[cfe-commits] r68208 - in /cfe/trunk: clang.xcodeproj/project.pbxproj include/clang/Basic/ConvertUTF.h lib/Basic/CMakeLists.txt lib/Basic/ConvertUTF.c lib/CodeGen/CodeGenModule.cpp
Steve Naroff
snaroff at apple.com
Wed Apr 1 08:50:34 PDT 2009
Author: snaroff
Date: Wed Apr 1 10:50:34 2009
New Revision: 68208
URL: http://llvm.org/viewvc/llvm-project?rev=68208&view=rev
Log:
Add ConvertUTF module from http://www.unicode.org/Public/PROGRAMS/CVTUTF.
#ifdef'd out the 5 conversion routines that we don't currently need.
Still need a bit more work in GetAddrOfConstantCFString(). Added a FIXME to indicate this.
Expect to remove the FIXME today...
Added:
cfe/trunk/include/clang/Basic/ConvertUTF.h
cfe/trunk/lib/Basic/ConvertUTF.c
Modified:
cfe/trunk/clang.xcodeproj/project.pbxproj
cfe/trunk/lib/Basic/CMakeLists.txt
cfe/trunk/lib/CodeGen/CodeGenModule.cpp
Modified: cfe/trunk/clang.xcodeproj/project.pbxproj
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/clang.xcodeproj/project.pbxproj?rev=68208&r1=68207&r2=68208&view=diff
==============================================================================
--- cfe/trunk/clang.xcodeproj/project.pbxproj (original)
+++ cfe/trunk/clang.xcodeproj/project.pbxproj Wed Apr 1 10:50:34 2009
@@ -87,6 +87,7 @@
84AF36A10CB17A3B00C820A5 /* DeclObjC.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = 84AF36A00CB17A3B00C820A5 /* DeclObjC.h */; };
84D9A8880C1A57E100AC7ABC /* AttributeList.cpp in Sources */ = {isa = PBXBuildFile; fileRef = 84D9A8870C1A57E100AC7ABC /* AttributeList.cpp */; };
84D9A88C0C1A581300AC7ABC /* AttributeList.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = 84D9A88B0C1A581300AC7ABC /* AttributeList.h */; };
+ 906BF4B00F83BA2E001071FA /* ConvertUTF.c in Sources */ = {isa = PBXBuildFile; fileRef = 906BF4AF0F83BA2E001071FA /* ConvertUTF.c */; };
DE01DA490B12ADA300AC22CE /* PPCallbacks.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DE01DA480B12ADA300AC22CE /* PPCallbacks.h */; };
DE06756C0C051CFE00EBBFD8 /* ParseExprCXX.cpp in Sources */ = {isa = PBXBuildFile; fileRef = DE06756B0C051CFE00EBBFD8 /* ParseExprCXX.cpp */; };
DE06B73E0A8307640050E87E /* LangOptions.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DE06B73D0A8307640050E87E /* LangOptions.h */; };
@@ -449,6 +450,8 @@
84D9A8870C1A57E100AC7ABC /* AttributeList.cpp */ = {isa = PBXFileReference; fileEncoding = 30; indentWidth = 2; lastKnownFileType = sourcecode.cpp.cpp; name = AttributeList.cpp; path = lib/Parse/AttributeList.cpp; sourceTree = "<group>"; tabWidth = 2; };
84D9A88B0C1A581300AC7ABC /* AttributeList.h */ = {isa = PBXFileReference; fileEncoding = 30; indentWidth = 2; lastKnownFileType = sourcecode.c.h; name = AttributeList.h; path = clang/Parse/AttributeList.h; sourceTree = "<group>"; tabWidth = 2; };
8DD76F6C0486A84900D96B5E /* clang */ = {isa = PBXFileReference; explicitFileType = "compiled.mach-o.executable"; includeInIndex = 0; path = clang; sourceTree = BUILT_PRODUCTS_DIR; };
+ 906BF4AE0F83BA16001071FA /* ConvertUTF.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = ConvertUTF.h; sourceTree = "<group>"; };
+ 906BF4AF0F83BA2E001071FA /* ConvertUTF.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = ConvertUTF.c; sourceTree = "<group>"; };
DE01DA480B12ADA300AC22CE /* PPCallbacks.h */ = {isa = PBXFileReference; fileEncoding = 30; lastKnownFileType = sourcecode.c.h; path = PPCallbacks.h; sourceTree = "<group>"; };
DE06756B0C051CFE00EBBFD8 /* ParseExprCXX.cpp */ = {isa = PBXFileReference; fileEncoding = 30; indentWidth = 2; lastKnownFileType = sourcecode.cpp.cpp; name = ParseExprCXX.cpp; path = lib/Parse/ParseExprCXX.cpp; sourceTree = "<group>"; tabWidth = 2; };
