From f73d5c238fc7defe88a594fa2954f5c2cf5d31f0 Mon Sep 17 00:00:00 2001 From: Michaël Zasso Date: Wed, 24 Apr 2019 17:33:26 +0200 Subject: deps: bump minimum icu version to 64 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Bump minimum version of ICU needed to build node to 64. Refs: https://github.com/v8/v8/commit/ccbe3d07fbac807e15c524257f3e79e04492f9cf PR-URL: https://github.com/nodejs/node/pull/27375 Reviewed-By: Michaël Zasso Reviewed-By: Ujjwal Sharma Reviewed-By: Refael Ackermann Reviewed-By: Matteo Collina Reviewed-By: Colin Ihrig Reviewed-By: Rich Trott --- tools/icu/icu_versions.json | 2 +- tools/icu/patches/54/source/io/ufile.c | 360 --- tools/icu/patches/55/source/io/ufile.c | 360 --- tools/icu/patches/58/source/i18n/digitlst.cpp | 1095 -------- tools/icu/patches/62/source/i18n/decimfmt.cpp | 1384 ---------- tools/icu/patches/63/source/i18n/dtptngen.cpp | 2752 -------------------- .../patches/63/source/tools/toolutil/pkg_genc.cpp | 1221 --------- 7 files changed, 1 insertion(+), 7173 deletions(-) delete mode 100644 tools/icu/patches/54/source/io/ufile.c delete mode 100644 tools/icu/patches/55/source/io/ufile.c delete mode 100644 tools/icu/patches/58/source/i18n/digitlst.cpp delete mode 100644 tools/icu/patches/62/source/i18n/decimfmt.cpp delete mode 100644 tools/icu/patches/63/source/i18n/dtptngen.cpp delete mode 100644 tools/icu/patches/63/source/tools/toolutil/pkg_genc.cpp (limited to 'tools') diff --git a/tools/icu/icu_versions.json b/tools/icu/icu_versions.json index 7224e9debe..e6929ac6a3 100644 --- a/tools/icu/icu_versions.json +++ b/tools/icu/icu_versions.json @@ -1,3 +1,3 @@ { - "minimum_icu": 63 + "minimum_icu": 64 } diff --git a/tools/icu/patches/54/source/io/ufile.c b/tools/icu/patches/54/source/io/ufile.c deleted file mode 100644 index ab9f70a3d5..0000000000 --- a/tools/icu/patches/54/source/io/ufile.c +++ /dev/null @@ -1,360 +0,0 @@ -/* -****************************************************************************** -* -* Copyright (C) 1998-2015, International Business Machines -* Corporation and others. All Rights Reserved. -* -****************************************************************************** -* -* File ufile.c -* -* Modification History: -* -* Date Name Description -* 11/19/98 stephen Creation. -* 03/12/99 stephen Modified for new C API. -* 06/16/99 stephen Changed T_LocaleBundle to u_locbund -* 07/19/99 stephen Fixed to use ucnv's default codepage. -****************************************************************************** -*/ - -/* - * fileno is not declared when building with GCC in strict mode. - */ -#if defined(__GNUC__) && defined(__STRICT_ANSI__) -#undef __STRICT_ANSI__ -#endif - -#include "locmap.h" -#include "unicode/ustdio.h" - -#if !UCONFIG_NO_CONVERSION - -#include "ufile.h" -#include "unicode/uloc.h" -#include "unicode/ures.h" -#include "unicode/ucnv.h" -#include "unicode/ustring.h" -#include "cstring.h" -#include "cmemory.h" - -#if U_PLATFORM_USES_ONLY_WIN32_API && !defined(fileno) -/* Windows likes to rename Unix-like functions */ -#define fileno _fileno -#endif - -static UFILE* -finit_owner(FILE *f, - const char *locale, - const char *codepage, - UBool takeOwnership - ) -{ - UErrorCode status = U_ZERO_ERROR; - UFILE *result; - if(f == NULL) { - return 0; - } - result = (UFILE*) uprv_malloc(sizeof(UFILE)); - if(result == NULL) { - return 0; - } - - uprv_memset(result, 0, sizeof(UFILE)); - result->fFileno = fileno(f); - -#if U_PLATFORM_USES_ONLY_WIN32_API && _MSC_VER < 1900 - /* - * Below is a very old workaround (ICU ticket:231). - * - * Previously, 'FILE*' from inside and outside ICU's DLL - * were different, because they pointed into local copies - * of the io block. At least by VS 2015 the implementation - * is something like: - * stdio = _acrt_iob_func(0) - * .. which is a function call, so should return the same pointer - * regardless of call site. - * As of _MSC_VER 1900 this patch is retired, at 16 years old. - */ - if (0 <= result->fFileno && result->fFileno <= 2) { - /* stdin, stdout and stderr need to be special cased for Windows 98 */ -#if _MSC_VER >= 1400 - result->fFile = &__iob_func()[_fileno(f)]; -#else - result->fFile = &_iob[_fileno(f)]; -#endif - } - else -#endif - { - result->fFile = f; - } - - result->str.fBuffer = result->fUCBuffer; - result->str.fPos = result->fUCBuffer; - result->str.fLimit = result->fUCBuffer; - -#if !UCONFIG_NO_FORMATTING - /* if locale is 0, use the default */ - if(u_locbund_init(&result->str.fBundle, locale) == 0) { - /* DO NOT FCLOSE HERE! */ - uprv_free(result); - return 0; - } -#endif - - /* If the codepage is not "" use the ucnv_open default behavior */ - if(codepage == NULL || *codepage != '\0') { - result->fConverter = ucnv_open(codepage, &status); - } - /* else result->fConverter is already memset'd to NULL. */ - - if(U_SUCCESS(status)) { - result->fOwnFile = takeOwnership; - } - else { -#if !UCONFIG_NO_FORMATTING - u_locbund_close(&result->str.fBundle); -#endif - /* DO NOT fclose here!!!!!! */ - uprv_free(result); - result = NULL; - } - - return result; -} - -U_CAPI UFILE* U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_finit(FILE *f, - const char *locale, - const char *codepage) -{ - return finit_owner(f, locale, codepage, FALSE); -} - -U_CAPI UFILE* U_EXPORT2 -u_fadopt(FILE *f, - const char *locale, - const char *codepage) -{ - return finit_owner(f, locale, codepage, TRUE); -} - -U_CAPI UFILE* U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fopen(const char *filename, - const char *perm, - const char *locale, - const char *codepage) -{ - UFILE *result; - FILE *systemFile = fopen(filename, perm); - if(systemFile == 0) { - return 0; - } - - result = finit_owner(systemFile, locale, codepage, TRUE); - - if (!result) { - /* Something bad happened. - Maybe the converter couldn't be opened. */ - fclose(systemFile); - } - - return result; /* not a file leak */ -} - -U_CAPI UFILE* U_EXPORT2 -u_fopen_u(const UChar *filename, - const char *perm, - const char *locale, - const char *codepage) -{ - UFILE *result; - char buffer[256]; - - u_austrcpy(buffer, filename); - - result = u_fopen(buffer, perm, locale, codepage); -#if U_PLATFORM_USES_ONLY_WIN32_API - /* Try Windows API _wfopen if the above fails. */ - if (!result) { - FILE *systemFile = _wfopen(filename, (UChar*)perm); - if (systemFile) { - result = finit_owner(systemFile, locale, codepage, TRUE); - } - if (!result) { - /* Something bad happened. - Maybe the converter couldn't be opened. */ - fclose(systemFile); - } - } -#endif - return result; /* not a file leak */ -} - -U_CAPI UFILE* U_EXPORT2 -u_fstropen(UChar *stringBuf, - int32_t capacity, - const char *locale) -{ - UFILE *result; - - if (capacity < 0) { - return NULL; - } - - result = (UFILE*) uprv_malloc(sizeof(UFILE)); - /* Null pointer test */ - if (result == NULL) { - return NULL; /* Just get out. */ - } - uprv_memset(result, 0, sizeof(UFILE)); - result->str.fBuffer = stringBuf; - result->str.fPos = stringBuf; - result->str.fLimit = stringBuf+capacity; - -#if !UCONFIG_NO_FORMATTING - /* if locale is 0, use the default */ - if(u_locbund_init(&result->str.fBundle, locale) == 0) { - /* DO NOT FCLOSE HERE! */ - uprv_free(result); - return 0; - } -#endif - - return result; -} - -U_CAPI UBool U_EXPORT2 -u_feof(UFILE *f) -{ - UBool endOfBuffer; - if (f == NULL) { - return TRUE; - } - endOfBuffer = (UBool)(f->str.fPos >= f->str.fLimit); - if (f->fFile != NULL) { - return endOfBuffer && feof(f->fFile); - } - return endOfBuffer; -} - -U_CAPI void U_EXPORT2 -u_fflush(UFILE *file) -{ - ufile_flush_translit(file); - ufile_flush_io(file); - if (file->fFile) { - fflush(file->fFile); - } - else if (file->str.fPos < file->str.fLimit) { - *(file->str.fPos++) = 0; - } - /* TODO: flush input */ -} - -U_CAPI void -u_frewind(UFILE *file) -{ - u_fflush(file); - ucnv_reset(file->fConverter); - if (file->fFile) { - rewind(file->fFile); - file->str.fLimit = file->fUCBuffer; - file->str.fPos = file->fUCBuffer; - } - else { - file->str.fPos = file->str.fBuffer; - } -} - -U_CAPI void U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fclose(UFILE *file) -{ - if (file) { - u_fflush(file); - ufile_close_translit(file); - - if(file->fOwnFile) - fclose(file->fFile); - -#if !UCONFIG_NO_FORMATTING - u_locbund_close(&file->str.fBundle); -#endif - - ucnv_close(file->fConverter); - uprv_free(file); - } -} - -U_CAPI FILE* U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fgetfile( UFILE *f) -{ - return f->fFile; -} - -#if !UCONFIG_NO_FORMATTING - -U_CAPI const char* U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fgetlocale( UFILE *file) -{ - return file->str.fBundle.fLocale; -} - -U_CAPI int32_t U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fsetlocale(UFILE *file, - const char *locale) -{ - u_locbund_close(&file->str.fBundle); - - return u_locbund_init(&file->str.fBundle, locale) == 0 ? -1 : 0; -} - -#endif - -U_CAPI const char* U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fgetcodepage(UFILE *file) -{ - UErrorCode status = U_ZERO_ERROR; - const char *codepage = NULL; - - if (file->fConverter) { - codepage = ucnv_getName(file->fConverter, &status); - if(U_FAILURE(status)) - return 0; - } - return codepage; -} - -U_CAPI int32_t U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fsetcodepage( const char *codepage, - UFILE *file) -{ - UErrorCode status = U_ZERO_ERROR; - int32_t retVal = -1; - - /* We use the normal default codepage for this system, and not the one for the locale. */ - if ((file->str.fPos == file->str.fBuffer) && (file->str.fLimit == file->str.fBuffer)) { - ucnv_close(file->fConverter); - file->fConverter = ucnv_open(codepage, &status); - if(U_SUCCESS(status)) { - retVal = 0; - } - } - return retVal; -} - - -U_CAPI UConverter * U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fgetConverter(UFILE *file) -{ - return file->fConverter; -} -#if !UCONFIG_NO_FORMATTING -U_CAPI const UNumberFormat* U_EXPORT2 u_fgetNumberFormat(UFILE *file) -{ - return u_locbund_getNumberFormat(&file->str.fBundle, UNUM_DECIMAL); -} -#endif - -#endif diff --git a/tools/icu/patches/55/source/io/ufile.c b/tools/icu/patches/55/source/io/ufile.c deleted file mode 100644 index ab9f70a3d5..0000000000 --- a/tools/icu/patches/55/source/io/ufile.c +++ /dev/null @@ -1,360 +0,0 @@ -/* -****************************************************************************** -* -* Copyright (C) 1998-2015, International Business Machines -* Corporation and others. All Rights Reserved. -* -****************************************************************************** -* -* File ufile.c -* -* Modification History: -* -* Date Name Description -* 11/19/98 stephen Creation. -* 03/12/99 stephen Modified for new C API. -* 06/16/99 stephen Changed T_LocaleBundle to u_locbund -* 07/19/99 stephen Fixed to use ucnv's default codepage. -****************************************************************************** -*/ - -/* - * fileno is not declared when building with GCC in strict mode. - */ -#if defined(__GNUC__) && defined(__STRICT_ANSI__) -#undef __STRICT_ANSI__ -#endif - -#include "locmap.h" -#include "unicode/ustdio.h" - -#if !UCONFIG_NO_CONVERSION - -#include "ufile.h" -#include "unicode/uloc.h" -#include "unicode/ures.h" -#include "unicode/ucnv.h" -#include "unicode/ustring.h" -#include "cstring.h" -#include "cmemory.h" - -#if U_PLATFORM_USES_ONLY_WIN32_API && !defined(fileno) -/* Windows likes to rename Unix-like functions */ -#define fileno _fileno -#endif - -static UFILE* -finit_owner(FILE *f, - const char *locale, - const char *codepage, - UBool takeOwnership - ) -{ - UErrorCode status = U_ZERO_ERROR; - UFILE *result; - if(f == NULL) { - return 0; - } - result = (UFILE*) uprv_malloc(sizeof(UFILE)); - if(result == NULL) { - return 0; - } - - uprv_memset(result, 0, sizeof(UFILE)); - result->fFileno = fileno(f); - -#if U_PLATFORM_USES_ONLY_WIN32_API && _MSC_VER < 1900 - /* - * Below is a very old workaround (ICU ticket:231). - * - * Previously, 'FILE*' from inside and outside ICU's DLL - * were different, because they pointed into local copies - * of the io block. At least by VS 2015 the implementation - * is something like: - * stdio = _acrt_iob_func(0) - * .. which is a function call, so should return the same pointer - * regardless of call site. - * As of _MSC_VER 1900 this patch is retired, at 16 years old. - */ - if (0 <= result->fFileno && result->fFileno <= 2) { - /* stdin, stdout and stderr need to be special cased for Windows 98 */ -#if _MSC_VER >= 1400 - result->fFile = &__iob_func()[_fileno(f)]; -#else - result->fFile = &_iob[_fileno(f)]; -#endif - } - else -#endif - { - result->fFile = f; - } - - result->str.fBuffer = result->fUCBuffer; - result->str.fPos = result->fUCBuffer; - result->str.fLimit = result->fUCBuffer; - -#if !UCONFIG_NO_FORMATTING - /* if locale is 0, use the default */ - if(u_locbund_init(&result->str.fBundle, locale) == 0) { - /* DO NOT FCLOSE HERE! */ - uprv_free(result); - return 0; - } -#endif - - /* If the codepage is not "" use the ucnv_open default behavior */ - if(codepage == NULL || *codepage != '\0') { - result->fConverter = ucnv_open(codepage, &status); - } - /* else