported function to return the total number of bytes in a codepoint
+ * represented in UTF-8, given the value of the first byte.
+ */
+unsigned getNumBytesForUTF8(UTF8 first) {
+  return trailingBytesForUTF8[first] + 1;
+}
+
+/* --------------------------------------------------------------------- */
+
+/*
+ * Exported function to return whether a UTF-8 string is legal or not.
+ * This is not used here; it's just exported.
+ */
+Boolean isLegalUTF8String(const UTF8 **source, const UTF8 *sourceEnd) {
+    while (*source != sourceEnd) {
+        int length = trailingBytesForUTF8[**source] + 1;
+        if (length > sourceEnd - *source || !isLegalUTF8(*source, length))
+            return false;
+        *source += length;
+    }
+    return true;
+}
+
+/* --------------------------------------------------------------------- */
+
+ConversionResult ConvertUTF8toUTF16 (
+        const UTF8** sourceStart, const UTF8* sourceEnd, 
+        UTF16** targetStart, UTF16* targetEnd, ConversionFlags flags) {
+    ConversionResult result = conversionOK;
+    const UTF8* source = *sourceStart;
+    UTF16* target = *targetStart;
+    while (source < sourceEnd) {
+        UTF32 ch = 0;
+        unsigned short extraBytesToRead = trailingBytesForUTF8[*source];
+        if (extraBytesToRead >= sourceEnd - source) {
+            result = sourceExhausted; break;
+        }
+        /* Do this check whether lenient or strict */
+        if (!isLegalUTF8(source, extraBytesToRead+1)) {
+            result = sourceIllegal;
+            break;
+        }
+        /*
+         * The cases all fall through. See "Note A" below.
+         */
+        switch (extraBytesToRead) {
+            case 5: ch += *source++; ch <<= 6; /* remember, illegal UTF-8 */
+            case 4: ch += *source++; ch <<= 6; /* remember, illegal UTF-8 */
+            case 3: ch += *source++; ch <<= 6;
+            case 2: ch += *source++; ch <<= 6;
+            case 1: ch += *source++; ch <<= 6;
+            case 0: ch += *source++;
+        }
+        ch -= offsetsFromUTF8[extraBytesToRead];
+
+        if (target >= targetEnd) {
+            source -= (extraBytesToRead+1); /* Back up source pointer! */
+            result = targetExhausted; break;
+        }
+        if (ch <= UNI_MAX_BMP) { /* Target is a character <= 0xFFFF */
+            /* UTF-16 surrogate values are illegal in UTF-32 */
+            if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_LOW_END) {
+                if (flags == strictConversion) {
+                    source -= (extraBytesToRead+1); /* return to the illegal value itself */
+                    result = sourceIllegal;
+                    break;
+                } else {
+                    *target++ = UNI_REPLACEMENT_CHAR;
+                }
+            } else {
+                *target++ = (UTF16)ch; /* normal case */
+            }
+        } else if (ch > UNI_MAX_UTF16) {
+            if (flags == strictConversion) {
+                result = sourceIllegal;
+                source -= (extraBytesToRead+1); /* return to the start */
+                break; /* Bail out; shouldn't continue */
+            } else {
+                *target++ = UNI_REPLACEMENT_CHAR;
+            }
+        } else {
+            /* target is a character in range 0xFFFF - 0x10FFFF. */
+            if (target + 1 >= targetEnd) {
+                source -= (extraBytesToRead+1); /* Back up source pointer! */
+                result = targetExhausted; break;
+            }
+            ch -= halfBase;
+            *target++ = (UTF16)((ch >> halfShift) + UNI_SUR_HIGH_START);
+            *target++ = (UTF16)((ch & halfMask) + UNI_SUR_LOW_START);
+        }
+    }
+    *sourceStart = source;
+    *targetStart = target;
+    return result;
+}
+
+/* --------------------------------------------------------------------- */
+
+static ConversionResult ConvertUTF8toUTF32Impl(
+        const UTF8** sourceStart, const UTF8* sourceEnd, 
+        UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags,
+        Boolean InputIsPartial) {
+    ConversionResult result = conversionOK;
+    const UTF8* source = *sourceStart;
+    UTF32* target = *targetStart;
+    while (source < sourceEnd) {
+        UTF32 ch = 0;
+        unsigned short extraBytesToRead = trailingBytesForUTF8[*source];
+        if (extraBytesToRead >= sourceEnd - source) {
+            if (flags == strictConversion || InputIsPartial) {
+                result = sourceExhausted;
+                break;
+            } else {
+                result = sourceIllegal;
+
+                /*
+                 * Replace the maximal subpart of ill-formed sequence with
+                 * replacement character.
