From 8f2f9b74e370b2314375d0347271c7282ecc762c Mon Sep 17 00:00:00 2001 From: Dennis Heimbigner Date: Tue, 6 Jun 2017 15:23:59 -0600 Subject: [PATCH 1/3] re: pull request https://github.com/Unidata/netcdf-c/pull/364 This is a follow-on in that the old utf8 code was still being used in ncgen to convert utf8->utf16 when converting cdl to Java (see genj.c). The new code apparently has no utf16 support, but it does have utf32 support. Converting utf32 -> utf16 can be approximated by truncating the 32bits to 16 bits, unless the top 16 bits are not zero. This latter condition is unlikely to be common because it implies use of some rather obscure characters. So solution is to convert to utf32 and truncate to 16 bits to get utf16. An error is reported if the high-order truncated 16 bits are not zero. If we get complaints, then I will figure out how to convert full utf32 to a utf16 pair. Also removed the old code from ncgen. --- cf | 3 +- include/ncutf8.h | 13 ++ libdispatch/dutf8.c | 80 ++++++- ncgen/ConvertUTF.c | 539 -------------------------------------------- ncgen/ConvertUTF.h | 149 ------------ ncgen/Makefile.am | 2 +- ncgen/escapes.c | 68 +----- 7 files changed, 104 insertions(+), 750 deletions(-) delete mode 100644 ncgen/ConvertUTF.c delete mode 100644 ncgen/ConvertUTF.h diff --git a/cf b/cf index 967868beb..6e66f166e 100644 --- a/cf +++ b/cf @@ -54,8 +54,7 @@ MALLOC_CHECK="" CPPFLAGS="" LDFLAGS="" -#CFLAGS="-g -O0 $CFLAGS" -CFLAGS="-g $CFLAGS" +CFLAGS="-g -O0 $CFLAGS -Wno-undefined" MAKE=make IGNORE="test 0 = 1" diff --git a/include/ncutf8.h b/include/ncutf8.h index 4ddbf755f..2e9b11460 100644 --- a/include/ncutf8.h +++ b/include/ncutf8.h @@ -30,5 +30,18 @@ extern int nc_utf8_validate(const unsigned char * name); */ extern int nc_utf8_normalize(const unsigned char* str, unsigned char** normalp); +/* + * Convert a normalized utf8 string to utf16. This is approximate + * because it just does the truncation version of conversion for + * each 32-bit codepoint to get the corresponding utf16. + * Return codes: + * NC_NOERR -- success + * NC_ENOMEM -- out of memory + * NC_EINVAL -- invalid argument or internal error + * NC_EBADNAME-- not valid utf16 + */ + +extern int nc_utf8_to_utf16(const unsigned char* s8, unsigned short** utf16p, size_t* lenp); + #endif /*NCUTF8_H*/ diff --git a/libdispatch/dutf8.c b/libdispatch/dutf8.c index 0ff99de48..08bea64ee 100644 --- a/libdispatch/dutf8.c +++ b/libdispatch/dutf8.c @@ -4,6 +4,12 @@ */ #include "config.h" +#ifdef HAVE_STDLIB_H +#include +#endif +#ifdef HAVE_STRING_H +#include +#endif #include "netcdf.h" #include "ncutf8.h" #include "utf8proc.h" @@ -23,13 +29,13 @@ int nc_utf8_validate(const unsigned char* name) { int ncstat = NC_NOERR; const nc_utf8proc_uint8_t *str; - nc_utf8proc_ssize_t strlen = -1; + nc_utf8proc_ssize_t nchars = -1; nc_utf8proc_int32_t codepoint; nc_utf8proc_ssize_t count; str = (const nc_utf8proc_uint8_t*)name; while(*str) { - count = nc_utf8proc_iterate(str,strlen,&codepoint); + count = nc_utf8proc_iterate(str,nchars,&codepoint); if(count < 0) { switch (count) { case UTF8PROC_ERROR_NOMEM: @@ -94,3 +100,73 @@ nc_utf8_normalize(const unsigned char* utf8, unsigned char** normalp) done: return ncstat; } + +/* + * Convert a normalized utf8 