result->fConverter is already memset'd to NULL. */ - - if(U_SUCCESS(status)) { - result->fOwnFile = takeOwnership; - } - else { -#if !UCONFIG_NO_FORMATTING - u_locbund_close(&result->str.fBundle); -#endif - /* DO NOT fclose here!!!!!! */ - uprv_free(result); - result = NULL; - } - - return result; -} - -U_CAPI UFILE* U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_finit(FILE *f, - const char *locale, - const char *codepage) -{ - return finit_owner(f, locale, codepage, FALSE); -} - -U_CAPI UFILE* U_EXPORT2 -u_fadopt(FILE *f, - const char *locale, - const char *codepage) -{ - return finit_owner(f, locale, codepage, TRUE); -} - -U_CAPI UFILE* U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fopen(const char *filename, - const char *perm, - const char *locale, - const char *codepage) -{ - UFILE *result; - FILE *systemFile = fopen(filename, perm); - if(systemFile == 0) { - return 0; - } - - result = finit_owner(systemFile, locale, codepage, TRUE); - - if (!result) { - /* Something bad happened. - Maybe the converter couldn't be opened. */ - fclose(systemFile); - } - - return result; /* not a file leak */ -} - -U_CAPI UFILE* U_EXPORT2 -u_fopen_u(const UChar *filename, - const char *perm, - const char *locale, - const char *codepage) -{ - UFILE *result; - char buffer[256]; - - u_austrcpy(buffer, filename); - - result = u_fopen(buffer, perm, locale, codepage); -#if U_PLATFORM_USES_ONLY_WIN32_API - /* Try Windows API _wfopen if the above fails. */ - if (!result) { - FILE *systemFile = _wfopen(filename, (UChar*)perm); - if (systemFile) { - result = finit_owner(systemFile, locale, codepage, TRUE); - } - if (!result) { - /* Something bad happened. - Maybe the converter couldn't be opened. */ - fclose(systemFile); - } - } -#endif - return result; /* not a file leak */ -} - -U_CAPI UFILE* U_EXPORT2 -u_fstropen(UChar *stringBuf, - int32_t capacity, - const char *locale) -{ - UFILE *result; - - if (capacity < 0) { - return NULL; - } - - result = (UFILE*) uprv_malloc(sizeof(UFILE)); - /* Null pointer test */ - if (result == NULL) { - return NULL; /* Just get out. */ - } - uprv_memset(result, 0, sizeof(UFILE)); - result->str.fBuffer = stringBuf; - result->str.fPos = stringBuf; - result->str.fLimit = stringBuf+capacity; - -#if !UCONFIG_NO_FORMATTING - /* if locale is 0, use the default */ - if(u_locbund_init(&result->str.fBundle, locale) == 0) { - /* DO NOT FCLOSE HERE! */ - uprv_free(result); - return 0; - } -#endif - - return result; -} - -U_CAPI UBool U_EXPORT2 -u_feof(UFILE *f) -{ - UBool endOfBuffer; - if (f == NULL) { - return TRUE; - } - endOfBuffer = (UBool)(f->str.fPos >= f->str.fLimit); - if (f->fFile != NULL) { - return endOfBuffer && feof(f->fFile); - } - return endOfBuffer; -} - -U_CAPI void U_EXPORT2 -u_fflush(UFILE *file) -{ - ufile_flush_translit(file); - ufile_flush_io(file); - if (file->fFile) { - fflush(file->fFile); - } - else if (file->str.fPos < file->str.fLimit) { - *(file->str.fPos++) = 0; - } - /* TODO: flush input */ -} - -U_CAPI void -u_frewind(UFILE *file) -{ - u_fflush(file); - ucnv_reset(file->fConverter); - if (file->fFile) { - rewind(file->fFile); - file->str.fLimit = file->fUCBuffer; - file->str.fPos = file->fUCBuffer; - } - else { - file->str.fPos = file->str.fBuffer; - } -} - -U_CAPI void U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fclose(UFILE *file) -{ - if (file) { - u_fflush(file); - ufile_close_translit(file); - - if(file->fOwnFile) - fclose(file->fFile); - -#if !UCONFIG_NO_FORMATTING - u_locbund_close(&file->str.fBundle); -#endif - - ucnv_close(file->fConverter); - uprv_free(file); - } -} - -U_CAPI FILE* U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fgetfile( UFILE *f) -{ - return f->fFile; -} - -#if !UCONFIG_NO_FORMATTING - -U_CAPI const char* U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fgetlocale( UFILE *file) -{ - return file->str.fBundle.fLocale; -} - -U_CAPI int32_t U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fsetlocale(UFILE *file, - const char *locale) -{ - u_locbund_close(&file->str.fBundle); - - return u_locbund_init(&file->str.fBundle, locale) == 0 ? -1 : 0; -} - -#endif - -U_CAPI const char* U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fgetcodepage(UFILE *file) -{ - UErrorCode status = U_ZERO_ERROR; - const char *codepage = NULL; - - if (file->fConverter) { - codepage = ucnv_getName(file->fConverter, &status); - if(U_FAILURE(status)) - return 0; - } - return codepage; -} - -U_CAPI int32_t U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fsetcodepage( const char *codepage, - UFILE *file) -{ - UErrorCode status = U_ZERO_ERROR; - int32_t retVal = -1; - - /* We use the normal default codepage for this system, and not the one for the locale. */ - if ((file->str.fPos == file->str.fBuffer) && (file->str.fLimit == file->str.fBuffer)) { - ucnv_close(file->fConverter); - file->fConverter = ucnv_open(codepage, &status); - if(U_SUCCESS(status)) { - retVal = 0; - } - } - return retVal; -} - - -U_CAPI UConverter * U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */ -u_fgetConverter(UFILE *file) -{ - return file->fConverter; -} -#if !UCONFIG_NO_FORMATTING -U_CAPI const UNumberFormat* U_EXPORT2 u_fgetNumberFormat(UFILE *file) -{ - return u_locbund_getNumberFormat(&file->str.fBundle, UNUM_DECIMAL); -} -#endif - -#endif diff --git a/tools/icu/patches/58/source/i18n/digitlst.cpp b/tools/icu/patches/58/source/i18n/digitlst.cpp deleted file mode 100644 index 9711a6cc22..0000000000 --- a/tools/icu/patches/58/source/i18n/digitlst.cpp +++ /dev/null @@ -1,1095 +0,0 @@ -// Copyright (C) 2016 and later: Unicode, Inc. and others. -// License & terms of use: http://www.unicode.org/copyright.html -/* -********************************************************************** -* Copyright (C) 1997-2015, International Business Machines -* Corporation and others. All Rights Reserved. -********************************************************************** -* -* File DIGITLST.CPP -* -* Modification History: -* -* Date Name Description -* 03/21/97 clhuang Converted from java. -* 03/21/97 clhuang Implemented with new APIs. -* 03/27/97 helena Updated to pass the simple test after code review. -* 03/31/97 aliu Moved isLONG_MIN to here, and fixed it. -* 04/15/97 aliu Changed MAX_COUNT to DBL_DIG. Changed Digit to char. -* Reworked representation by replacing fDecimalAt -* with fExponent. -* 04/16/97 aliu Rewrote set() and getDouble() to use sprintf/atof -* to do digit conversion. -* 09/09/97 aliu Modified for exponential notation support. -* 08/02/98 stephen Added nearest/even rounding -* Fixed bug in fitsIntoLong -****************************************************************************** -*/ - -#if defined(__CYGWIN__) && !defined(_GNU_SOURCE) -#define _GNU_SOURCE -#endif - -#include "digitlst.h" - -#if !UCONFIG_NO_FORMATTING - -#include "unicode/putil.h" -#include "charstr.h" -#include "cmemory.h" -#include "cstring.h" -#include "mutex.h" -#include "putilimp.h" -#include "uassert.h" -#include "digitinterval.h" -#include "ucln_in.h" -#include "umutex.h" -#include -#include -#include -#include -#include - -#if !defined(U_USE_STRTOD_L) -# if U_PLATFORM_USES_ONLY_WIN32_API -# define U_USE_STRTOD_L 1 -# elif defined(U_HAVE_STRTOD_L) -# define U_USE_STRTOD_L U_HAVE_STRTOD_L -# else -# define U_USE_STRTOD_L 0 -# endif -#endif - -// Patch for http://bugs.icu-project.org/trac/ticket/12822 -#if U_USE_STRTOD_L -# if U_PLATFORM_USES_ONLY_WIN32_API || U_PLATFORM == U_PF_CYGWIN -# include -# else -# include -# endif -#endif - -// *************************************************************************** -// class DigitList -// A wrapper onto decNumber. -// Used to be standalone. -// *************************************************************************** - -/** - * This is the zero digit. The base for the digits returned by getDigit() - * Note that it is the platform invariant digit, and is not Unicode. - */ -#define kZero '0' - - -/* Only for 32 bit numbers. Ignore the negative sign. */ -//static const char LONG_MIN_REP[] = "2147483648"; -//static const char I64_MIN_REP[] = "9223372036854775808"; - - -U_NAMESPACE_BEGIN - -// ------------------------------------- -// default constructor - -DigitList::DigitList() -{ - uprv_decContextDefault(&fContext, DEC_INIT_BASE); - fContext.traps = 0; - uprv_decContextSetRounding(&fContext, DEC_ROUND_HALF_EVEN); - fContext.digits = fStorage.getCapacity(); - - fDecNumber = fStorage.getAlias(); - uprv_decNumberZero(fDecNumber); - - internalSetDouble(0.0); -} - -// ------------------------------------- - -DigitList::~DigitList() -{ -} - -// ------------------------------------- -// copy constructor - -DigitList::DigitList(const DigitList &other) -{ - fDecNumber = fStorage.getAlias(); - *this = other; -} - - -// ------------------------------------- -// assignment operator - -DigitList& -DigitList::operator=(const DigitList& other) -{ - if (this != &other) - { - uprv_memcpy(&fContext, &other.fContext, sizeof(decContext)); - - if (other.fStorage.getCapacity() > fStorage.getCapacity()) { - fDecNumber = fStorage.resize(other.fStorage.getCapacity()); - } - // Always reset the fContext.digits, even if fDecNumber was not reallocated, - // because above we copied fContext from other.fContext. - fContext.digits = fStorage.getCapacity(); - uprv_decNumberCopy(fDecNumber, other.fDecNumber); - - { - // fDouble is lazily created and cached. - // Avoid potential races with that happening with other.fDouble - // while we are doing the assignment. - Mutex mutex; - - if(other.fHave==kDouble) { - fUnion.fDouble = other.fUnion.fDouble; - } - fHave = other.fHave; - } - } - return *this; -} - -// ------------------------------------- -// operator == (does not exactly match the old DigitList function) - -UBool -DigitList::operator==(const DigitList& that) const -{ - if (this == &that) { - return TRUE; - } - decNumber n; // Has space for only a none digit value. - decContext c; - uprv_decContextDefault(&c, DEC_INIT_BASE); - c.digits = 1; - c.traps = 0; - - uprv_decNumberCompare(&n, this->fDecNumber, that.fDecNumber, &c); - UBool result = decNumberIsZero(&n); - return result; -} - -// ------------------------------------- -// comparison function. Returns -// Not Comparable : -2 -// < : -1 -// == : 0 -// > : +1 -int32_t DigitList::compare(const DigitList &other) { - decNumber result; - int32_t savedDigits = fContext.digits; - fContext.digits = 1; - uprv_decNumberCompare(&result, this->fDecNumber, other.fDecNumber, &fContext); - fContext.digits = savedDigits; - if (decNumberIsZero(&result)) { - return 0; - } else if (decNumberIsSpecial(&result)) { - return -2; - } else if (result.bits & DECNEG) { - return -1; - } else { - return 1; - } -} - - -// ------------------------------------- -// Reduce - remove trailing zero digits. -void -DigitList::reduce() { - uprv_decNumberReduce(fDecNumber, fDecNumber, &fContext); -} - - -// ------------------------------------- -// trim - remove trailing fraction zero digits. -void -DigitList::trim() { - uprv_decNumberTrim(fDecNumber); -} - -// ------------------------------------- -// Resets the digit list; sets all the digits to zero. - -void -DigitList::clear() -{ - uprv_decNumberZero(fDecNumber); - uprv_decContextSetRounding(&fContext, DEC_ROUND_HALF_EVEN); - internalSetDouble(0.0); -} - - -/** - * Formats a int64_t number into a base 10 string representation, and NULL terminates it. - * @param number The number to format - * @param outputStr The string to output to. Must be at least MAX_DIGITS+2 in length (21), - * to hold the longest int64_t value. - * @return the number of digits written, not including the sign. - */ -static int32_t -formatBase10(int64_t number, char *outputStr) { - // The number is output backwards, starting with the LSD. - // Fill the buffer from the far end. After the number is complete, - // slide the string contents to the front. - - const int32_t MAX_IDX = MAX_DIGITS+2; - int32_t destIdx = MAX_IDX; - outputStr[--destIdx] = 0; - - int64_t n = number; - if (number < 0) { // Negative numbers are slightly larger than a postive - outputStr[--destIdx] = (char)(-(n % 10) + kZero); - n /= -10; - } - do { - outputStr[--destIdx] = (char)(n % 10 + kZero); - n /= 10; - } while (n > 0); - - if (number < 0) { - outputStr[--destIdx] = '-'; - } - - // Slide the number to the start of the output str - U_ASSERT(destIdx >= 0); - int32_t length = MAX_IDX - destIdx; - uprv_memmove(outputStr, outputStr+MAX_IDX-length, length); - - return length; -} - - -// ------------------------------------- -// -// setRoundingMode() -// For most modes, the meaning and names are the same between the decNumber library -// (which DigitList follows) and the ICU Formatting Rounding Mode values. -// The flag constants are different, however. -// -// Note that ICU's kRoundingUnnecessary is not implemented directly by DigitList. -// This mode, inherited from Java, means that numbers that would not format exactly -// will return an error when formatting is attempted. - -void -DigitList::setRoundingMode(DecimalFormat::ERoundingMode m) { - enum rounding r; - - switch (m) { - case DecimalFormat::kRoundCeiling: r = DEC_ROUND_CEILING; break; - case DecimalFormat::kRoundFloor: r = DEC_ROUND_FLOOR; break; - case DecimalFormat::kRoundDown: r = DEC_ROUND_DOWN; break; - case DecimalFormat::kRoundUp: r = DEC_ROUND_UP; break; - case DecimalFormat::kRoundHalfEven: r = DEC_ROUND_HALF_EVEN; break; - case DecimalFormat::kRoundHalfDown: r = DEC_ROUND_HALF_DOWN; break; - case DecimalFormat::kRoundHalfUp: r = DEC_ROUND_HALF_UP; break; - case DecimalFormat::kRoundUnnecessary: r = DEC_ROUND_HALF_EVEN; break; - default: - // TODO: how to report the problem? - // Leave existing mode unchanged. - r = uprv_decContextGetRounding(&fContext); - } - uprv_decContextSetRounding(&fContext, r); - -} - - -// ------------------------------------- - -void -DigitList::setPositive(UBool s) { - if (s) { - fDecNumber->bits &= ~DECNEG; - } else { - fDecNumber->bits |= DECNEG; - } - internalClear(); -} -// ------------------------------------- - -void -DigitList::setDecimalAt(int32_t d) { - U_ASSERT((fDecNumber->bits & DECSPECIAL) == 0); // Not Infinity or NaN - U_ASSERT(d-1>-999999999); - U_ASSERT(d-1< 999999999); - int32_t adjustedDigits = fDecNumber->digits; - if (decNumberIsZero(fDecNumber)) { - // Account for difference in how zero is represented between DigitList & decNumber. - adjustedDigits = 0; - } - fDecNumber->exponent = d - adjustedDigits; - internalClear(); -} - -int32_t -DigitList::getDecimalAt() { - U_ASSERT((fDecNumber->bits & DECSPECIAL) == 0); // Not Infinity or NaN - if (decNumberIsZero(fDecNumber) || ((fDecNumber->bits & DECSPECIAL) != 0)) { - return fDecNumber->exponent; // Exponent should be zero for these cases. - } - return fDecNumber->exponent + fDecNumber->digits; -} - -void -DigitList::setCount(int32_t c) { - U_ASSERT(c <= fContext.digits); - if (c == 0) { - // For a value of zero, DigitList sets all fields to zero, while - // decNumber keeps one digit (with that digit being a zero) - c = 1; - fDecNumber->lsu[0] = 0; - } - fDecNumber->digits = c; - internalClear(); -} - -int32_t -DigitList::getCount() const { - if (decNumberIsZero(fDecNumber) && fDecNumber->exponent==0) { - // The extra test for exponent==0 is needed because parsing sometimes appends - // zero digits. It's bogus, decimalFormatter parsing needs to be cleaned up. - return 0; - } else { - return fDecNumber->digits; - } -} - -void -DigitList::setDigit(int32_t i, char v) { - int32_t count = fDecNumber->digits; - U_ASSERT(i='0' && v<='9'); - v &= 0x0f; - fDecNumber->lsu[count-i-1] = v; - internalClear(); -} - -char -DigitList::getDigit(int32_t i) { - int32_t count = fDecNumber->digits; - U_ASSERT(ilsu[count-i-1] + '0'; -} - -// copied from DigitList::getDigit() -uint8_t -DigitList::getDigitValue(int32_t i) { - int32_t count = fDecNumber->digits; - U_ASSERT(ilsu[count-i-1]; -} - -// ------------------------------------- -// Appends the digit to the digit list if it's not out of scope. -// Ignores the digit, otherwise. -// -// This function is horribly inefficient to implement with decNumber because -// the digits are stored least significant first, which requires moving all -// existing digits down one to make space for the new one to be appended. -// -void -DigitList::append(char digit) -{ - U_ASSERT(digit>='0' && digit<='9'); - // Ignore digits which exceed the precision we can represent - // And don't fix for larger precision. Fix callers instead. - if (decNumberIsZero(fDecNumber)) { - // Zero needs to be special cased because of the difference in the way - // that the old DigitList and decNumber represent it. - // digit cout was zero for digitList, is one for decNumber - fDecNumber->lsu[0] = digit & 0x0f; - fDecNumber->digits = 1; - fDecNumber->exponent--; // To match the old digit list implementation. - } else { - int32_t nDigits = fDecNumber->digits; - if (nDigits < fContext.digits) { - int i; - for (i=nDigits; i>0; i--) { - fDecNumber->lsu[i] = fDecNumber->lsu[i-1]; - } - fDecNumber->lsu[0] = digit & 0x0f; - fDecNumber->digits++; - // DigitList emulation - appending doesn't change the magnitude of existing - // digits. With decNumber's decimal being after the - // least signficant digit, we need to adjust the exponent. - fDecNumber->exponent--; - } - } - internalClear(); -} - -// ------------------------------------- - -/** - * Currently, getDouble() depends on strtod() to do its conversion. - * - * WARNING!! - * This is an extremely costly function. ~1/2 of the conversion time - * can be linked to this function. - */ -double -DigitList::getDouble() const -{ - { - Mutex mutex; - if (fHave == kDouble) { - return fUnion.fDouble; - } - } - - double tDouble = 0.0; - if (isZero()) { - tDouble = 0.0; - if (decNumberIsNegative(fDecNumber)) { - tDouble /= -1; - } - } else if (isInfinite()) { - if (std::numeric_limits::has_infinity) { - tDouble = std::numeric_limits::infinity(); - } else { - tDouble = std::numeric_limits::max(); - } - if (!isPositive()) { - tDouble = -tDouble; //this was incorrectly "-fDouble" originally. - } - } else { - MaybeStackArray s; - // Note: 14 is a magic constant from the decNumber library documentation, - // the max number of extra characters beyond the number of digits - // needed to represent the number in string form. Add a few more - // for the additional digits we retain. - - // Round down to appx. double precision, if the number is longer than that. - // Copy the number first, so that we don't modify the original. - if (getCount() > MAX_DBL_DIGITS + 3) { - DigitList numToConvert(*this); - numToConvert.reduce(); // Removes any trailing zeros, so that digit count is good. - numToConvert.round(MAX_DBL_DIGITS+3); - uprv_decNumberToString(numToConvert.fDecNumber, s.getAlias()); - // TODO: how many extra digits should be included for an accurate conversion? - } else { - uprv_decNumberToString(this->fDecNumber, s.getAlias()); - } - U_ASSERT(uprv_strlen(&s[0]) < MAX_DBL_DIGITS+18); - - char *end = NULL; - tDouble = decimalStrToDouble(s.getAlias(), &end); - } - { - Mutex mutex; - DigitList *nonConstThis = const_cast(this); - nonConstThis->internalSetDouble(tDouble); - } - return tDouble; -} - -#if U_USE_STRTOD_L && U_PLATFORM_USES_ONLY_WIN32_API -# define locale_t _locale_t -# define freelocale _free_locale -# define strtod_l _strtod_l -#endif - -#if U_USE_STRTOD_L -static locale_t gCLocale = (locale_t)0; -#endif -static icu::UInitOnce gCLocaleInitOnce = U_INITONCE_INITIALIZER; - -U_CDECL_BEGIN -// Cleanup callback func -static UBool U_CALLCONV digitList_cleanup(void) -{ -#if U_USE_STRTOD_L - if (gCLocale != (locale_t)0) { - freelocale(gCLocale); - } -#endif - return TRUE; -} -// C Locale initialization func -static void U_CALLCONV initCLocale(void) { - ucln_i18n_registerCleanup(UCLN_I18N_DIGITLIST, digitList_cleanup); -#if U_USE_STRTOD_L -# if U_PLATFORM_USES_ONLY_WIN32_API - gCLocale = _create_locale(LC_ALL, "C"); -# else - gCLocale = newlocale(LC_ALL_MASK, "C", (locale_t)0); -# endif -#endif -} -U_CDECL_END - -double -DigitList::decimalStrToDouble(char *decstr, char **end) { - umtx_initOnce(gCLocaleInitOnce, &initCLocale); -#if U_USE_STRTOD_L - return strtod_l(decstr, end, gCLocale); -#else - char *decimalPt = strchr(decstr, '.'); - if (decimalPt) { - // We need to know the decimal separator character that will be used with strtod(). - // Depends on the C runtime global locale. - // Most commonly is '.' - char rep[MAX_DIGITS]; - sprintf(rep, "%+1.1f", 1.0); - *decimalPt = rep[2]; - } - return uprv_strtod(decstr, end); -#endif -} - -// ------------------------------------- - -/** - * convert this number to an int32_t. Round if there is a fractional part. - * Return zero if the number cannot be represented. - */ -int32_t DigitList::getLong() /*const*/ -{ - int32_t result = 0; - if (getUpperExponent() > 10) { - // Overflow, absolute value too big. - return result; - } - if (fDecNumber->exponent != 0) { - // Force to an integer, with zero exponent, rounding if necessary. - // (decNumberToInt32 will only work if the exponent is exactly zero.) - DigitList copy(*this); - DigitList zero; - uprv_decNumberQuantize(copy.fDecNumber, copy.fDecNumber, zero.fDecNumber, &fContext); - result = uprv_decNumberToInt32(copy.fDecNumber, &fContext); - } else { - result = uprv_decNumberToInt32(fDecNumber, &fContext); - } - return result; -} - - -/** - * convert this number to an int64_t. Truncate if there is a fractional part. - * Return zero if the number cannot be represented. - */ -int64_t DigitList::getInt64() /*const*/ { - // TODO: fast conversion if fHave == fDouble - - // Truncate if non-integer. - // Return 0 if out of range. - // Range of in64_t is -9223372036854775808 to 9223372036854775807 (19 digits) - // - if (getUpperExponent() > 19) { - // Overflow, absolute value too big. - return 0; - } - - // The number of integer digits may differ from the number of digits stored - // in the decimal number. - // for 12.345 numIntDigits = 2, number->digits = 5 - // for 12E4 numIntDigits = 6, number->digits = 2 - // The conversion ignores the fraction digits in the first case, - // and fakes up extra zero digits in the second. - // TODO: It would be faster to store a table of powers of ten to multiply by - // instead of looping over zero digits, multiplying each time. - - int32_t numIntDigits = getUpperExponent(); - uint64_t value = 0; - for (int32_t i = 0; i < numIntDigits; i++) { - // Loop is iterating over digits starting with the most significant. - // Numbers are stored with the least significant digit at index zero. - int32_t digitIndex = fDecNumber->digits - i - 1; - int32_t v = (digitIndex >= 0) ? fDecNumber->lsu[digitIndex] : 0; - value = value * (uint64_t)10 + (uint64_t)v; - } - - if (decNumberIsNegative(fDecNumber)) { - value = ~value; - value += 1; - } - int64_t svalue = (int64_t)value; - - // Check overflow. It's convenient that the MSD is 9 only on overflow, the amount of - // overflow can't wrap too far. The test will also fail -0, but - // that does no harm; the right answer is 0. - if (numIntDigits == 19) { - if (( decNumberIsNegative(fDecNumber) && svalue>0) || - (!decNumberIsNegative(fDecNumber) && svalue<0)) { - svalue = 0; - } - } - - return svalue; -} - - -/** - * Return a string form of this number. - * Format is as defined by the decNumber library, for interchange of - * decimal numbers. - */ -void DigitList::getDecimal(CharString &str, UErrorCode &status) { - if (U_FAILURE(status)) { - return; - } - - // A decimal number in string form can, worst case, be 14 characters longer - // than the number of digits. So says the decNumber library doc. - int32_t maxLength = fDecNumber->digits + 14; - int32_t capacity = 0; - char *buffer = str.clear().getAppendBuffer(maxLength, 0, capacity, status); - if (U_FAILURE(status)) { - return; // Memory allocation error on growing the string. - } - U_ASSERT(capacity >= maxLength); - uprv_decNumberToString(this->fDecNumber, buffer); - U_ASSERT((int32_t)uprv_strlen(buffer) <= maxLength); - str.append(buffer, -1, status); -} - -/** - * Return true if this is an integer value that can be held - * by an int32_t type. - */ -UBool -DigitList::fitsIntoLong(UBool ignoreNegativeZero) /*const*/ -{ - if (decNumberIsSpecial(this->fDecNumber)) { - // NaN or Infinity. Does not fit in int32. - return FALSE; - } - uprv_decNumberTrim(this->fDecNumber); - if (fDecNumber->exponent < 0) { - // Number contains fraction digits. - return FALSE; - } - if (decNumberIsZero(this->fDecNumber) && !ignoreNegativeZero && - (fDecNumber->bits & DECNEG) != 0) { - // Negative Zero, not ingored. Cannot represent as a long. - return FALSE; - } - if (getUpperExponent() < 10) { - // The number is 9 or fewer digits. - // The max and min int32 are 10 digts, so this number fits. - // This is the common case. - return TRUE; - } - - // TODO: Should cache these constants; construction is relatively costly. - // But not of huge consequence; they're only needed for 10 digit ints. - UErrorCode status = U_ZERO_ERROR; - DigitList min32; min32.set("-2147483648", status); - if (this->compare(min32) < 0) { - return FALSE; - } - DigitList max32; max32.set("2147483647", status); - if (this->compare(max32) > 0) { - return FALSE; - } - if (U_FAILURE(status)) { - return FALSE; - } - return true; -} - - - -/** - * Return true if the number represented by this object can fit into - * a long. - */ -UBool -DigitList::fitsIntoInt64(UBool ignoreNegativeZero) /*const*/ -{ - if (decNumberIsSpecial(this->fDecNumber)) { - // NaN or Infinity. Does not fit in int32. - return FALSE; - } - uprv_decNumberTrim(this->fDecNumber); - if (fDecNumber->exponent < 0) { - // Number contains fraction digits. - return FALSE; - } - if (decNumberIsZero(this->fDecNumber) && !ignoreNegativeZero && - (fDecNumber->bits & DECNEG) != 0) { - // Negative Zero, not ingored. Cannot represent as a long. - return FALSE; - } - if (getUpperExponent() < 19) { - // The number is 18 or fewer digits. - // The max and min int64 are 19 digts, so this number fits. - // This is the common case. - return TRUE; - } - - // TODO: Should cache these constants; construction is relatively costly. - // But not of huge consequence; they're only needed for 19 digit ints. - UErrorCode status = U_ZERO_ERROR; - DigitList