+                 */
+                source += findMaximalSubpartOfIllFormedUTF8Sequence(source,
+                                                                    sourceEnd);
+                *target++ = UNI_REPLACEMENT_CHAR;
+                continue;
+            }
+        }
+        if (target >= targetEnd) {
+            result = targetExhausted; break;
+        }
+
+        /* Do this check whether lenient or strict */
+        if (!isLegalUTF8(source, extraBytesToRead+1)) {
+            result = sourceIllegal;
+            if (flags == strictConversion) {
+                /* Abort conversion. */
+                break;
+            } else {
+                /*
+                 * Replace the maximal subpart of ill-formed sequence with
+                 * replacement character.
+                 */
+                source += findMaximalSubpartOfIllFormedUTF8Sequence(source,
+                                                                    sourceEnd);
+                *target++ = UNI_REPLACEMENT_CHAR;
+                continue;
+            }
+        }
+        /*
+         * The cases all fall through. See "Note A" below.
+         */
+        switch (extraBytesToRead) {
+            case 5: ch += *source++; ch <<= 6;
+            case 4: ch += *source++; ch <<= 6;
+            case 3: ch += *source++; ch <<= 6;
+            case 2: ch += *source++; ch <<= 6;
+            case 1: ch += *source++; ch <<= 6;
+            case 0: ch += *source++;
+        }
+        ch -= offsetsFromUTF8[extraBytesToRead];
+
+        if (ch <= UNI_MAX_LEGAL_UTF32) {
+            /*
+             * UTF-16 surrogate values are illegal in UTF-32, and anything
+             * over Plane 17 (> 0x10FFFF) is illegal.
+             */
+            if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_LOW_END) {
+                if (flags == strictConversion) {
+                    source -= (extraBytesToRead+1); /* return to the illegal value itself */
+                    result = sourceIllegal;
+                    break;
+                } else {
+                    *target++ = UNI_REPLACEMENT_CHAR;
+                }
+            } else {
+                *target++ = ch;
+            }
+        } else { /* i.e., ch > UNI_MAX_LEGAL_UTF32 */
+            result = sourceIllegal;
+            *target++ = UNI_REPLACEMENT_CHAR;
+        }
+    }
+    *sourceStart = source;
+    *targetStart = target;
+    return result;
+}
+
+ConversionResult ConvertUTF8toUTF32Partial(const UTF8 **sourceStart,
+                                           const UTF8 *sourceEnd,
+                                           UTF32 **targetStart,
+                                           UTF32 *targetEnd,
+                                           ConversionFlags flags) {
+  return ConvertUTF8toUTF32Impl(sourceStart, sourceEnd, targetStart, targetEnd,
+                                flags, /*InputIsPartial=*/true);
+}
+
+ConversionResult ConvertUTF8toUTF32(const UTF8 **sourceStart,
+                                    const UTF8 *sourceEnd, UTF32 **targetStart,
+                                    UTF32 *targetEnd, ConversionFlags flags) {
+  return ConvertUTF8toUTF32Impl(sourceStart, sourceEnd, targetStart, targetEnd,
+                                flags, /*InputIsPartial=*/false);
+}
+
+/* ---------------------------------------------------------------------
+
+    Note A.
+    The fall-through switches in UTF-8 reading code save a
+    temp variable, some decrements & conditionals.  The switches
+    are equivalent to the following loop:
+        {
+            int tmpBytesToRead = extraBytesToRead+1;
+            do {
+                ch += *source++;
+                --tmpBytesToRead;
+                if (tmpBytesToRead) ch <<= 6;
+            } while (tmpBytesToRead > 0);
+        }
+    In UTF-8 writing code, the switches on "bytesToWrite" are
+    similarly unrolled loops.