DE06B73D0A8307640050E87E /* LangOptions.h */ = {isa = PBXFileReference; fileEncoding = 30; indentWidth = 2; lastKnownFileType = sourcecode.c.h; path = LangOptions.h; sourceTree = "<group>"; tabWidth = 2; };
@@ -1128,6 +1131,7 @@
DED7D7300A524295003AD0FB /* Basic */ = {
isa = PBXGroup;
children = (
+ 906BF4AE0F83BA16001071FA /* ConvertUTF.h */,
DED7D7310A524295003AD0FB /* Diagnostic.h */,
1A7019E90F79BC1100FEC4D1 /* DiagnosticAnalysisKinds.td */,
1A7019EA0F79BC1100FEC4D1 /* DiagnosticASTKinds.td */,
@@ -1183,6 +1187,7 @@
DED7D7500A5242C7003AD0FB /* Basic */ = {
isa = PBXGroup;
children = (
+ 906BF4AF0F83BA2E001071FA /* ConvertUTF.c */,
DED7D75D0A5242C7003AD0FB /* Diagnostic.cpp */,
DED7D75E0A5242C7003AD0FB /* FileManager.cpp */,
DE3986F30CB8D50C00223765 /* IdentifierTable.cpp */,
@@ -1557,6 +1562,7 @@
DEDFE6670F7B3B830035BD10 /* HostInfo.cpp in Sources */,
DEDFE6680F7B3B830035BD10 /* Driver.cpp in Sources */,
1A701B640F7C8FE400FEC4D1 /* SemaAccess.cpp in Sources */,
+ 906BF4B00F83BA2E001071FA /* ConvertUTF.c in Sources */,
);
runOnlyForDeploymentPostprocessing = 0;
};
Added: cfe/trunk/include/clang/Basic/ConvertUTF.h
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Basic/ConvertUTF.h?rev=68208&view=auto
==============================================================================
--- cfe/trunk/include/clang/Basic/ConvertUTF.h (added)
+++ cfe/trunk/include/clang/Basic/ConvertUTF.h Wed Apr 1 10:50:34 2009
@@ -0,0 +1,159 @@
+/*===--- 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.
+
+------------------------------------------------------------------------ */
+
+/* ---------------------------------------------------------------------
+ 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 long 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
+
+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);
+
+#ifdef CLANG_NEEDS_THESE_ONE_DAY
+ConversionResult ConvertUTF16toUTF8 (
+ const UTF16** sourceStart, const UTF16* sourceEnd,
+ UTF8** targetStart, UTF8* targetEnd, ConversionFlags flags);
+
+ConversionResult ConvertUTF8toUTF32 (
+ const UTF8** sourceStart, const UTF8* sourceEnd,
+ UTF32** targetStart, UTF32* 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);
+#endif
+
+Boolean isLegalUTF8Sequence(const UTF8 *source, const UTF8 *sourceEnd);
+
+#ifdef __cplusplus
+}
+#endif
+
+/* --------------------------------------------------------------------- */
Modified: cfe/trunk/lib/Basic/CMakeLists.txt
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Basic/CMakeLists.txt?rev=68208&r1=68207&r2=68208&view=diff
==============================================================================
--- cfe/trunk/lib/Basic/CMakeLists.txt (original)
+++ cfe/trunk/lib/Basic/CMakeLists.txt Wed Apr 1 10:50:34 2009
@@ -1,6 +1,7 @@
set(LLVM_NO_RTTI 1)
add_clang_library(clangBasic
+ ConvertUTF.c
Diagnostic.cpp
FileManager.cpp
IdentifierTable.cpp
Added: cfe/trunk/lib/Basic/ConvertUTF.c
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Basic/ConvertUTF.c?rev=68208&view=auto
==============================================================================
--- cfe/trunk/lib/Basic/ConvertUTF.c (added)
+++ cfe/trunk/lib/Basic/ConvertUTF.c Wed Apr 1 10:50:34 2009
@@ -0,0 +1,547 @@
+/*===--- 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 "clang/Basic/ConvertUTF.h"
+#ifdef CVTUTF_DEBUG
+#include <stdio.h>
+#endif
+
+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.