min64; min64.set("-9223372036854775808", status); - if (this->compare(min64) < 0) { - return FALSE; - } - DigitList max64; max64.set("9223372036854775807", status); - if (this->compare(max64) > 0) { - return FALSE; - } - if (U_FAILURE(status)) { - return FALSE; - } - return true; -} - - -// ------------------------------------- - -void -DigitList::set(int32_t source) -{ - set((int64_t)source); - internalSetDouble(source); -} - -// ------------------------------------- -/** - * Set an int64, via decnumber - */ -void -DigitList::set(int64_t source) -{ - char str[MAX_DIGITS+2]; // Leave room for sign and trailing nul. - formatBase10(source, str); - U_ASSERT(uprv_strlen(str) < sizeof(str)); - - uprv_decNumberFromString(fDecNumber, str, &fContext); - internalSetDouble(static_cast(source)); -} - -// ------------------------------------- -/** - * Set the DigitList from a decimal number string. - * - * The incoming string _must_ be nul terminated, even though it is arriving - * as a StringPiece because that is what the decNumber library wants. - * We can get away with this for an internal function; it would not - * be acceptable for a public API. - */ -void -DigitList::set(StringPiece source, UErrorCode &status, uint32_t /*fastpathBits*/) { - if (U_FAILURE(status)) { - return; - } - -#if 0 - if(fastpathBits==(kFastpathOk|kNoDecimal)) { - int32_t size = source.size(); - const char *data = source.data(); - int64_t r = 0; - int64_t m = 1; - // fast parse - while(size>0) { - char ch = data[--size]; - if(ch=='+') { - break; - } else if(ch=='-') { - r = -r; - break; - } else { - int64_t d = ch-'0'; - //printf("CH[%d]=%c, %d, *=%d\n", size,ch, (int)d, (int)m); - r+=(d)*m; - m *= 10; - } - } - //printf("R=%d\n", r); - set(r); - } else -#endif - { - // Figure out a max number of digits to use during the conversion, and - // resize the number up if necessary. - int32_t numDigits = source.length(); - if (numDigits > fContext.digits) { - // fContext.digits == fStorage.getCapacity() - decNumber *t = fStorage.resize(numDigits, fStorage.getCapacity()); - if (t == NULL) { - status = U_MEMORY_ALLOCATION_ERROR; - return; - } - fDecNumber = t; - fContext.digits = numDigits; - } - - fContext.status = 0; - uprv_decNumberFromString(fDecNumber, source.data(), &fContext); - if ((fContext.status & DEC_Conversion_syntax) != 0) { - status = U_DECIMAL_NUMBER_SYNTAX_ERROR; - } - } - internalClear(); -} - -/** - * Set the digit list to a representation of the given double value. - * This method supports both fixed-point and exponential notation. - * @param source Value to be converted. - */ -void -DigitList::set(double source) -{ - // for now, simple implementation; later, do proper IEEE stuff - char rep[MAX_DIGITS + 8]; // Extra space for '+', '.', e+NNN, and '\0' (actually +8 is enough) - - // Generate a representation of the form /[+-][0-9].[0-9]+e[+-][0-9]+/ - // Can also generate /[+-]nan/ or /[+-]inf/ - // TODO: Use something other than sprintf() here, since it's behavior is somewhat platform specific. - // That is why infinity is special cased here. - if (uprv_isInfinite(source)) { - if (uprv_isNegativeInfinity(source)) { - uprv_strcpy(rep,"-inf"); // Handle negative infinity - } else { - uprv_strcpy(rep,"inf"); - } - } else { - sprintf(rep, "%+1.*e", MAX_DBL_DIGITS - 1, source); - } - U_ASSERT(uprv_strlen(rep) < sizeof(rep)); - - // uprv_decNumberFromString() will parse the string expecting '.' as a - // decimal separator, however sprintf() can use ',' in certain locales. - // Overwrite a ',' with '.' here before proceeding. - char *decimalSeparator = strchr(rep, ','); - if (decimalSeparator != NULL) { - *decimalSeparator = '.'; - } - - // Create a decNumber from the string. - uprv_decNumberFromString(fDecNumber, rep, &fContext); - uprv_decNumberTrim(fDecNumber); - internalSetDouble(source); -} - -// ------------------------------------- - -/* - * Multiply - * The number will be expanded if need be to retain full precision. - * In practice, for formatting, multiply is by 10, 100 or 1000, so more digits - * will not be required for this use. - */ -void -DigitList::mult(const DigitList &other, UErrorCode &status) { - if (U_FAILURE(status)) { - return; - } - fContext.status = 0; - int32_t requiredDigits = this->digits() + other.digits(); - if (requiredDigits > fContext.digits) { - reduce(); // Remove any trailing zeros - int32_t requiredDigits = this->digits() + other.digits(); - ensureCapacity(requiredDigits, status); - } - uprv_decNumberMultiply(fDecNumber, fDecNumber, other.fDecNumber, &fContext); - internalClear(); -} - -// ------------------------------------- - -/* - * Divide - * The number will _not_ be expanded for inexact results. - * TODO: probably should expand some, for rounding increments that - * could add a few digits, e.g. .25, but not expand arbitrarily. - */ -void -DigitList::div(const DigitList &other, UErrorCode &status) { - if (U_FAILURE(status)) { - return; - } - uprv_decNumberDivide(fDecNumber, fDecNumber, other.fDecNumber, &fContext); - internalClear(); -} - -// ------------------------------------- - -/* - * ensureCapacity. Grow the digit storage for the number if it's less than the requested - * amount. Never reduce it. Available size is kept in fContext.digits. - */ -void -DigitList::ensureCapacity(int32_t requestedCapacity, UErrorCode &status) { - if (U_FAILURE(status)) { - return; - } - if (requestedCapacity <= 0) { - status = U_ILLEGAL_ARGUMENT_ERROR; - return; - } - if (requestedCapacity > DEC_MAX_DIGITS) { - // Don't report an error for requesting too much. - // Arithemetic Results will be rounded to what can be supported. - // At 999,999,999 max digits, exceeding the limit is not too likely! - requestedCapacity = DEC_MAX_DIGITS; - } - if (requestedCapacity > fContext.digits) { - decNumber *newBuffer = fStorage.resize(requestedCapacity, fStorage.getCapacity()); - if (newBuffer == NULL) { - status = U_MEMORY_ALLOCATION_ERROR; - return; - } - fContext.digits = requestedCapacity; - fDecNumber = newBuffer; - } -} - -// ------------------------------------- - -/** - * Round the representation to the given number of digits. - * @param maximumDigits The maximum number of digits to be shown. - * Upon return, count will be less than or equal to maximumDigits. - */ -void -DigitList::round(int32_t maximumDigits) -{ - reduce(); - if (maximumDigits >= fDecNumber->digits) { - return; - } - int32_t savedDigits = fContext.digits; - fContext.digits = maximumDigits; - uprv_decNumberPlus(fDecNumber, fDecNumber, &fContext); - fContext.digits = savedDigits; - uprv_decNumberTrim(fDecNumber); - reduce(); - internalClear(); -} - - -void -DigitList::roundFixedPoint(int32_t maximumFractionDigits) { - reduce(); // Remove trailing zeros. - if (fDecNumber->exponent >= -maximumFractionDigits) { - return; - } - decNumber scale; // Dummy decimal number, but with the desired number of - uprv_decNumberZero(&scale); // fraction digits. - scale.exponent = -maximumFractionDigits; - scale.lsu[0] = 1; - - uprv_decNumberQuantize(fDecNumber, fDecNumber, &scale, &fContext); - reduce(); - internalClear(); -} - -// ------------------------------------- - -void -DigitList::toIntegralValue() { - uprv_decNumberToIntegralValue(fDecNumber, fDecNumber, &fContext); -} - - -// ------------------------------------- -UBool -DigitList::isZero() const -{ - return decNumberIsZero(fDecNumber); -} - -// ------------------------------------- -int32_t -DigitList::getUpperExponent() const { - return fDecNumber->digits + fDecNumber->exponent; -} - -DigitInterval & -DigitList::getSmallestInterval(DigitInterval &result) const { - result.setLeastSignificantInclusive(fDecNumber->exponent); - result.setMostSignificantExclusive(getUpperExponent()); - return result; -} - -uint8_t -DigitList::getDigitByExponent(int32_t exponent) const { - int32_t idx = exponent - fDecNumber->exponent; - if (idx < 0 || idx >= fDecNumber->digits) { - return 0; - } - return fDecNumber->lsu[idx]; -} - -void -DigitList::appendDigitsTo(CharString &str, UErrorCode &status) const { - str.append((const char *) fDecNumber->lsu, fDecNumber->digits, status); -} - -void -DigitList::roundAtExponent(int32_t exponent, int32_t maxSigDigits) { - reduce(); - if (maxSigDigits < fDecNumber->digits) { - int32_t minExponent = getUpperExponent() - maxSigDigits; - if (exponent < minExponent) { - exponent = minExponent; - } - } - if (exponent <= fDecNumber->exponent) { - return; - } - int32_t digits = getUpperExponent() - exponent; - if (digits > 0) { - round(digits); - } else { - roundFixedPoint(-exponent); - } -} - -void -DigitList::quantize(const DigitList &quantity, UErrorCode &status) { - if (U_FAILURE(status)) { - return; - } - div(quantity, status); - roundAtExponent(0); - mult(quantity, status); - reduce(); -} - -int32_t -DigitList::getScientificExponent( - int32_t minIntDigitCount, int32_t exponentMultiplier) const { - // The exponent for zero is always zero. - if (isZero()) { - return 0; - } - int32_t intDigitCount = getUpperExponent(); - int32_t exponent; - if (intDigitCount >= minIntDigitCount) { - int32_t maxAdjustment = intDigitCount - minIntDigitCount; - exponent = (maxAdjustment / exponentMultiplier) * exponentMultiplier; - } else { - int32_t minAdjustment = minIntDigitCount - intDigitCount; - exponent = ((minAdjustment + exponentMultiplier - 1) / exponentMultiplier) * -exponentMultiplier; - } - return exponent; -} - -int32_t -DigitList::toScientific( - int32_t minIntDigitCount, int32_t exponentMultiplier) { - int32_t exponent = getScientificExponent( - minIntDigitCount, exponentMultiplier); - shiftDecimalRight(-exponent); - return exponent; -} - -void -DigitList::shiftDecimalRight(int32_t n) { - fDecNumber->exponent += n; - internalClear(); -} - -U_NAMESPACE_END -#endif // #if !UCONFIG_NO_FORMATTING - -//eof diff --git a/tools/icu/patches/62/source/i18n/decimfmt.cpp b/tools/icu/patches/62/source/i18n/decimfmt.cpp deleted file mode 100644 index 8ae773b75c..0000000000 --- a/tools/icu/patches/62/source/i18n/decimfmt.cpp +++ /dev/null @@ -1,1384 +0,0 @@ -// © 2018 and later: Unicode, Inc. and others. -// License & terms of use: http://www.unicode.org/copyright.html - -#include "unicode/utypes.h" - -#if !UCONFIG_NO_FORMATTING - -// Allow implicit conversion from char16_t* to UnicodeString for this file: -// Helpful in toString methods and elsewhere. -#define UNISTR_FROM_STRING_EXPLICIT - -#include -#include -#include -#include "unicode/errorcode.h" -#include "unicode/decimfmt.h" -#include "number_decimalquantity.h" -#include "number_types.h" -#include "numparse_impl.h" -#include "number_mapper.h" -#include "number_patternstring.h" -#include "putilimp.h" -#include "number_utils.h" -#include "number_utypes.h" - -using namespace icu; -using namespace icu::number; -using namespace icu::number::impl; -using namespace icu::numparse; -using namespace icu::numparse::impl; -using ERoundingMode = icu::DecimalFormat::ERoundingMode; -using EPadPosition = icu::DecimalFormat::EPadPosition; - -// MSVC warns C4805 when comparing bool with UBool -// TODO: Move this macro into a better place? -#if U_PF_WINDOWS <= U_PLATFORM && U_PLATFORM <= U_PF_CYGWIN -#define UBOOL_TO_BOOL(b) static_cast(b) -#else -#define UBOOL_TO_BOOL(b) b -#endif - - -UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DecimalFormat) - - -DecimalFormat::DecimalFormat(UErrorCode& status) - : DecimalFormat(nullptr, status) { - // Use the default locale and decimal pattern. - const char* localeName = Locale::getDefault().getName(); - LocalPointer ns(NumberingSystem::createInstance(status)); - UnicodeString patternString = utils::getPatternForStyle( - localeName, - ns->getName(), - CLDR_PATTERN_STYLE_DECIMAL, - status); - setPropertiesFromPattern(patternString, IGNORE_ROUNDING_IF_CURRENCY, status); - touch(status); -} - -DecimalFormat::DecimalFormat(const UnicodeString& pattern, UErrorCode& status) - : DecimalFormat(nullptr, status) { - setPropertiesFromPattern(pattern, IGNORE_ROUNDING_IF_CURRENCY, status); - touch(status); -} - -DecimalFormat::DecimalFormat(const UnicodeString& pattern, DecimalFormatSymbols* symbolsToAdopt, - UErrorCode& status) - : DecimalFormat(symbolsToAdopt, status) { - setPropertiesFromPattern(pattern, IGNORE_ROUNDING_IF_CURRENCY, status); - touch(status); -} - -DecimalFormat::DecimalFormat(const UnicodeString& pattern, DecimalFormatSymbols* symbolsToAdopt, - UNumberFormatStyle style, UErrorCode& status) - : DecimalFormat(symbolsToAdopt, status) { - // If choice is a currency type, ignore the rounding information. - if (style == UNumberFormatStyle::UNUM_CURRENCY || style == UNumberFormatStyle::UNUM_CURRENCY_ISO || - style == UNumberFormatStyle::UNUM_CURRENCY_ACCOUNTING || - style == UNumberFormatStyle::UNUM_CASH_CURRENCY || - style == UNumberFormatStyle::UNUM_CURRENCY_STANDARD || - style == UNumberFormatStyle::UNUM_CURRENCY_PLURAL) { - setPropertiesFromPattern(pattern, IGNORE_ROUNDING_ALWAYS, status); - } else { - setPropertiesFromPattern(pattern, IGNORE_ROUNDING_IF_CURRENCY, status); - } - // Note: in Java, CurrencyPluralInfo is set in NumberFormat.java, but in C++, it is not set there, - // so we have to set it here. - if (style == UNumberFormatStyle::UNUM_CURRENCY_PLURAL) { - LocalPointer cpi( - new CurrencyPluralInfo(fields->symbols->getLocale(), status), - status); - if (U_FAILURE(status)) { return; } - fields->properties->currencyPluralInfo.fPtr.adoptInstead(cpi.orphan()); - } - touch(status); -} - -DecimalFormat::DecimalFormat(const DecimalFormatSymbols* symbolsToAdopt, UErrorCode& status) { - LocalPointer adoptedSymbols(symbolsToAdopt); - fields = new DecimalFormatFields(); - if (U_FAILURE(status)) { - return; - } - if (fields == nullptr) { - status = U_MEMORY_ALLOCATION_ERROR; - return; - } - fields->properties.adoptInsteadAndCheckErrorCode(new DecimalFormatProperties(), status); - fields->exportedProperties.adoptInsteadAndCheckErrorCode(new DecimalFormatProperties(), status); - if (adoptedSymbols.isNull()) { - fields->symbols.adoptInsteadAndCheckErrorCode(new DecimalFormatSymbols(status), status); - } else { - fields->symbols.adoptInsteadAndCheckErrorCode(adoptedSymbols.orphan(), status); - } -} - -#if UCONFIG_HAVE_PARSEALLINPUT - -void DecimalFormat::setParseAllInput(UNumberFormatAttributeValue value) { - if (value == fields->properties->parseAllInput) { return; } - fields->properties->parseAllInput = value; -} - -#endif - -DecimalFormat& -DecimalFormat::setAttribute(UNumberFormatAttribute attr, int32_t newValue, UErrorCode& status) { - if (U_FAILURE(status)) { return *this; } - - switch (attr) { - case UNUM_LENIENT_PARSE: - setLenient(newValue != 0); - break; - - case UNUM_PARSE_INT_ONLY: - setParseIntegerOnly(newValue != 0); - break; - - case UNUM_GROUPING_USED: - setGroupingUsed(newValue != 0); - break; - - case UNUM_DECIMAL_ALWAYS_SHOWN: - setDecimalSeparatorAlwaysShown(newValue != 0); - break; - - case UNUM_MAX_INTEGER_DIGITS: - setMaximumIntegerDigits(newValue); - break; - - case UNUM_MIN_INTEGER_DIGITS: - setMinimumIntegerDigits(newValue); - break; - - case UNUM_INTEGER_DIGITS: - setMinimumIntegerDigits(newValue); - setMaximumIntegerDigits(newValue); - break; - - case UNUM_MAX_FRACTION_DIGITS: - setMaximumFractionDigits(newValue); - break; - - case UNUM_MIN_FRACTION_DIGITS: - setMinimumFractionDigits(newValue); - break; - - case UNUM_FRACTION_DIGITS: - setMinimumFractionDigits(newValue); - setMaximumFractionDigits(newValue); - break; - - case UNUM_SIGNIFICANT_DIGITS_USED: - setSignificantDigitsUsed(newValue != 0); - break; - - case UNUM_MAX_SIGNIFICANT_DIGITS: - setMaximumSignificantDigits(newValue); - break; - - case UNUM_MIN_SIGNIFICANT_DIGITS: - setMinimumSignificantDigits(newValue); - break; - - case UNUM_MULTIPLIER: - setMultiplier(newValue); - break; - - case UNUM_SCALE: - setMultiplierScale(newValue); - break; - - case UNUM_GROUPING_SIZE: - setGroupingSize(newValue); - break; - - case UNUM_ROUNDING_MODE: - setRoundingMode((DecimalFormat::ERoundingMode) newValue); - break; - - case UNUM_FORMAT_WIDTH: - setFormatWidth(newValue); - break; - - case UNUM_PADDING_POSITION: - /** The position at which padding will take place. */ - setPadPosition((DecimalFormat::EPadPosition) newValue); - break; - - case UNUM_SECONDARY_GROUPING_SIZE: - setSecondaryGroupingSize(newValue); - break; - -#if UCONFIG_HAVE_PARSEALLINPUT - case UNUM_PARSE_ALL_INPUT: - setParseAllInput((UNumberFormatAttributeValue) newValue); - break; -#endif - - case UNUM_PARSE_NO_EXPONENT: - setParseNoExponent((UBool) newValue); - break; - - case UNUM_PARSE_DECIMAL_MARK_REQUIRED: - setDecimalPatternMatchRequired((UBool) newValue); - break; - - case UNUM_CURRENCY_USAGE: - setCurrencyUsage((UCurrencyUsage) newValue, &status); - break; - - case UNUM_MINIMUM_GROUPING_DIGITS: - setMinimumGroupingDigits(newValue); - break; - - case UNUM_PARSE_CASE_SENSITIVE: - setParseCaseSensitive(static_cast(newValue)); - break; - - case UNUM_SIGN_ALWAYS_SHOWN: - setSignAlwaysShown(static_cast(newValue)); - break; - - case UNUM_FORMAT_FAIL_IF_MORE_THAN_MAX_DIGITS: - setFormatFailIfMoreThanMaxDigits(static_cast(newValue)); - break; - - default: - status = U_UNSUPPORTED_ERROR; - break; - } - return *this; -} - -int32_t DecimalFormat::getAttribute(UNumberFormatAttribute attr, UErrorCode& status) const { - if (U_FAILURE(status)) { return -1; } - switch (attr) { - case UNUM_LENIENT_PARSE: - return isLenient(); - - case UNUM_PARSE_INT_ONLY: - return isParseIntegerOnly(); - - case UNUM_GROUPING_USED: - return isGroupingUsed(); - - case UNUM_DECIMAL_ALWAYS_SHOWN: - return isDecimalSeparatorAlwaysShown(); - - case UNUM_MAX_INTEGER_DIGITS: - return getMaximumIntegerDigits(); - - case UNUM_MIN_INTEGER_DIGITS: - return getMinimumIntegerDigits(); - - case UNUM_INTEGER_DIGITS: - // TBD: what should this return? - return getMinimumIntegerDigits(); - - case UNUM_MAX_FRACTION_DIGITS: - return getMaximumFractionDigits(); - - case UNUM_MIN_FRACTION_DIGITS: - return getMinimumFractionDigits(); - - case UNUM_FRACTION_DIGITS: - // TBD: what should this return? - return getMinimumFractionDigits(); - - case UNUM_SIGNIFICANT_DIGITS_USED: - return areSignificantDigitsUsed(); - - case UNUM_MAX_SIGNIFICANT_DIGITS: - return getMaximumSignificantDigits(); - - case UNUM_MIN_SIGNIFICANT_DIGITS: - return getMinimumSignificantDigits(); - - case UNUM_MULTIPLIER: - return getMultiplier(); - - case UNUM_SCALE: - return getMultiplierScale(); - - case UNUM_GROUPING_SIZE: - return getGroupingSize(); - - case UNUM_ROUNDING_MODE: - return getRoundingMode(); - - case UNUM_FORMAT_WIDTH: - return getFormatWidth(); - - case UNUM_PADDING_POSITION: - return getPadPosition(); - - case UNUM_SECONDARY_GROUPING_SIZE: - return getSecondaryGroupingSize(); - - case UNUM_PARSE_NO_EXPONENT: - return isParseNoExponent(); - - case UNUM_PARSE_DECIMAL_MARK_REQUIRED: - return isDecimalPatternMatchRequired(); - - case UNUM_CURRENCY_USAGE: - return getCurrencyUsage(); - - case UNUM_MINIMUM_GROUPING_DIGITS: - return getMinimumGroupingDigits(); - - case UNUM_PARSE_CASE_SENSITIVE: - return isParseCaseSensitive(); - - case UNUM_SIGN_ALWAYS_SHOWN: - return isSignAlwaysShown(); - - case UNUM_FORMAT_FAIL_IF_MORE_THAN_MAX_DIGITS: - return isFormatFailIfMoreThanMaxDigits(); - - default: - status = U_UNSUPPORTED_ERROR; - break; - } - - return -1; /* undefined */ -} - -void DecimalFormat::setGroupingUsed(UBool enabled) { - if (UBOOL_TO_BOOL(enabled) == fields->properties->groupingUsed) { return; } - NumberFormat::setGroupingUsed(enabled); // to set field for compatibility - fields->properties->groupingUsed = enabled; - touchNoError(); -} - -void DecimalFormat::setParseIntegerOnly(UBool value) { - if (UBOOL_TO_BOOL(value) == fields->properties->parseIntegerOnly) { return; } - NumberFormat::setParseIntegerOnly(value); // to set field for compatibility - fields->properties->parseIntegerOnly = value; - touchNoError(); -} - -void DecimalFormat::setLenient(UBool enable) { - ParseMode mode = enable ? PARSE_MODE_LENIENT : PARSE_MODE_STRICT; - if (!fields->properties->parseMode.isNull() && mode == fields->properties->parseMode.getNoError()) { return; } - NumberFormat::setLenient(enable); // to set field for compatibility - fields->properties->parseMode = mode; - touchNoError(); -} - -DecimalFormat::DecimalFormat(const UnicodeString& pattern, DecimalFormatSymbols* symbolsToAdopt, - UParseError&, UErrorCode& status) - : DecimalFormat(symbolsToAdopt, status) { - // TODO: What is parseError for? - setPropertiesFromPattern(pattern, IGNORE_ROUNDING_IF_CURRENCY, status); - touch(status); -} - -DecimalFormat::DecimalFormat(const UnicodeString& pattern, const DecimalFormatSymbols& symbols, - UErrorCode& status) - : DecimalFormat(new DecimalFormatSymbols(symbols), status) { - setPropertiesFromPattern(pattern, IGNORE_ROUNDING_IF_CURRENCY, status); - touch(status); -} - -DecimalFormat::DecimalFormat(const DecimalFormat& source) : NumberFormat(source) { - // Note: it is not safe to copy fields->formatter or fWarehouse directly because fields->formatter might have - // dangling pointers to fields inside fWarehouse. The safe thing is to re-construct fields->formatter from - // the property bag, despite being somewhat slower. - fields = new DecimalFormatFields(); - if (fields == nullptr) { - return; - } - fields->properties.adoptInstead(new DecimalFormatProperties(*source.fields->properties)); - fields->symbols.adoptInstead(new DecimalFormatSymbols(*source.fields->symbols)); - fields->exportedProperties.adoptInstead(new DecimalFormatProperties()); - if (fields->properties == nullptr || fields->symbols == nullptr || fields->exportedProperties == nullptr) { - return; - } - touchNoError(); -} - -DecimalFormat& DecimalFormat::operator=(const DecimalFormat& rhs) { - *fields->properties = *rhs.fields->properties; - fields->exportedProperties->clear(); - fields->symbols.adoptInstead(new DecimalFormatSymbols(*rhs.fields->symbols)); - touchNoError(); - return *this; -} - -DecimalFormat::~DecimalFormat() { - delete fields->atomicParser.exchange(nullptr); - delete fields->atomicCurrencyParser.exchange(nullptr); - delete fields; -} - -Format* DecimalFormat::clone() const { - return new DecimalFormat(*this); -} - -UBool DecimalFormat::operator==(const Format& other) const { - auto* otherDF = dynamic_cast(&other); - if (otherDF == nullptr) { - return false; - } - return *fields->properties == *otherDF->fields->properties && *fields->symbols == *otherDF->fields->symbols; -} - -UnicodeString& DecimalFormat::format(double number, UnicodeString& appendTo, FieldPosition& pos) const { - if (pos.getField() == FieldPosition::DONT_CARE && fastFormatDouble(number, appendTo)) { - return appendTo; - } - UErrorCode localStatus = U_ZERO_ERROR; - FormattedNumber output = fields->formatter->formatDouble(number, localStatus); - fieldPositionHelper(output, pos, appendTo.length(), localStatus); - auto appendable = UnicodeStringAppendable(appendTo); - output.appendTo(appendable); - return appendTo; -} - -UnicodeString& DecimalFormat::format(double number, UnicodeString& appendTo, FieldPosition& pos, - UErrorCode& status) const { - if (pos.getField() == FieldPosition::DONT_CARE && fastFormatDouble(number, appendTo)) { - return appendTo; - } - FormattedNumber output = fields->formatter->formatDouble(number, status); - fieldPositionHelper(output, pos, appendTo.length(), status); - auto appendable = UnicodeStringAppendable(appendTo); - output.appendTo(appendable); - return appendTo; -} - -UnicodeString& -DecimalFormat::format(double number, UnicodeString& appendTo, FieldPositionIterator* posIter, - UErrorCode& status) const { - if (posIter == nullptr && fastFormatDouble(number, appendTo)) { - return appendTo; - } - FormattedNumber output = fields->formatter->formatDouble(number, status); - fieldPositionIteratorHelper(output, posIter, appendTo.length(), status); - auto appendable = UnicodeStringAppendable(appendTo); - output.appendTo(appendable); - return appendTo; -} - -UnicodeString& DecimalFormat::format(int32_t number, UnicodeString& appendTo, FieldPosition& pos) const { - return format(static_cast (number), appendTo, pos); -} - -UnicodeString& DecimalFormat::format(int32_t number, UnicodeString& appendTo, FieldPosition& pos, - UErrorCode& status) const { - return format(static_cast (number), appendTo, pos, status); -} - -UnicodeString& -DecimalFormat::format(int32_t number, UnicodeString& appendTo, FieldPositionIterator* posIter, - UErrorCode& status) const { - return format(static_cast (number), appendTo, posIter, status); -} - -UnicodeString& DecimalFormat::format(int64_t number, UnicodeString& appendTo, FieldPosition& pos) const { - if (pos.getField() == FieldPosition::DONT_CARE && fastFormatInt64(number, appendTo)) { - return appendTo; - } - UErrorCode localStatus = U_ZERO_ERROR; - FormattedNumber output = fields->formatter->formatInt(number, localStatus); - fieldPositionHelper(output, pos, appendTo.length(), localStatus); - auto appendable = UnicodeStringAppendable(appendTo); - output.appendTo(appendable); - return appendTo; -} - -UnicodeString& DecimalFormat::format(int64_t number, UnicodeString& appendTo, FieldPosition& pos, - UErrorCode& status) const { - if (pos.getField() == FieldPosition::DONT_CARE && fastFormatInt64(number, appendTo)) { - return appendTo; - } - FormattedNumber output = fields->formatter->formatInt(number, status); - fieldPositionHelper(output, pos, appendTo.length(), status); - auto appendable = UnicodeStringAppendable(appendTo); - output.appendTo(appendable); - return appendTo; -} - -UnicodeString& -DecimalFormat::format(int64_t number, UnicodeString& appendTo, FieldPositionIterator* posIter, - UErrorCode& status) const { - if (posIter == nullptr && fastFormatInt64(number, appendTo)) { - return appendTo; - } - FormattedNumber output = fields->formatter->formatInt(number, status); - fieldPositionIteratorHelper(output, posIter, appendTo.length(), status); - auto appendable = UnicodeStringAppendable(appendTo); - output.appendTo(appendable); - return appendTo; -} - -UnicodeString& -DecimalFormat::format(StringPiece number, UnicodeString& appendTo, FieldPositionIterator* posIter, - UErrorCode& status) const { - FormattedNumber output = fields->formatter->formatDecimal(number, status); - fieldPositionIteratorHelper(output, posIter, appendTo.length(), status); - auto appendable = UnicodeStringAppendable(appendTo); - output.appendTo(appendable); - return appendTo; -} - -UnicodeString& DecimalFormat::format(const DecimalQuantity& number, UnicodeString& appendTo, - FieldPositionIterator* posIter, UErrorCode& status) const { - FormattedNumber output = fields->formatter->formatDecimalQuantity(number, status); - fieldPositionIteratorHelper(output, posIter, appendTo.length(), status); - auto appendable = UnicodeStringAppendable(appendTo); - output.appendTo(appendable); - return appendTo; -} - -UnicodeString& -DecimalFormat::format(const DecimalQuantity& number, UnicodeString& appendTo, FieldPosition& pos, - UErrorCode& status) const { - FormattedNumber output = fields->formatter->formatDecimalQuantity(number, status); - fieldPositionHelper(output, pos, appendTo.length(), status); - auto appendable = UnicodeStringAppendable(appendTo); - output.appendTo(appendable); - return appendTo; -} - -void DecimalFormat::parse(const UnicodeString& text, Formattable& output, - ParsePosition& parsePosition) const { - if (parsePosition.getIndex() < 0 || parsePosition.getIndex() >= text.length()) { - return; - } - - ErrorCode status; - ParsedNumber result; - // Note: if this is a currency instance, currencies will be matched despite the fact that we are not in the - // parseCurrency method (backwards compatibility) - int32_t startIndex = parsePosition.getIndex(); - const NumberParserImpl* parser = getParser(status); - if (U_FAILURE(status)) { return; } - parser->parse(text, startIndex, true, result, status); - // TODO: Do we need to check for fImpl->properties->parseAllInput (UCONFIG_HAVE_PARSEALLINPUT) here? - if (result.success()) { - parsePosition.setIndex(result.charEnd); - result.populateFormattable(output, parser->getParseFlags()); - } else { - parsePosition.setErrorIndex(startIndex + result.charEnd); - } -} - -CurrencyAmount* DecimalFormat::parseCurrency(const UnicodeString& text, ParsePosition& parsePosition) const { - if (parsePosition.getIndex() < 0 || parsePosition.getIndex() >= text.length()) { - return nullptr; - } - - ErrorCode status; - ParsedNumber result; - // Note: if this is a currency instance, currencies will be matched despite the fact that we are not in the - // parseCurrency method (backwards compatibility) - int32_t startIndex = parsePosition.getIndex(); - const NumberParserImpl* parser = getCurrencyParser(status); - if (U_FAILURE(status)) { return nullptr; } - parser->parse(text, startIndex, true, result, status); - // TODO: Do we need to check for fImpl->properties->parseAllInput (UCONFIG_HAVE_PARSEALLINPUT) here? - if (result.success()) { - parsePosition.setIndex(result.charEnd); - Formattable formattable; - result.populateFormattable(formattable, parser->getParseFlags()); - return new CurrencyAmount(formattable, result.currencyCode, status); - } else { - parsePosition.setErrorIndex(startIndex + result.charEnd); - return nullptr; - } -} - -const DecimalFormatSymbols* DecimalFormat::getDecimalFormatSymbols(void) const { - return fields->symbols.getAlias(); -} - -void DecimalFormat::adoptDecimalFormatSymbols(DecimalFormatSymbols* symbolsToAdopt) { - if (symbolsToAdopt == nullptr) { - return; // do not allow caller to set fields->symbols to NULL - } - fields->symbols.adoptInstead(symbolsToAdopt); - touchNoError(); -} - -void DecimalFormat::setDecimalFormatSymbols(const DecimalFormatSymbols& symbols) { - fields->symbols.adoptInstead(new DecimalFormatSymbols(symbols)); - touchNoError(); -} - -const CurrencyPluralInfo* DecimalFormat::getCurrencyPluralInfo(void) const { - return fields->properties->currencyPluralInfo.fPtr.getAlias(); -} - -void DecimalFormat::adoptCurrencyPluralInfo(CurrencyPluralInfo* toAdopt) { - fields->properties->currencyPluralInfo.fPtr.adoptInstead(toAdopt); - touchNoError(); -} - -void DecimalFormat::setCurrencyPluralInfo(const CurrencyPluralInfo& info) { - if (fields->properties->currencyPluralInfo.fPtr.isNull()) { - fields->properties->currencyPluralInfo.fPtr.adoptInstead(info.clone()); - } else { - *fields->properties->currencyPluralInfo.fPtr = info; // copy-assignment operator - } - touchNoError(); -} - -UnicodeString& DecimalFormat::getPositivePrefix(UnicodeString& result) const { - ErrorCode localStatus; - fields->formatter->getAffixImpl(true, false, result, localStatus); - return result; -} - -void DecimalFormat::setPositivePrefix(const UnicodeString& newValue) { - if (newValue == fields->properties->positivePrefix) { return; } - fields->properties->positivePrefix = newValue; - touchNoError(); -} - -UnicodeString& DecimalFormat::getNegativePrefix(UnicodeString& result) const { - ErrorCode localStatus; - fields->formatter->getAffixImpl(true, true, result, localStatus); - return result; -} - -void DecimalFormat::setNegativePrefix(const UnicodeString& newValue) { - if (newValue == fields->properties->negativePrefix) { return; } - fields->properties->negativePrefix = newValue; - touchNoError(); -} - -UnicodeString& DecimalFormat::getPositiveSuffix(UnicodeString& result) const { - ErrorCode localStatus; - fields->formatter->getAffixImpl(false, false, result, localStatus); - return result; -} - -void DecimalFormat::setPositiveSuffix(const UnicodeString& newValue) { - if (newValue == fields->properties->positiveSuffix) { return; } - fields->properties->positiveSuffix = newValue; - touchNoError(); -} - -UnicodeString& DecimalFormat::getNegativeSuffix(UnicodeString& result) const { - ErrorCode localStatus; - fields->formatter->getAffixImpl(false, true, result, localStatus); - return result; -} - -void DecimalFormat::setNegativeSuffix(const UnicodeString& newValue) { - if (newValue == fields->properties->negativeSuffix) { return; } - fields->properties->negativeSuffix = newValue; - touchNoError(); -} - -UBool DecimalFormat::isSignAlwaysShown() const { - return fields->properties->signAlwaysShown; -} - -void DecimalFormat::setSignAlwaysShown(UBool value) { - if (UBOOL_TO_BOOL(value) == fields->properties->signAlwaysShown) { return; } - fields->properties->signAlwaysShown = value; - touchNoError(); -} - -int32_t DecimalFormat::getMultiplier(void) const { - if (fields->properties->multiplier != 1) { - return fields->properties->multiplier; - } else if (fields->properties->magnitudeMultiplier != 0) { - return static_cast(uprv_pow10(fields->properties->magnitudeMultiplier)); - } else { - return 1; - } -} - -void DecimalFormat::setMultiplier(int32_t multiplier) { - if (multiplier == 0) { - multiplier = 1; // one being the benign default value for a multiplier. - } - - // Try to convert to a magnitude multiplier first - int delta = 0; - int value = multiplier; - while (value != 1) { - delta++; - int temp = value / 10; - if (temp * 10 != value) { - delta = -1; - break; - } - value = temp; - } - if (delta != -1) { - fields->properties->magnitudeMultiplier = delta; - fields->properties->multiplier = 1; - } else { - fields->properties->magnitudeMultiplier = 0; - fields->properties->multiplier = multiplier; - } - touchNoError(); -} - -int32_t DecimalFormat::getMultiplierScale() const { - return fields->properties->multiplierScale; -} - -void DecimalFormat::setMultiplierScale(int32_t newValue) { - if (newValue == fields->properties->multiplierScale) { return; } - fields->properties->multiplierScale = newValue; - touchNoError(); -} - -double DecimalFormat::getRoundingIncrement(void) const { - return fields->exportedProperties->roundingIncrement; -} - -void DecimalFormat::setRoundingIncrement(double newValue) { - if (newValue == fields->properties->roundingIncrement) { return; } - fields->properties->roundingIncrement = newValue; - touchNoError(); -} - -ERoundingMode DecimalFormat::getRoundingMode(void) const { - // UNumberFormatRoundingMode and ERoundingMode have the same values. - return static_cast(fields->exportedProperties->roundingMode.getNoError()); -} - -void DecimalFormat::setRoundingMode(ERoundingMode roundingMode) { - auto uRoundingMode = static_cast(roundingMode); - if (!fields->properties->roundingMode.isNull() && uRoundingMode == fields->properties->roundingMode.getNoError()) { - return; - } - NumberFormat::setMaximumIntegerDigits(roundingMode); // to set field for compatibility - fields->properties->roundingMode = uRoundingMode; - touchNoError(); -} - -int32_t DecimalFormat::getFormatWidth(void) const { - return fields->properties->formatWidth; -} - -void DecimalFormat::setFormatWidth(int32_t width) { - if (width == fields->properties->formatWidth) { return; } - fields->properties->formatWidth = width; - touchNoError(); -} - -UnicodeString DecimalFormat::getPadCharacterString() const { - if (fields->properties->padString.isBogus()) { - // Readonly-alias the static string kFallbackPaddingString - return {TRUE, kFallbackPaddingString, -1}; - } else { - return fields->properties->padString; - } -} - -void DecimalFormat::setPadCharacter(const UnicodeString& padChar) { - if (padChar == fields->properties->padString) { return; } - if (padChar.length() > 0) { - fields->properties->padString = UnicodeString(padChar.char32At(0)); - } else { - fields->properties->padString.setToBogus(); - } - touchNoError(); -} - -EPadPosition DecimalFormat::getPadPosition(void) const { - if (fields->properties->padPosition.isNull()) { - return EPadPosition::kPadBeforePrefix; - } else { - // UNumberFormatPadPosition and EPadPosition have the same values. - return static_cast(fields->properties->padPosition.getNoError()); - } -} - -void DecimalFormat::setPadPosition(EPadPosition padPos) { - auto uPadPos = static_cast(padPos); - if (!fields->properties->padPosition.isNull() && uPadPos == fields->properties->padPosition.getNoError()) { - return; - } - fields->properties->padPosition = uPadPos; - touchNoError(); -} - -UBool DecimalFormat::isScientificNotation(void) const { - return fields->properties->minimumExponentDigits != -1; -} - -void DecimalFormat::setScientificNotation(UBool useScientific) { - int32_t minExp = useScientific ? 1 : -1; - if (fields->properties->minimumExponentDigits == minExp) { return; } - if (useScientific) { - fields->properties->minimumExponentDigits = 1; - } else { - fields->properties->minimumExponentDigits = -1; - } - touchNoError(); -} - -int8_t DecimalFormat::getMinimumExponentDigits(void) const { - return static_cast(fields->properties->minimumExponentDigits); -} - -void DecimalFormat::setMinimumExponentDigits(int8_t minExpDig) { - if (minExpDig == fields->properties->minimumExponentDigits) { return; } - fields->properties->minimumExponentDigits = minExpDig; - touchNoError(); -} - -UBool DecimalFormat::isExponentSignAlwaysShown(void) const { - return fields->properties->exponentSignAlwaysShown; -} - -void DecimalFormat::setExponentSignAlwaysShown(UBool expSignAlways) { - if (UBOOL_TO_BOOL(expSignAlways) == fields->properties->exponentSignAlwaysShown) { return; } - fields->properties->exponentSignAlwaysShown = expSignAlways; - touchNoError(); -} - -int32_t DecimalFormat::getGroupingSize(void) const { - if (fields->properties->groupingSize < 0) { - return 0; - } - return fields->properties->groupingSize; -} - -void DecimalFormat::setGroupingSize(int32_t newValue) { - if (newValue == fields->properties->groupingSize) { return; } - fields->properties->groupingSize = newValue; - touchNoError(); -} - -int32_t DecimalFormat::getSecondaryGroupingSize(void) const { - int grouping2 = fields->properties->secondaryGroupingSize; - if (grouping2 < 0) { - return 0; - } - return grouping2; -} - -void DecimalFormat::setSecondaryGroupingSize(int32_t newValue) { - if (newValue == fields->properties->secondaryGroupingSize) { return; } - fields->properties->secondaryGroupingSize = newValue; - touchNoError(); -} - -int32_t DecimalFormat::getMinimumGroupingDigits() const { - return fields->properties->minimumGroupingDigits; -} - -void DecimalFormat::setMinimumGroupingDigits(int32_t newValue) { - if (newValue == fields->properties->minimumGroupingDigits) { return; } - fields->properties->minimumGroupingDigits = newValue; - touchNoError(); -} - -UBool DecimalFormat::isDecimalSeparatorAlwaysShown(void) const { - return fields->properties->decimalSeparatorAlwaysShown; -} - -void DecimalFormat::setDecimalSeparatorAlwaysShown(UBool newValue) { - if (UBOOL_TO_BOOL(newValue) == fields->properties->decimalSeparatorAlwaysShown) { return; } - fields->properties->decimalSeparatorAlwaysShown = newValue; - touchNoError(); -} - -UBool DecimalFormat::isDecimalPatternMatchRequired(void) const { - return fields->properties->decimalPatternMatchRequired; -} - -void DecimalFormat::setDecimalPatternMatchRequired(UBool newValue) { - if (UBOOL_TO_BOOL(newValue) == fields->properties->decimalPatternMatchRequired) { return; } - fields->properties->decimalPatternMatchRequired = newValue; - touchNoError(); -} - -UBool DecimalFormat::isParseNoExponent() const { - return fields->properties->parseNoExponent; -} - -void DecimalFormat::setParseNoExponent(UBool value) { - if (UBOOL_TO_BOOL(value) == fields->properties->parseNoExponent) { return; } - fields->properties->parseNoExponent = value; - touchNoError(); -} - -UBool DecimalFormat::isParseCaseSensitive() const { - return fields->properties->parseCaseSensitive; -} - -void