+
+   --------------------------------------------------------------------- */
diff --git a/src/core/basetypes/ConvertUTF.h b/src/core/basetypes/ConvertUTF.h
new file mode 100644
index 00000000..2ea4d2c3
--- /dev/null
+++ b/src/core/basetypes/ConvertUTF.h
@@ -0,0 +1,262 @@
+/*===--- ConvertUTF.h - Universal Character Names conversions ---------------===
+ *
+ *                     The LLVM Compiler Infrastructure
+ *
+ * This file is distributed under the University of Illinois Open Source
+ * License. See LICENSE.TXT for details.
+ *
+ *==------------------------------------------------------------------------==*/
+/*
+ * Copyright 2001-2004 Unicode, Inc.
+ *
+ * Disclaimer
+ *
+ * This source code is provided as is by Unicode, Inc. No claims are
+ * made as to fitness for any particular purpose. No warranties of any
+ * kind are expressed or implied. The recipient agrees to determine
+ * applicability of information provided. If this file has been
+ * purchased on magnetic or optical media from Unicode, Inc., the
+ * sole remedy for any claim will be exchange of defective media
+ * within 90 days of receipt.
+ *
+ * Limitations on Rights to Redistribute This Code
+ *
+ * Unicode, Inc. hereby grants the right to freely use the information
+ * supplied in this file in the creation of products supporting the
+ * Unicode Standard, and to make copies of this file in any form
+ * for internal or external distribution as long as this notice
+ * remains attached.
+ */
+
+/* ---------------------------------------------------------------------
+
+    Conversions between UTF32, UTF-16, and UTF-8.  Header file.
+
+    Several funtions are included here, forming a complete set of
+    conversions between the three formats.  UTF-7 is not included
+    here, but is handled in a separate source file.
+
+    Each of these routines takes pointers to input buffers and output
+    buffers.  The input buffers are const.
+
+    Each routine converts the text between *sourceStart and sourceEnd,
+    putting the result into the buffer between *targetStart and
+    targetEnd. Note: the end pointers are *after* the last item: e.g.
+    *(sourceEnd - 1) is the last item.
+
+    The return result indicates whether the conversion was successful,
+    and if not, whether the problem was in the source or target buffers.
+    (Only the first encountered problem is indicated.)
+
+    After the conversion, *sourceStart and *targetStart are both
+    updated to point to the end of last text successfully converted in
+    the respective buffers.
+
+    Input parameters:
+        sourceStart - pointer to a pointer to the source buffer.
+                The contents of this are modified on return so that
+                it points at the next thing to be converted.
+        targetStart - similarly, pointer to pointer to the target buffer.
+        sourceEnd, targetEnd - respectively pointers to the ends of the
+                two buffers, for overflow checking only.
+
+    These conversion functions take a ConversionFlags argument. When this
+    flag is set to strict, both irregular sequences and isolated surrogates
+    will cause an error.  When the flag is set to lenient, both irregular
+    sequences and isolated surrogates are converted.
+
+    Whether the flag is strict or lenient, all illegal sequences will cause
+    an error return. This includes sequences such as: <F4 90 80 80>, <C0 80>,
+    or <A0> in UTF-8, and values above 0x10FFFF in UTF-32. Conformant code
+    must check for illegal sequences.
+
+    When the flag is set to lenient, characters over 0x10FFFF are converted
+    to the replacement character; otherwise (when the flag is set to strict)
+    they constitute an error.
+
+    Output parameters:
+        The value "sourceIllegal" is returned from some routines if the input
+        sequence is malformed.  When "sourceIllegal" is returned, the source
+        value will point to the illegal value that caused the problem. E.g.,
+        in UTF-8 when a sequence is malformed, it points to the start of the
+        malformed sequence.
+
+    Author: Mark E. Davis, 1994.
+    Rev History: Rick McGowan, fixes & updates May 2001.
+         Fixes & updates, Sept 2001.