+ */
+
+#ifdef CLANG_NEEDS_THESE_ONE_DAY
+
+/* --------------------------------------------------------------------- */
+
+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;
+}
+
+/* --------------------------------------------------------------------- */
+
+ConversionResult ConvertUTF8toUTF32 (
+ const UTF8** sourceStart, const UTF8* sourceEnd,
+ UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags) {
+ ConversionResult result = conversionOK;
+ const UTF8* source = *sourceStart;
+ UTF32* target = *targetStart;
+ while (source < sourceEnd) {
+ UTF32 ch = 0;
+ unsigned short extraBytesToRead = trailingBytesForUTF8[*source];
+ if (source + extraBytesToRead >= sourceEnd) {
+ 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;
+ 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 (target >= targetEnd) {
+ source -= (extraBytesToRead+1); /* Back up the source pointer! */
+ result = targetExhausted; break;
+ }
+ 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;
+}
+#endif
+
+/* --------------------------------------------------------------------- */
+
+/*
+ * 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)) > 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 (source+length > sourceEnd) {
+ return false;
+ }
+ return isLegalUTF8(source, length);
+}
+
+/* --------------------------------------------------------------------- */
+
+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 (source + extraBytesToRead >= sourceEnd) {
+ 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;
+}
+
+/* ---------------------------------------------------------------------
+
+ 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.
+
+ --------------------------------------------------------------------- */
Modified: cfe/trunk/lib/CodeGen/CodeGenModule.cpp
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/CodeGen/CodeGenModule.cpp?rev=68208&r1=68207&r2=68208&view=diff
==============================================================================
--- cfe/trunk/lib/CodeGen/CodeGenModule.cpp (original)
+++ cfe/trunk/lib/CodeGen/CodeGenModule.cpp Wed Apr 1 10:50:34 2009
@@ -24,6 +24,7 @@
#include "clang/Basic/Diagnostic.h"
#include "clang/Basic/SourceManager.h"
#include "clang/Basic/TargetInfo.h"
+#include "clang/Basic/ConvertUTF.h"
#include "llvm/CallingConv.h"
#include "llvm/Module.h"
#include "llvm/Intrinsics.h"
@@ -1003,9 +1004,21 @@
// See: <rdr://2996215>
llvm::Constant *CodeGenModule::
GetAddrOfConstantCFString(const StringLiteral *Literal) {
- // if (Literal->containsNonAsciiOrNull()) {
- // // FIXME: Convert from UTF-8 to UTF-16.
- // }
+ bool isUTF16 = false;
+ if (Literal->containsNonAsciiOrNull()) {
+ // Convert from UTF-8 to UTF-16.
+ llvm::SmallVector<UTF16, 128> ToBuf(Literal->getByteLength());
+ const UTF8 *FromPtr = (UTF8 *)Literal->getStrData();
+ UTF16 *ToPtr = &ToBuf[0];
+
+ ConversionResult Result;
+ Result = ConvertUTF8toUTF16(&FromPtr, FromPtr+Literal->getByteLength(),
+ &ToPtr, ToPtr+Literal->getByteLength(),
+ strictConversion);
+ assert(Result == conversionOK && "UTF-8 to UTF-16 conversion failed");
+ isUTF16 = true;
+ // FIXME: Do something with the converted value!
+ }
std::string str(Literal->getStrData(), Literal->getByteLength());
llvm::StringMapEntry<llvm::Constant *> &Entry =
CFConstantStringMap.GetOrCreateValue(&str[0], &str[str.length()]);
@@ -1056,7 +1069,9 @@
NextField = *Field++;
const llvm::Type *Ty = getTypes().ConvertType(getContext().UnsignedIntTy);
appendFieldAndPadding(*this, Fields, CurField, NextField,
- llvm::ConstantInt::get(Ty, 0x07C8), CFRD, STy);
+ isUTF16 ? llvm::ConstantInt::get(Ty, 0x07d0)
+ : llvm::ConstantInt::get(Ty, 0x07C8),
+ CFRD, STy);
// String pointer.
CurField = NextField;
More information about the cfe-commits
mailing list