DecimalFormat::setParseCaseSensitive(UBool value) { - if (UBOOL_TO_BOOL(value) == fields->properties->parseCaseSensitive) { return; } - fields->properties->parseCaseSensitive = value; - touchNoError(); -} - -UBool DecimalFormat::isFormatFailIfMoreThanMaxDigits() const { - return fields->properties->formatFailIfMoreThanMaxDigits; -} - -void DecimalFormat::setFormatFailIfMoreThanMaxDigits(UBool value) { - if (UBOOL_TO_BOOL(value) == fields->properties->formatFailIfMoreThanMaxDigits) { return; } - fields->properties->formatFailIfMoreThanMaxDigits = value; - touchNoError(); -} - -UnicodeString& DecimalFormat::toPattern(UnicodeString& result) const { - // Pull some properties from exportedProperties and others from properties - // to keep affix patterns intact. In particular, pull rounding properties - // so that CurrencyUsage is reflected properly. - // TODO: Consider putting this logic in number_patternstring.cpp instead. - ErrorCode localStatus; - DecimalFormatProperties tprops(*fields->properties); - bool useCurrency = ((!tprops.currency.isNull()) || !tprops.currencyPluralInfo.fPtr.isNull() || - !tprops.currencyUsage.isNull() || AffixUtils::hasCurrencySymbols( - tprops.positivePrefixPattern, localStatus) || AffixUtils::hasCurrencySymbols( - tprops.positiveSuffixPattern, localStatus) || AffixUtils::hasCurrencySymbols( - tprops.negativePrefixPattern, localStatus) || AffixUtils::hasCurrencySymbols( - tprops.negativeSuffixPattern, localStatus)); - if (useCurrency) { - tprops.minimumFractionDigits = fields->exportedProperties->minimumFractionDigits; - tprops.maximumFractionDigits = fields->exportedProperties->maximumFractionDigits; - tprops.roundingIncrement = fields->exportedProperties->roundingIncrement; - } - result = PatternStringUtils::propertiesToPatternString(tprops, localStatus); - return result; -} - -UnicodeString& DecimalFormat::toLocalizedPattern(UnicodeString& result) const { - ErrorCode localStatus; - result = toPattern(result); - result = PatternStringUtils::convertLocalized(result, *fields->symbols, true, localStatus); - return result; -} - -void DecimalFormat::applyPattern(const UnicodeString& pattern, UParseError&, UErrorCode& status) { - // TODO: What is parseError for? - applyPattern(pattern, status); -} - -void DecimalFormat::applyPattern(const UnicodeString& pattern, UErrorCode& status) { - setPropertiesFromPattern(pattern, IGNORE_ROUNDING_NEVER, status); - touch(status); -} - -void DecimalFormat::applyLocalizedPattern(const UnicodeString& localizedPattern, UParseError&, - UErrorCode& status) { - // TODO: What is parseError for? - applyLocalizedPattern(localizedPattern, status); -} - -void DecimalFormat::applyLocalizedPattern(const UnicodeString& localizedPattern, UErrorCode& status) { - if (U_SUCCESS(status)) { - UnicodeString pattern = PatternStringUtils::convertLocalized( - localizedPattern, *fields->symbols, false, status); - applyPattern(pattern, status); - } -} - -void DecimalFormat::setMaximumIntegerDigits(int32_t newValue) { - if (newValue == fields->properties->maximumIntegerDigits) { return; } - // For backwards compatibility, conflicting min/max need to keep the most recent setting. - int32_t min = fields->properties->minimumIntegerDigits; - if (min >= 0 && min > newValue) { - fields->properties->minimumIntegerDigits = newValue; - } - fields->properties->maximumIntegerDigits = newValue; - touchNoError(); -} - -void DecimalFormat::setMinimumIntegerDigits(int32_t newValue) { - if (newValue == fields->properties->minimumIntegerDigits) { return; } - // For backwards compatibility, conflicting min/max need to keep the most recent setting. - int32_t max = fields->properties->maximumIntegerDigits; - if (max >= 0 && max < newValue) { - fields->properties->maximumIntegerDigits = newValue; - } - fields->properties->minimumIntegerDigits = newValue; - touchNoError(); -} - -void DecimalFormat::setMaximumFractionDigits(int32_t newValue) { - if (newValue == fields->properties->maximumFractionDigits) { return; } - // For backwards compatibility, conflicting min/max need to keep the most recent setting. - int32_t min = fields->properties->minimumFractionDigits; - if (min >= 0 && min > newValue) { - fields->properties->minimumFractionDigits = newValue; - } - fields->properties->maximumFractionDigits = newValue; - touchNoError(); -} - -void DecimalFormat::setMinimumFractionDigits(int32_t newValue) { - if (newValue == fields->properties->minimumFractionDigits) { return; } - // For backwards compatibility, conflicting min/max need to keep the most recent setting. - int32_t max = fields->properties->maximumFractionDigits; - if (max >= 0 && max < newValue) { - fields->properties->maximumFractionDigits = newValue; - } - fields->properties->minimumFractionDigits = newValue; - touchNoError(); -} - -int32_t DecimalFormat::getMinimumSignificantDigits() const { - return fields->exportedProperties->minimumSignificantDigits; -} - -int32_t DecimalFormat::getMaximumSignificantDigits() const { - return fields->exportedProperties->maximumSignificantDigits; -} - -void DecimalFormat::setMinimumSignificantDigits(int32_t value) { - if (value == fields->properties->minimumSignificantDigits) { return; } - int32_t max = fields->properties->maximumSignificantDigits; - if (max >= 0 && max < value) { - fields->properties->maximumSignificantDigits = value; - } - fields->properties->minimumSignificantDigits = value; - touchNoError(); -} - -void DecimalFormat::setMaximumSignificantDigits(int32_t value) { - if (value == fields->properties->maximumSignificantDigits) { return; } - int32_t min = fields->properties->minimumSignificantDigits; - if (min >= 0 && min > value) { - fields->properties->minimumSignificantDigits = value; - } - fields->properties->maximumSignificantDigits = value; - touchNoError(); -} - -UBool DecimalFormat::areSignificantDigitsUsed() const { - return fields->properties->minimumSignificantDigits != -1 || fields->properties->maximumSignificantDigits != -1; -} - -void DecimalFormat::setSignificantDigitsUsed(UBool useSignificantDigits) { - if (areSignificantDigitsUsed()) return; - // These are the default values from the old implementation. - int32_t minSig = useSignificantDigits ? 1 : -1; - int32_t maxSig = useSignificantDigits ? 6 : -1; - if (fields->properties->minimumSignificantDigits == minSig && - fields->properties->maximumSignificantDigits == maxSig) { - return; - } - fields->properties->minimumSignificantDigits = minSig; - fields->properties->maximumSignificantDigits = maxSig; - touchNoError(); -} - -void DecimalFormat::setCurrency(const char16_t* theCurrency, UErrorCode& ec) { - CurrencyUnit currencyUnit(theCurrency, ec); - if (U_FAILURE(ec)) { return; } - if (!fields->properties->currency.isNull() && fields->properties->currency.getNoError() == currencyUnit) { - return; - } - NumberFormat::setCurrency(theCurrency, ec); // to set field for compatibility - fields->properties->currency = currencyUnit; - // TODO: Set values in fields->symbols, too? - touchNoError(); -} - -void DecimalFormat::setCurrency(const char16_t* theCurrency) { - ErrorCode localStatus; - setCurrency(theCurrency, localStatus); -} - -void DecimalFormat::setCurrencyUsage(UCurrencyUsage newUsage, UErrorCode* ec) { - if (U_FAILURE(*ec)) { - return; - } - if (!fields->properties->currencyUsage.isNull() && newUsage == fields->properties->currencyUsage.getNoError()) { - return; - } - fields->properties->currencyUsage = newUsage; - touch(*ec); -} - -UCurrencyUsage DecimalFormat::getCurrencyUsage() const { - // CurrencyUsage is not exported, so we have to get it from the input property bag. - // TODO: Should we export CurrencyUsage instead? - if (fields->properties->currencyUsage.isNull()) { - return UCURR_USAGE_STANDARD; - } - return fields->properties->currencyUsage.getNoError(); -} - -void -DecimalFormat::formatToDecimalQuantity(double number, DecimalQuantity& output, UErrorCode& status) const { - fields->formatter->formatDouble(number, status).getDecimalQuantity(output, status); -} - -void DecimalFormat::formatToDecimalQuantity(const Formattable& number, DecimalQuantity& output, - UErrorCode& status) const { - UFormattedNumberData obj; - number.populateDecimalQuantity(obj.quantity, status); - fields->formatter->formatImpl(&obj, status); - output = std::move(obj.quantity); -} - -const number::LocalizedNumberFormatter& DecimalFormat::toNumberFormatter() const { - return *fields->formatter; -} - -/** Rebuilds the formatter object from the property bag. */ -void DecimalFormat::touch(UErrorCode& status) { - if (fields->exportedProperties == nullptr) { - // fields->exportedProperties is null only when the formatter is not ready yet. - // The only time when this happens is during legacy deserialization. - return; - } - - // In C++, fields->symbols is the source of truth for the locale. - Locale locale = fields->symbols->getLocale(); - - // Note: The formatter is relatively cheap to create, and we need it to populate fields->exportedProperties, - // so automatically compute it here. The parser is a bit more expensive and is not needed until the - // parse method is called, so defer that until needed. - // TODO: Only update the pieces that changed instead of re-computing the whole formatter? - fields->formatter.adoptInstead( - new LocalizedNumberFormatter( - NumberPropertyMapper::create( - *fields->properties, *fields->symbols, fields->warehouse, *fields->exportedProperties, status).locale( - locale))); - - // Do this after fields->exportedProperties are set up - setupFastFormat(); - - // Delete the parsers if they were made previously - delete fields->atomicParser.exchange(nullptr); - delete fields->atomicCurrencyParser.exchange(nullptr); - - // In order for the getters to work, we need to populate some fields in NumberFormat. - NumberFormat::setCurrency(fields->exportedProperties->currency.get(status).getISOCurrency(), status); - NumberFormat::setMaximumIntegerDigits(fields->exportedProperties->maximumIntegerDigits); - NumberFormat::setMinimumIntegerDigits(fields->exportedProperties->minimumIntegerDigits); - NumberFormat::setMaximumFractionDigits(fields->exportedProperties->maximumFractionDigits); - NumberFormat::setMinimumFractionDigits(fields->exportedProperties->minimumFractionDigits); - // fImpl->properties, not fields->exportedProperties, since this information comes from the pattern: - NumberFormat::setGroupingUsed(fields->properties->groupingUsed); -} - -void DecimalFormat::touchNoError() { - UErrorCode localStatus = U_ZERO_ERROR; - touch(localStatus); -} - -void DecimalFormat::setPropertiesFromPattern(const UnicodeString& pattern, int32_t ignoreRounding, - UErrorCode& status) { - if (U_SUCCESS(status)) { - // Cast workaround to get around putting the enum in the public header file - auto actualIgnoreRounding = static_cast(ignoreRounding); - PatternParser::parseToExistingProperties(pattern, *fields->properties, actualIgnoreRounding, status); - } -} - -const numparse::impl::NumberParserImpl* DecimalFormat::getParser(UErrorCode& status) const { - if (U_FAILURE(status)) { return nullptr; } - - // First try to get the pre-computed parser - auto* ptr = fields->atomicParser.load(); - if (ptr != nullptr) { - return ptr; - } - - // Try computing the parser on our own - auto* temp = NumberParserImpl::createParserFromProperties(*fields->properties, *fields->symbols, false, status); - if (temp == nullptr) { - status = U_MEMORY_ALLOCATION_ERROR; - // although we may still dereference, call sites should be guarded - } - - // Note: ptr starts as nullptr; during compare_exchange, it is set to what is actually stored in the - // atomic if another thread beat us to computing the parser object. - auto* nonConstThis = const_cast(this); - if (!nonConstThis->fields->atomicParser.compare_exchange_strong(ptr, temp)) { - // Another thread beat us to computing the parser - delete temp; - return ptr; - } else { - // Our copy of the parser got stored in the atomic - return temp; - } -} - -const numparse::impl::NumberParserImpl* DecimalFormat::getCurrencyParser(UErrorCode& status) const { - if (U_FAILURE(status)) { return nullptr; } - - // First try to get the pre-computed parser - auto* ptr = fields->atomicCurrencyParser.load(); - if (ptr != nullptr) { - return ptr; - } - - // Try computing the parser on our own - auto* temp = NumberParserImpl::createParserFromProperties(*fields->properties, *fields->symbols, true, status); - if (temp == nullptr) { - status = U_MEMORY_ALLOCATION_ERROR; - // although we may still dereference, call sites should be guarded - } - - // Note: ptr starts as nullptr; during compare_exchange, it is set to what is actually stored in the - // atomic if another thread beat us to computing the parser object. - auto* nonConstThis = const_cast(this); - if (!nonConstThis->fields->atomicCurrencyParser.compare_exchange_strong(ptr, temp)) { - // Another thread beat us to computing the parser - delete temp; - return ptr; - } else { - // Our copy of the parser got stored in the atomic - return temp; - } -} - -void -DecimalFormat::fieldPositionHelper(const number::FormattedNumber& formatted, FieldPosition& fieldPosition, - int32_t