+
+------------------------------------------------------------------------ */
+
+#ifndef LLVM_SUPPORT_CONVERTUTF_H
+#define LLVM_SUPPORT_CONVERTUTF_H
+
+/* ---------------------------------------------------------------------
+    The following 4 definitions are compiler-specific.
+    The C standard does not guarantee that wchar_t has at least
+    16 bits, so wchar_t is no less portable than unsigned short!
+    All should be unsigned values to avoid sign extension during
+    bit mask & shift operations.
+------------------------------------------------------------------------ */
+
+typedef unsigned int    UTF32;  /* at least 32 bits */
+typedef unsigned short  UTF16;  /* at least 16 bits */
+typedef unsigned char   UTF8;   /* typically 8 bits */
+typedef unsigned char   Boolean; /* 0 or 1 */
+
+/* Some fundamental constants */
+#define UNI_REPLACEMENT_CHAR (UTF32)0x0000FFFD
+#define UNI_MAX_BMP (UTF32)0x0000FFFF
+#define UNI_MAX_UTF16 (UTF32)0x0010FFFF
+#define UNI_MAX_UTF32 (UTF32)0x7FFFFFFF
+#define UNI_MAX_LEGAL_UTF32 (UTF32)0x0010FFFF
+
+#define UNI_MAX_UTF8_BYTES_PER_CODE_POINT 4
+
+#define UNI_UTF16_BYTE_ORDER_MARK_NATIVE  0xFEFF
+#define UNI_UTF16_BYTE_ORDER_MARK_SWAPPED 0xFFFE
+
+typedef enum {
+  conversionOK,           /* conversion successful */
+  sourceExhausted,        /* partial character in source, but hit end */
+  targetExhausted,        /* insuff. room in target for conversion */
+  sourceIllegal           /* source sequence is illegal/malformed */
+} ConversionResult;
+
+typedef enum {
+  strictConversion = 0,
+  lenientConversion
+} ConversionFlags;
+
+/* This is for C++ and does no harm in C */
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+ConversionResult ConvertUTF8toUTF16 (
+  const UTF8** sourceStart, const UTF8* sourceEnd,
+  UTF16** targetStart, UTF16* targetEnd, ConversionFlags flags);
+
+/**
+ * Convert a partial UTF8 sequence to UTF32.  If the sequence ends in an
+ * incomplete code unit sequence, returns \c sourceExhausted.
+ */
+ConversionResult ConvertUTF8toUTF32Partial(
+  const UTF8** sourceStart, const UTF8* sourceEnd,
+  UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags);
+
+/**
+ * Convert a partial UTF8 sequence to UTF32.  If the sequence ends in an
+ * incomplete code unit sequence, returns \c sourceIllegal.
+ */
+ConversionResult ConvertUTF8toUTF32(
+  const UTF8** sourceStart, const UTF8* sourceEnd,
+  UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags);
+
+ConversionResult ConvertUTF16toUTF8 (
+  const UTF16** sourceStart, const UTF16* sourceEnd,
+  UTF8** targetStart, UTF8* targetEnd, ConversionFlags flags);
+
+ConversionResult ConvertUTF32toUTF8 (
+  const UTF32** sourceStart, const UTF32* sourceEnd,
+  UTF8** targetStart, UTF8* targetEnd, ConversionFlags flags);
+
+ConversionResult ConvertUTF16toUTF32 (
+  const UTF16** sourceStart, const UTF16* sourceEnd,
+  UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags);
+
+ConversionResult ConvertUTF32toUTF16 (
+  const UTF32** sourceStart, const UTF32* sourceEnd,
+  UTF16** targetStart, UTF16* targetEnd, ConversionFlags flags);
+
+Boolean isLegalUTF8Sequence(const UTF8 *source, const UTF8 *sourceEnd);
+
+Boolean isLegalUTF8String(const UTF8 **source, const UTF8 *sourceEnd);
+
+unsigned getNumBytesForUTF8(UTF8 firstByte);
+
+#ifdef __cplusplus
+}
+
+#if 0 // ignored in mailcore2
+/*************************************************************************/
+/* Below are LLVM-specific wrappers of the functions above. */
+
+#include "llvm/ADT/ArrayRef.h"
+#include "llvm/ADT/StringRef.h"
+
+namespace llvm {
+
+/**
+ * Convert an UTF8 StringRef to UTF8, UTF16, or UTF32 depending on
+ * WideCharWidth. The converted data is written to ResultPtr, which needs to
+ * point to at least WideCharWidth * (Source.Size() + 1) bytes. On success,
+ * ResultPtr will point one after the end of the copied string. On failure,
+ * ResultPtr will not be changed, and ErrorPtr will be set to the location of
+ * the first character which could not be converted.