string to utf16. This is approximate + * because it just does the truncation version of conversion for + * each 32-bit codepoint to get the corresponding utf16. + * Return codes: + * NC_NOERR -- success + * NC_ENOMEM -- out of memory + * NC_EINVAL -- invalid argument or internal error + * NC_EBADNAME-- not valid utf16 + */ + +int nc_utf8_to_utf16(const unsigned char* s8, unsigned short** utf16p, size_t* len16p) +{ + int ncstat = NC_NOERR; + const nc_utf8proc_uint8_t *str; + nc_utf8proc_ssize_t nchars = -1; + nc_utf8proc_int32_t codepoint; + nc_utf8proc_ssize_t count; + size_t len8, len16; + unsigned short* utf16; + unsigned short* p16; + + len8 = strlen((char*)s8); + utf16 = (unsigned short*)malloc(sizeof(unsigned short)*(len8+1)); + if(utf16 == NULL) { + ncstat = NC_ENOMEM; + goto done; + } + str = (const nc_utf8proc_uint8_t*)s8; + /* Walk the string and convert each codepoint */ + p16 = utf16; + len16 = 0; + while(*str) { + count = nc_utf8proc_iterate(str,nchars,&codepoint); + if(count < 0) { + switch (count) { + case UTF8PROC_ERROR_NOMEM: + case UTF8PROC_ERROR_OVERFLOW: + ncstat = NC_ENOMEM; + break; + case UTF8PROC_ERROR_INVALIDOPTS: + ncstat = NC_EINVAL; + break; + case UTF8PROC_ERROR_INVALIDUTF8: + case UTF8PROC_ERROR_NOTASSIGNED: + default: + ncstat = NC_EBADNAME; + break; + } + goto done; + } else { /* move to next char */ + /* Complain if top 16 bits not zero */ + if((codepoint & 0x0000FFFF) != 0) { + ncstat = NC_EBADNAME; + goto done; + } + /* Truncate codepoint to 16 bits and store */ + *p16++ = (unsigned short)(codepoint & 0x0000FFFF); + str += count; + len16++; + } + } + *p16++ = (unsigned short)0; + if(utf16p) *utf16p = utf16; + if(len16p) *len16p = len16; +done: + if(ncstat) free(utf16); + return ncstat; +} diff --git a/ncgen/ConvertUTF.c b/ncgen/ConvertUTF.c deleted file mode 100644 index 0be73c363..000000000 --- a/ncgen/ConvertUTF.c +++ /dev/null @@ -1,539 +0,0 @@ -/* - * 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 "config.h" -#include "ConvertUTF.h" -#ifdef CVTUTF_DEBUG -#include -#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 - -/* --------------------------------------------------------------------- */ - -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; -} - -/* --------------------------------------------------------------------- */ - -/* - * 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 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; -} - -/* --------------------------------------------------------------------- */ - -/* - * 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; -} - -/* --------------------------------------------------------------------- */ - -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; -} - -/* --------------------------------------------------------------------- - - 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/ncgen/ConvertUTF.h b/ncgen/ConvertUTF.h deleted file mode 100644 index e26e7c978..000000000 --- a/ncgen/ConvertUTF.h +++ /dev/null @@ -1,149 +0,0 @@ -/* - * 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 functions 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: , , - or 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); - -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); - -Boolean