+ * \return true on success.
+ */
+bool ConvertUTF8toWide(unsigned WideCharWidth, llvm::StringRef Source,
+                       char *&ResultPtr, const UTF8 *&ErrorPtr);
+
+/**
+ * Convert an Unicode code point to UTF8 sequence.
+ *
+ * \param Source a Unicode code point.
+ * \param [in,out] ResultPtr pointer to the output buffer, needs to be at least
+ * \c UNI_MAX_UTF8_BYTES_PER_CODE_POINT bytes.  On success \c ResultPtr is
+ * updated one past end of the converted sequence.
+ *
+ * \returns true on success.
+ */
+bool ConvertCodePointToUTF8(unsigned Source, char *&ResultPtr);
+
+/**
+ * Convert the first UTF8 sequence in the given source buffer to a UTF32
+ * code point.
+ *
+ * \param [in,out] source A pointer to the source buffer. If the conversion
+ * succeeds, this pointer will be updated to point to the byte just past the
+ * end of the converted sequence.
+ * \param sourceEnd A pointer just past the end of the source buffer.
+ * \param [out] target The converted code
+ * \param flags Whether the conversion is strict or lenient.
+ *
+ * \returns conversionOK on success
+ *
+ * \sa ConvertUTF8toUTF32
+ */
+static inline ConversionResult convertUTF8Sequence(const UTF8 **source,
+                                                   const UTF8 *sourceEnd,
+                                                   UTF32 *target,
+                                                   ConversionFlags flags) {
+  if (*source == sourceEnd)
+    return sourceExhausted;
+  unsigned size = getNumBytesForUTF8(**source);
+  if ((ptrdiff_t)size > sourceEnd - *source)
+    return sourceExhausted;
+  return ConvertUTF8toUTF32(source, *source + size, &target, target + 1, flags);
+}
+
+/**
+ * Returns true if a blob of text starts with a UTF-16 big or little endian byte
+ * order mark.
+ */
+bool hasUTF16ByteOrderMark(ArrayRef<char> SrcBytes);
+
+/**
+ * Converts a stream of raw bytes assumed to be UTF16 into a UTF8 std::string.
+ *
+ * \param [in] SrcBytes A buffer of what is assumed to be UTF-16 encoded text.
+ * \param [out] Out Converted UTF-8 is stored here on success.
+ * \returns true on success
+ */
+bool convertUTF16ToUTF8String(ArrayRef<char> SrcBytes, std::string &Out);
+
+} /* end namespace llvm */
+#endif // ignored in mailcore2
+
+#endif
+
+/* --------------------------------------------------------------------- */
+
+#endif
diff --git a/src/core/basetypes/MCString.cc b/src/core/basetypes/MCString.cc
index c005ceb1..faee0005 100644
--- a/src/core/basetypes/MCString.cc
+++ b/src/core/basetypes/MCString.cc
@@ -33,6 +33,7 @@
 #include "MCHTMLCleaner.h"
 #include "MCBase64.h"
 #include "MCIterator.h"
+#include "ConvertUTF.h"
 
 using namespace mailcore;
 
@@ -960,17 +961,13 @@ void String::appendUTF8CharactersLength(const char * UTF8Characters, unsigned in
     }
 
 #if DISABLE_ICU
-    CFStringRef cfStr = CFStringCreateWithBytes(NULL, (const UInt8 *) UTF8Characters, length,
-                                                kCFStringEncodingUTF8, false);
-    if (cfStr == NULL) {
-        // Data could not be converted to UTF-8.