isLegalUTF8Sequence(const UTF8 *source, const UTF8 *sourceEnd); - -#ifdef __cplusplus -} -#endif - -/* --------------------------------------------------------------------- */ diff --git a/ncgen/Makefile.am b/ncgen/Makefile.am index f7a8a0509..ee9ec3488 100644 --- a/ncgen/Makefile.am +++ b/ncgen/Makefile.am @@ -14,7 +14,7 @@ escapes.c genc.c genbin.c generr.c genlib.c getfill.c odom.c \ semantics.c dump.c util.c bytebuffer.c list.c data.h \ debug.h generr.h genlib.h includes.h ncgen.h odom.h dump.h \ util.h bytebuffer.h list.h genf77.c f77data.c genj.c jdata.c nc_iter.h \ -nc_iter.c ConvertUTF.c ConvertUTF.h \ +nc_iter.c \ ncgeny.c ncgeny.h # This is the man page. diff --git a/ncgen/escapes.c b/ncgen/escapes.c index 5f24fa930..3211728b3 100644 --- a/ncgen/escapes.c +++ b/ncgen/escapes.c @@ -5,7 +5,7 @@ *********************************************************************/ #include "includes.h" -#include "ConvertUTF.h" +#include "ncutf8.h" #define HEXCHARS "0123456789abcdefABCDEF" #define OCTCHARS "01234567" @@ -16,48 +16,6 @@ static char* ccodify(const char*); static char* f77codify(const char*); static char* jcodify(const char*); -#if 0 -/* - * Replace escaped chars in CDL representation of name such as - * 'abc\:def\ gh\\i' with unescaped version, such as 'abc:def gh\i'. - */ -/* ?? This seems redundant over expand_escapes*/ -void -deescapify(char* name) -{ - const char *cp = name; - char *sp; - size_t len = strlen(name); - char *newname; - - if(strchr(name, '\\') == NULL) - return; - - newname = (char *) emalloc(len + 1); - cp = name; - sp = newname; - while(*cp != '\0') { /* delete '\' chars, except change '\\' to '\' */ - switch (*cp) { - case '\\': - if(*(cp+1) == '\\') { - *sp++ = '\\'; - cp++; - } - break; - default: - *sp++ = *cp; - break; - } - cp++; - } - *sp = '\0'; - /* ASSERT(strlen(newname) <= strlen(name)); */ - strncpy(name, newname, len+1); /* watch out for trailing null*/ - efree(newname); - return; -} -#endif /*0*/ - /* Given a character c, fill s with the character suitably escaped. E.g. c = '\t' => s="\t" @@ -368,7 +326,7 @@ to suitably escape c for use with Java. */ static void -jescapifychar(UTF16 c, int quote, Bytebuffer* s) +jescapifychar(unsigned short c, int quote, Bytebuffer* s) { /* Separate out ascii from UTF16 */ if(c <= '\177') { @@ -422,25 +380,21 @@ jescapifychar(UTF16 c, int quote, Bytebuffer* s) char* jescapify(char* s0, int quote, size_t len) { + int stat = NC_NOERR; int i; - char* result; - UTF8* s8; - UTF16* s16; /* for storing the utf16 string */ - UTF16* tmp16; /* for storing the utf16 string */ - ConversionResult status; + char* result = NULL; + unsigned char* s8 = NULL; + unsigned short* s16 = NULL; /* for storing the utf16 string */ Bytebuffer* escaped = bbNew(); size_t len16; - s16 = emalloc((1+len)*sizeof(UTF16)); - s8 = (UTF8*)s0; - tmp16 = s16; - status = ConvertUTF8toUTF16((const UTF8**)&s8,s8+len,&tmp16,tmp16+len,lenientConversion); - if(status != conversionOK) { + s8 = (unsigned char*)s0; + stat = nc_utf8_to_utf16(s8, &s16, &len16); + if(stat != NC_NOERR) { derror("Cannot convert UTF8 string to UTF16: %s",s0); - return NULL; + if(s16) efree(s16); + return NULL; } - /* Get the length of the utf16 string */ - len16 = (tmp16 - s16); for(i=0;i Date: Tue, 6 Jun 2017 16:21:04 -0600 Subject: [PATCH 2/3] Cmake --- cf.cmake | 2 +- include/ncutf8.h | 1 - ncgen/CMakeLists.txt | 2 +- 3 files changed, 2 insertions(+), 3 deletions(-) diff --git a/cf.cmake b/cf.cmake index dcb842ba1..8820e0eb3 100644 --- a/cf.cmake +++ b/cf.cmake @@ -1,5 +1,5 @@ # Visual Studio -#VS=1 +VS=1 # Is netcdf-4 and/or DAP enabled? NC4=1 diff --git a/include/ncutf8.h b/include/ncutf8.h index 2e9b11460..163f93158 100644 --- a/include/ncutf8.h +++ b/include/ncutf8.h @@ -44,4 +44,3 @@ extern int nc_utf8_normalize(const unsigned char* str, unsigned char** normalp); extern int nc_utf8_to_utf16(const unsigned char* s8, unsigned short** utf16p, size_t* lenp); #endif /*NCUTF8_H*/ - diff --git a/ncgen/CMakeLists.txt b/ncgen/CMakeLists.txt index 6feb4b939..7314acef5 100644 --- a/ncgen/CMakeLists.txt +++ b/ncgen/CMakeLists.txt @@ -3,7 +3,7 @@ IF(BUILD_SHARED_LIBS AND WIN32) remove_definitions(-DDLL_NETCDF) ENDIF() -SET(ncgen_FILES generate.c main.c cdata.c bindata.c genchar.c cvt.c data.c debug.c escapes.c genc.c genbin.c generr.c genlib.c getfill.c odom.c semantics.c ncgeny.c dump.c util.c bytebuffer.c list.c genf77.c f77data.c genj.c jdata.c nc_iter.c ConvertUTF.c ncgen.h) +SET(ncgen_FILES generate.c main.c cdata.c bindata.c genchar.c cvt.c data.c debug.c escapes.c genc.c genbin.c generr.c genlib.c getfill.c odom.c semantics.c ncgeny.c dump.c util.c bytebuffer.c list.c genf77.c f77data.c genj.c jdata.c nc_iter.c ncgen.h) # don't add the automatically determined parts of the RPATH # which point to directories outside the build tree to the install RPATH From 9cde91687ad37a4638701eb3a1afd41637a6b326 Mon Sep 17 00:00:00 2001 From: Dennis Heimbigner Date: Sun, 18 Jun 2017 21:06:19 -0600 Subject: [PATCH 3/3] re: pull request https://github.com/Unidata/netcdf-c/pull/364 This is a follow-on in that the old utf8 code was still being used in ncgen to convert utf8->utf16 when converting cdl to Java (see genj.c). The new code apparently has no utf16 support, but it does have utf32 support. Converting utf32 -> utf16 can be approximated by truncating the 32bits to 16 bits, unless the top 16 bits are not zero. This latter condition is unlikely to be common because it implies use of some rather obscure characters. So solution is to convert to utf32 and truncate to 16 bits to get utf16. An error is reported if the high-order truncated 16 bits are not zero. If we get complaints, then I will figure out how to convert full utf32 to a utf16 pair. Other changes: 1. removed the old code from ncgen. 2. changed UTF8PROC_DLLEXPORT (in utf8proc) to EXTERNL and added appropriate includes. This should fix issue https://github.com/Unidata/netcdf-c/issues/404, but since we cannot duplicate the failure, I am not quite sure. --- include/ncutf8.h | 8 +++++--- libdispatch/utf8proc.h | 25 ++++++------------------- 2 files changed, 11 insertions(+), 22 deletions(-) diff --git a/include/ncutf8.h b/include/ncutf8.h index 163f93158..fca22ca14 100644 --- a/include/ncutf8.h +++ b/include/ncutf8.h @@ -6,6 +6,8 @@ #ifndef NCUTF8_H #define NCUTF8_H 1 +#include "ncexternl.h" + /* Provide a wrapper around whatever utf8 library we use. */ /* @@ -15,7 +17,7 @@ * NC_ENOMEM -- out of memory * NC_EBADNAME-- not