-        return;
-    }
-    UniChar * characters = (UniChar *) malloc(sizeof(* characters) * CFStringGetLength(cfStr));
-    CFStringGetCharacters(cfStr, CFRangeMake(0, CFStringGetLength(cfStr)), characters);
-    appendCharactersLength(characters, (unsigned int) CFStringGetLength(cfStr));
-    free(characters);
-    CFRelease(cfStr);
+    const UTF8 * source = (const UTF8 *) UTF8Characters;
+    UTF16 * target = (UTF16 *) malloc(length * sizeof(* target));
+    UTF16 * targetStart = target;
+    ConvertUTF8toUTF16(&source, source + length,
+                       &targetStart, targetStart + length, lenientConversion);
+    appendCharactersLength((UChar *) target, (unsigned int ) (targetStart - target));
+    free(target);
 #else
     UChar * dest;
     int32_t destLength;
@@ -1018,13 +1015,15 @@ const UChar * String::unicodeCharacters()
 const char * String::UTF8Characters()
 {
 #if DISABLE_ICU
-    CFStringRef cfStr = CFStringCreateWithCharactersNoCopy(NULL, mUnicodeChars, mLength, kCFAllocatorNull);
-    char * buffer = (char *) malloc(mLength * 6 + 1);
-    CFIndex usedBufLen;
-    CFStringGetBytes(cfStr, CFRangeMake(0, mLength), kCFStringEncodingUTF8, '?', false, (UInt8 *) buffer, mLength * 6, &usedBufLen);
-    buffer[usedBufLen] = 0;
-    Data * data = Data::dataWithBytes(buffer, (unsigned int) usedBufLen + 1);
-    free(buffer);
+    const UTF16 * source = (const UTF16 *) mUnicodeChars;
+    UTF8 * target = (UTF8 *) malloc(mLength * 6 + 1);
+    UTF8 * targetStart = target;
+    ConvertUTF16toUTF8(&source, source + mLength,
+                       &targetStart, targetStart + mLength * 6 + 1, lenientConversion);
+    unsigned int utf8length = (unsigned int) (targetStart - target);
+    target[utf8length] = 0;
+    Data * data = Data::dataWithBytes((const char *) target, utf8length + 1);
+    free(target);
     
     return data->bytes();
 #else
diff --git a/unittest/data/parser/output/mbox/jwz/58 b/unittest/data/parser/output/mbox/jwz/58
index fa967dda..ba0337d5 100644
--- a/unittest/data/parser/output/mbox/jwz/58
+++ b/unittest/data/parser/output/mbox/jwz/58
@@ -1 +1 @@
-{"header":{"to":{"class":"mailcore::Array","items":[{"class":"mailcore::Address","displayName":"Shlomi Fish","mailbox":"shlomi@Medusa.Cortext.co.il"}]},"from":{"class":"mailcore::Address","displayName":"Vladim\u00edr Solnick\u00fd","mailbox":"vs@utia.cas.cz"},"extraHeaders":{"Disposition-Notification-To":"vs+disposition@utia.cas.cz","X-Chameleon-Return-To":"vs+reading@utia.cas.cz","X-Confirm-Delivery-To":"vs+delivery@utia.cas.cz","Acknowledge-To":"vs+delivery@utia.cas.cz","Return-Receipt-To":"vs+delivery@utia.cas.cz","X-Confirm-Reading-To":"vs+reading@utia.cas.cz","Content-Type":"TEXT\/PLAIN; charset=ISO-8859-2","Lines":"35","X-Postal-Address":"=?ISO-8859-2?Q?=DATIA_AV_=C8R=3B_Pod_vod=E1renskou_v=EC=BE=ED?=\u000a =?ISO-8859-2?Q?_4=3B_182_08__Praha_8-Libe=F2=3B_Czech_Rep?=\u000a =?ISO-8859-2?Q?ublic?=","NNTP-Posting-Host":"visla.utia.cas.cz","Newsgroups":"comp.mail.mime","Transport-Options":"\/delivery \/return","Content-Transfer-Encoding":"8BIT","Mime-Version":"1.0","Path":"tera.mcom.com!news.Stanford.EDU!bloom-beacon.mit.edu!news.mathworks.com!nntp.primenet.com!uunet!inXS.uu.net!voskovec.radio.cz!visla.utia.cas.cz!vs","Organization":"Institute of Information Theory and Automation (UTIA AV CR)","Read-Receipt-To":"vs+reading@utia.cas.cz","Organizace":"=?ISO-8859-2?Q?=DAstav_teorie_informace_a_automatizace_AV_=C8?=\u000a =?ISO-8859-2?Q?R_=28=DATIA=29?="},"date":"978307200","messageID":"Pine.HPP.3.94.1.960713165148.28138A-100000@visla.utia.cas.cz","receivedDate":"978307200","subject":"Re: Sending a Hebrew message with sendmail","class":"mailcore::MessageHeader","replyTo":{"class":"mailcore::Array","items":[{"class":"mailcore::Address","displayName":"Bc. Vladim\u00edr Solnick\u00fd","mailbox":"vs@utia.cas.cz"}]},"inReplyTo":["4s36v9$gi0@news.NetVision.net.il"],"references":["4s36v9$gi0@news.NetVision.net.il"]},"class":"mailcore::MessageParser","mainPart":{"partType":"single","uniqueID":"0","mimeType":"text\/PLAIN","charset":"ISO-8859-2","class":"mailcore::Attachment"}}
\ No newline at end of file
+{"header":{"to":{"class":"mailcore::Array","items":[{"class":"mailcore::Address","displayName":"Shlomi Fish","mailbox":"shlomi@Medusa.Cortext.co.il"}]},"from":{"class":"mailcore::Address","displayName":"Vladim\u00edr Solnick\u00fd","mailbox":"vs@utia.cas.cz"},"extraHeaders":{"Disposition-Notification-To":"vs+disposition@utia.cas.cz","X-Chameleon-Return-To":"vs+reading@utia.cas.cz","X-Confirm-Delivery-To":"vs+delivery@utia.cas.cz","Acknowledge-To":"vs+delivery@utia.cas.cz","Return-Receipt-To":"vs+delivery@utia.cas.cz","X-Confirm-Reading-To":"vs+reading@utia.cas.cz","Content-Type":"TEXT\/PLAIN; charset=ISO-8859-2","Lines":"35","X-Postal-Address":"\u00daTIA AV \u010cR; Pod vod\u00e1renskou v\u011b\u017e\u00ed 4; 182 08  Praha 8-Libe\u0148; Czech Republic","NNTP-Posting-Host":"visla.utia.cas.cz","Newsgroups":"comp.mail.mime","Transport-Options":"\/delivery \/return","Content-Transfer-Encoding":"8BIT","Mime-Version":"1.0","Path":"tera.mcom.com!news.Stanford.EDU!bloom-beacon.mit.edu!news.mathworks.com!nntp.primenet.com!uunet!inXS.uu.net!voskovec.radio.cz!visla.utia.cas.cz!vs","Organization":"Institute of Information Theory and Automation (UTIA AV CR)","Read-Receipt-To":"vs+reading@utia.cas.cz","Organizace":"\u00dastav teorie informace a automatizace AV \u010cR (\u00daTIA)"},"date":"978307200","messageID":"Pine.HPP.3.94.1.960713165148.28138A-100000@visla.utia.cas.cz","receivedDate":"978307200","subject":"Re: Sending a Hebrew message with sendmail","class":"mailcore::MessageHeader","replyTo":{"class":"mailcore::Array","items":[{"class":"mailcore::Address","displayName":"Bc. Vladim\u00edr Solnick\u00fd","mailbox":"vs@utia.cas.cz"}]},"inReplyTo":["4s36v9$gi0@news.NetVision.net.il"],"references":["4s36v9$gi0@news.NetVision.net.il"]},"class":"mailcore::MessageParser","mainPart":{"partType":"single","uniqueID":"0","mimeType":"text\/PLAIN","charset":"ISO-8859-2","class":"mailcore::Attachment"}}
\ No newline at end of file