valid utf8 */ -extern int nc_utf8_validate(const unsigned char * name); +EXTERNL int nc_utf8_validate(const unsigned char * name); /* * Apply NFC normalization to a string. @@ -28,7 +30,7 @@ extern int nc_utf8_validate(const unsigned char * name); * NC_ENOMEM -- out of memory * NC_EBADNAME -- other failure */ -extern int nc_utf8_normalize(const unsigned char* str, unsigned char** normalp); +EXTERNL int nc_utf8_normalize(const unsigned char* str, unsigned char** normalp); /* * Convert a normalized utf8 string to utf16. This is approximate @@ -41,6 +43,6 @@ extern int nc_utf8_normalize(const unsigned char* str, unsigned char** normalp); * NC_EBADNAME-- not valid utf16 */ -extern int nc_utf8_to_utf16(const unsigned char* s8, unsigned short** utf16p, size_t* lenp); +EXTERNL int nc_utf8_to_utf16(const unsigned char* s8, unsigned short** utf16p, size_t* lenp); #endif /*NCUTF8_H*/ diff --git a/libdispatch/utf8proc.h b/libdispatch/utf8proc.h index cee988264..ceea425a7 100644 --- a/libdispatch/utf8proc.h +++ b/libdispatch/utf8proc.h @@ -77,6 +77,7 @@ /** @} */ #include +#include "ncexternl.h" #if defined(_MSC_VER) && _MSC_VER < 1800 // MSVC prior to 2013 lacked stdbool.h and inttypes.h @@ -120,20 +121,6 @@ typedef bool nc_utf8proc_bool; #endif #include -#ifndef UTF8PROC_EXPORTS -#ifdef _WIN32 -# ifdef UTF8PROC_EXPORTS -# define UTF8PROC_DLLEXPORT __declspec(dllexport) -# else -# define UTF8PROC_DLLEXPORT __declspec(dllimport) -# endif -#elif __GNUC__ >= 4 -# define UTF8PROC_DLLEXPORT __attribute__ ((visibility("default"))) -#else -# define UTF8PROC_DLLEXPORT -#endif -#endif - #ifdef __cplusplus extern "C" { #endif @@ -418,7 +405,7 @@ const char *nc_utf8proc_errmsg(nc_utf8proc_ssize_t errcode); * In case of success, the number of bytes read is returned; otherwise, a * negative error code is returned. */ -UTF8PROC_DLLEXPORT nc_utf8proc_ssize_t nc_utf8proc_iterate(const nc_utf8proc_uint8_t *str, nc_utf8proc_ssize_t strlen, nc_utf8proc_int32_t *codepoint_ref); +EXTERNL nc_utf8proc_ssize_t nc_utf8proc_iterate(const nc_utf8proc_uint8_t *str, nc_utf8proc_ssize_t strlen, nc_utf8proc_int32_t *codepoint_ref); /** * Check if a codepoint is valid (regardless of whether it has been @@ -685,13 +672,13 @@ nc_utf8proc_ssize_t nc_utf8proc_map_custom( */ /** @{ */ /** NFD normalization (@ref UTF8PROC_DECOMPOSE). */ -UTF8PROC_DLLEXPORT nc_utf8proc_uint8_t *nc_utf8proc_NFD(const nc_utf8proc_uint8_t *str); +EXTERNL nc_utf8proc_uint8_t *nc_utf8proc_NFD(const nc_utf8proc_uint8_t *str); /** NFC normalization (@ref UTF8PROC_COMPOSE). */ -UTF8PROC_DLLEXPORT nc_utf8proc_uint8_t *nc_utf8proc_NFC(const nc_utf8proc_uint8_t *str); +EXTERNL nc_utf8proc_uint8_t *nc_utf8proc_NFC(const nc_utf8proc_uint8_t *str); /** NFKD normalization (@ref UTF8PROC_DECOMPOSE and @ref UTF8PROC_COMPAT). */ -UTF8PROC_DLLEXPORT nc_utf8proc_uint8_t *nc_utf8proc_NFKD(const nc_utf8proc_uint8_t *str); +EXTERNL nc_utf8proc_uint8_t *nc_utf8proc_NFKD(const nc_utf8proc_uint8_t *str); /** NFKC normalization (@ref UTF8PROC_COMPOSE and @ref UTF8PROC_COMPAT). */ -UTF8PROC_DLLEXPORT nc_utf8proc_uint8_t *nc_utf8proc_NFKC(const nc_utf8proc_uint8_t *str); +EXTERNL nc_utf8proc_uint8_t *nc_utf8proc_NFKC(const nc_utf8proc_uint8_t *str); /** @} */ #ifdef __cplusplus