offset, UErrorCode& status) { - // always return first occurrence: - fieldPosition.setBeginIndex(0); - fieldPosition.setEndIndex(0); - bool found = formatted.nextFieldPosition(fieldPosition, status); - if (found && offset != 0) { - FieldPositionOnlyHandler fpoh(fieldPosition); - fpoh.shiftLast(offset); - } -} - -void -DecimalFormat::fieldPositionIteratorHelper(const number::FormattedNumber& formatted, FieldPositionIterator* fpi, - int32_t offset, UErrorCode& status) { - if (fpi != nullptr) { - FieldPositionIteratorHandler fpih(fpi, status); - fpih.setShift(offset); - formatted.getAllFieldPositionsImpl(fpih, status); - } -} - -// To debug fast-format, change void(x) to printf(x) -#define trace(x) void(x) - -void DecimalFormat::setupFastFormat() { - // Check the majority of properties: - if (!fields->properties->equalsDefaultExceptFastFormat()) { - trace("no fast format: equality\n"); - fields->canUseFastFormat = false; - return; - } - - // Now check the remaining properties. - // Nontrivial affixes: - UBool trivialPP = fields->properties->positivePrefixPattern.isEmpty(); - UBool trivialPS = fields->properties->positiveSuffixPattern.isEmpty(); - UBool trivialNP = fields->properties->negativePrefixPattern.isBogus() || ( - fields->properties->negativePrefixPattern.length() == 1 && - fields->properties->negativePrefixPattern.charAt(0) == u'-'); - UBool trivialNS = fields->properties->negativeSuffixPattern.isEmpty(); - if (!trivialPP || !trivialPS || !trivialNP || !trivialNS) { - trace("no fast format: affixes\n"); - fields->canUseFastFormat = false; - return; - } - - // Grouping (secondary grouping is forbidden in equalsDefaultExceptFastFormat): - bool groupingUsed = fields->properties->groupingUsed; - int32_t groupingSize = fields->properties->groupingSize; - bool unusualGroupingSize = groupingSize > 0 && groupingSize != 3; - const UnicodeString& groupingString = fields->symbols->getConstSymbol(DecimalFormatSymbols::kGroupingSeparatorSymbol); - if (groupingUsed && (unusualGroupingSize || groupingString.length() != 1)) { - trace("no fast format: grouping\n"); - fields->canUseFastFormat = false; - return; - } - - // Integer length: - int32_t minInt = fields->exportedProperties->minimumIntegerDigits; - int32_t maxInt = fields->exportedProperties->maximumIntegerDigits; - // Fastpath supports up to only 10 digits (length of INT32_MIN) - if (minInt > 10) { - trace("no fast format: integer\n"); - fields->canUseFastFormat = false; - return; - } - - // Fraction length (no fraction part allowed in fast path): - int32_t minFrac = fields->exportedProperties->minimumFractionDigits; - if (minFrac > 0) { - trace("no fast format: fraction\n"); - fields->canUseFastFormat = false; - return; - } - - // Other symbols: - const UnicodeString& minusSignString = fields->symbols->getConstSymbol(DecimalFormatSymbols::kMinusSignSymbol); - UChar32 codePointZero = fields->symbols->getCodePointZero(); - if (minusSignString.length() != 1 || U16_LENGTH(codePointZero) != 1) { - trace("no fast format: symbols\n"); - fields->canUseFastFormat = false; - return; - } - - // Good to go! - trace("can use fast format!\n"); - fields->canUseFastFormat = true; - fields->fastData.cpZero = static_cast(codePointZero); - fields->fastData.cpGroupingSeparator = groupingUsed && groupingSize == 3 ? groupingString.charAt(0) : 0; - fields->fastData.cpMinusSign = minusSignString.charAt(0); - fields->fastData.minInt = (minInt < 0 || minInt > 127) ? 0 : static_cast(minInt); - fields->fastData.maxInt = (maxInt < 0 || maxInt > 127) ? 127 : static_cast(maxInt); -} - -bool DecimalFormat::fastFormatDouble(double input, UnicodeString& output) const { - if (!fields->canUseFastFormat) { - return false; - } - if (std::isnan(input) - || std::trunc(input) != input - || input <= INT32_MIN - || input > INT32_MAX) { - return false; - } - doFastFormatInt32(static_cast(input), std::signbit(input), output); - return true; -} - -bool DecimalFormat::fastFormatInt64(int64_t input, UnicodeString& output) const { - if (!fields->canUseFastFormat) { - return false; - } - if (input <= INT32_MIN || input > INT32_MAX) { - return false; - } - doFastFormatInt32(static_cast(input), input < 0, output); - return true; -} - -void DecimalFormat::doFastFormatInt32(int32_t input, bool isNegative, UnicodeString& output) const { - U_ASSERT(fields->canUseFastFormat); - if (isNegative) { - output.append(fields->fastData.cpMinusSign); - U_ASSERT(input != INT32_MIN); // handled by callers - input = -input; - } - // Cap at int32_t to make the buffer small and operations fast. - // Longest string: "2,147,483,648" (13 chars in length) - static constexpr int32_t localCapacity = 13; - char16_t localBuffer[localCapacity]; - char16_t* ptr = localBuffer + localCapacity; - int8_t group = 0; - for (int8_t i = 0; i < fields->fastData.maxInt && (input != 0 || i < fields->fastData.minInt); i++) { - if (group++ == 3 && fields->fastData.cpGroupingSeparator != 0) { - *(--ptr) = fields->fastData.cpGroupingSeparator; - group = 1; - } - std::div_t res = std::div(input, 10); - *(--ptr) = static_cast(fields->fastData.cpZero + res.rem); - input = res.quot; - } - int32_t len = localCapacity - static_cast(ptr - localBuffer); - output.append(ptr, len); -} - - -#endif /* #if !UCONFIG_NO_FORMATTING */ diff --git a/tools/icu/patches/63/source/i18n/dtptngen.cpp b/tools/icu/patches/63/source/i18n/dtptngen.cpp deleted file mode 100644 index 4ed724b6a4..0000000000 --- a/tools/icu/patches/63/source/i18n/dtptngen.cpp +++ /dev/null @@ -1,2752 +0,0 @@ -// © 2016 and later: Unicode, Inc. and others. -// License & terms of use: http://www.unicode.org/copyright.html -/* -******************************************************************************* -* Copyright (C) 2007-2016, International Business Machines Corporation and -* others. All Rights Reserved. -******************************************************************************* -* -* File DTPTNGEN.CPP -* -******************************************************************************* -*/ - -#include "unicode/utypes.h" -#if !UCONFIG_NO_FORMATTING - -#include "unicode/datefmt.h" -#include "unicode/decimfmt.h" -#include "unicode/dtfmtsym.h" -#include "unicode/dtptngen.h" -#include "unicode/localpointer.h" -#include "unicode/simpleformatter.h" -#include "unicode/smpdtfmt.h" -#include "unicode/udat.h" -#include "unicode/udatpg.h" -#include "unicode/uniset.h" -#include "unicode/uloc.h" -#include "unicode/ures.h" -#include "unicode/ustring.h" -#include "unicode/rep.h" -#include "cpputils.h" -#include "mutex.h" -#include "umutex.h" -#include "cmemory.h" -#include "cstring.h" -#include "locbased.h" -#include "hash.h" -#include "uhash.h" -#include "uresimp.h" -#include "dtptngen_impl.h" -#include "ucln_in.h" -#include "charstr.h" -#include "uassert.h" - -#if U_CHARSET_FAMILY==U_EBCDIC_FAMILY -/** - * If we are on EBCDIC, use an iterator which will - * traverse the bundles in ASCII order. - */ -#define U_USE_ASCII_BUNDLE_ITERATOR -#define U_SORT_ASCII_BUNDLE_ITERATOR -#endif - -#if defined(U_USE_ASCII_BUNDLE_ITERATOR) - -#include "unicode/ustring.h" -#include "uarrsort.h" - -struct UResAEntry { - UChar *key; - UResourceBundle *item; -}; - -struct UResourceBundleAIterator { - UResourceBundle *bund; - UResAEntry *entries; - int32_t num; - int32_t cursor; -}; - -/* Must be C linkage to pass function pointer to the sort function */ - -U_CDECL_BEGIN - -static int32_t U_CALLCONV -ures_a_codepointSort(const void *context, const void *left, const void *right) { - //CompareContext *cmp=(CompareContext *)context; - return u_strcmp(((const UResAEntry *)left)->key, - ((const UResAEntry *)right)->key); -} - -U_CDECL_END - -static void ures_a_open(UResourceBundleAIterator *aiter, UResourceBundle *bund, UErrorCode *status) { - if(U_FAILURE(*status)) { - return; - } - aiter->bund = bund; - aiter->num = ures_getSize(aiter->bund); - aiter->cursor = 0; -#if !defined(U_SORT_ASCII_BUNDLE_ITERATOR) - aiter->entries = nullptr; -#else - aiter->entries = (UResAEntry*)uprv_malloc(sizeof(UResAEntry)*aiter->num); - for(int i=0;inum;i++) { - aiter->entries[i].item = ures_getByIndex(aiter->bund, i, nullptr, status); - const char *akey = ures_getKey(aiter->entries[i].item); - int32_t len = uprv_strlen(akey)+1; - aiter->entries[i].key = (UChar*)uprv_malloc(len*sizeof(UChar)); - u_charsToUChars(akey, aiter->entries[i].key, len); - } - uprv_sortArray(aiter->entries, aiter->num, sizeof(UResAEntry), ures_a_codepointSort, nullptr, TRUE, status); -#endif -} - -static void ures_a_close(UResourceBundleAIterator *aiter) { -#if defined(U_SORT_ASCII_BUNDLE_ITERATOR) - for(int i=0;inum;i++) { - uprv_free(aiter->entries[i].key); - ures_close(aiter->entries[i].item); - } -#endif -} - -static const UChar *ures_a_getNextString(UResourceBundleAIterator *aiter, int32_t *len, const char **key, UErrorCode *err) { -#if !defined(U_SORT_ASCII_BUNDLE_ITERATOR) - return ures_getNextString(aiter->bund, len, key, err); -#else - if(U_FAILURE(*err)) return nullptr; - UResourceBundle *item = aiter->entries[aiter->cursor].item; - const UChar* ret = ures_getString(item, len, err); - *key = ures_getKey(item); - aiter->cursor++; - return ret; -#endif -} - - -#endif - - -U_NAMESPACE_BEGIN - -// ***************************************************************************** -// class DateTimePatternGenerator -// ***************************************************************************** -static const UChar Canonical_Items[] = { - // GyQMwWEDFdaHmsSv - CAP_G, LOW_Y, CAP_Q, CAP_M, LOW_W, CAP_W, CAP_E, - CAP_D, CAP_F, LOW_D, LOW_A, // The UDATPG_x_FIELD constants and these fields have a different order than in ICU4J - CAP_H, LOW_M, LOW_S, CAP_S, LOW_V, 0 -}; - -static const dtTypeElem dtTypes[] = { - // patternChar, field, type, minLen, weight - {CAP_G, UDATPG_ERA_FIELD, DT_SHORT, 1, 3,}, - {CAP_G, UDATPG_ERA_FIELD, DT_LONG, 4, 0}, - {CAP_G, UDATPG_ERA_FIELD, DT_NARROW, 5, 0}, - - {LOW_Y, UDATPG_YEAR_FIELD, DT_NUMERIC, 1, 20}, - {CAP_Y, UDATPG_YEAR_FIELD, DT_NUMERIC + DT_DELTA, 1, 20}, - {LOW_U, UDATPG_YEAR_FIELD, DT_NUMERIC + 2*DT_DELTA, 1, 20}, - {LOW_R, UDATPG_YEAR_FIELD, DT_NUMERIC + 3*DT_DELTA, 1, 20}, - {CAP_U, UDATPG_YEAR_FIELD, DT_SHORT, 1, 3}, - {CAP_U, UDATPG_YEAR_FIELD, DT_LONG, 4, 0}, - {CAP_U, UDATPG_YEAR_FIELD, DT_NARROW, 5, 0}, - - {CAP_Q, UDATPG_QUARTER_FIELD, DT_NUMERIC, 1, 2}, - {CAP_Q, UDATPG_QUARTER_FIELD, DT_SHORT, 3, 0}, - {CAP_Q, UDATPG_QUARTER_FIELD, DT_LONG, 4, 0}, - {CAP_Q, UDATPG_QUARTER_FIELD, DT_NARROW, 5, 0}, - {LOW_Q, UDATPG_QUARTER_FIELD, DT_NUMERIC + DT_DELTA, 1, 2}, - {LOW_Q, UDATPG_QUARTER_FIELD, DT_SHORT - DT_DELTA, 3, 0}, - {LOW_Q, UDATPG_QUARTER_FIELD, DT_LONG - DT_DELTA, 4, 0}, - {LOW_Q, UDATPG_QUARTER_FIELD, DT_NARROW - DT_DELTA, 5, 0}, - - {CAP_M, UDATPG_MONTH_FIELD, DT_NUMERIC, 1, 2}, - {CAP_M, UDATPG_MONTH_FIELD, DT_SHORT, 3, 0}, - {CAP_M, UDATPG_MONTH_FIELD, DT_LONG, 4, 0}, - {CAP_M, UDATPG_MONTH_FIELD, DT_NARROW, 5, 0}, - {CAP_L, UDATPG_MONTH_FIELD, DT_NUMERIC + DT_DELTA, 1, 2}, - {CAP_L, UDATPG_MONTH_FIELD, DT_SHORT - DT_DELTA, 3, 0}, - {CAP_L, UDATPG_MONTH_FIELD, DT_LONG - DT_DELTA, 4, 0}, - {CAP_L, UDATPG_MONTH_FIELD, DT_NARROW - DT_DELTA, 5, 0}, - {LOW_L, UDATPG_MONTH_FIELD, DT_NUMERIC + DT_DELTA, 1, 1}, - - {LOW_W, UDATPG_WEEK_OF_YEAR_FIELD, DT_NUMERIC, 1, 2}, - - {CAP_W, UDATPG_WEEK_OF_MONTH_FIELD, DT_NUMERIC, 1, 0}, - - {CAP_E, UDATPG_WEEKDAY_FIELD, DT_SHORT, 1, 3}, - {CAP_E, UDATPG_WEEKDAY_FIELD, DT_LONG, 4, 0}, - {CAP_E, UDATPG_WEEKDAY_FIELD, DT_NARROW, 5, 0}, - {CAP_E, UDATPG_WEEKDAY_FIELD, DT_SHORTER, 6, 0}, - {LOW_C, UDATPG_WEEKDAY_FIELD, DT_NUMERIC + 2*DT_DELTA, 1, 2}, - {LOW_C, UDATPG_WEEKDAY_FIELD, DT_SHORT - 2*DT_DELTA, 3, 0}, - {LOW_C, UDATPG_WEEKDAY_FIELD, DT_LONG - 2*DT_DELTA, 4, 0}, - {LOW_C, UDATPG_WEEKDAY_FIELD, DT_NARROW - 2*DT_DELTA, 5, 0}, - {LOW_C, UDATPG_WEEKDAY_FIELD, DT_SHORTER - 2*DT_DELTA, 6, 0}, - {LOW_E, UDATPG_WEEKDAY_FIELD, DT_NUMERIC + DT_DELTA, 1, 2}, // LOW_E is currently not used in CLDR data, should not be canonical - {LOW_E, UDATPG_WEEKDAY_FIELD, DT_SHORT - DT_DELTA, 3, 0}, - {LOW_E, UDATPG_WEEKDAY_FIELD, DT_LONG - DT_DELTA, 4, 0}, - {LOW_E, UDATPG_WEEKDAY_FIELD, DT_NARROW - DT_DELTA, 5, 0}, - {LOW_E, UDATPG_WEEKDAY_FIELD, DT_SHORTER - DT_DELTA, 6, 0}, - - {LOW_D, UDATPG_DAY_FIELD, DT_NUMERIC, 1, 2}, - {LOW_G, UDATPG_DAY_FIELD, DT_NUMERIC + DT_DELTA, 1, 20}, // really internal use, so we don't care - - {CAP_D, UDATPG_DAY_OF_YEAR_FIELD, DT_NUMERIC, 1, 3}, - - {CAP_F, UDATPG_DAY_OF_WEEK_IN_MONTH_FIELD, DT_NUMERIC, 1, 0}, - - {LOW_A, UDATPG_DAYPERIOD_FIELD, DT_SHORT, 1, 3}, - {LOW_A, UDATPG_DAYPERIOD_FIELD, DT_LONG, 4, 0}, - {LOW_A, UDATPG_DAYPERIOD_FIELD, DT_NARROW, 5, 0}, - {LOW_B, UDATPG_DAYPERIOD_FIELD, DT_SHORT - DT_DELTA, 1, 3}, - {LOW_B, UDATPG_DAYPERIOD_FIELD, DT_LONG - DT_DELTA, 4, 0}, - {LOW_B, UDATPG_DAYPERIOD_FIELD, DT_NARROW - DT_DELTA, 5, 0}, - // b needs to be closer to a than to B, so we make this 3*DT_DELTA - {CAP_B, UDATPG_DAYPERIOD_FIELD, DT_SHORT - 3*DT_DELTA, 1, 3}, - {CAP_B, UDATPG_DAYPERIOD_FIELD, DT_LONG - 3*DT_DELTA, 4, 0}, - {CAP_B, UDATPG_DAYPERIOD_FIELD, DT_NARROW - 3*DT_DELTA, 5, 0}, - - {CAP_H, UDATPG_HOUR_FIELD, DT_NUMERIC + 10*DT_DELTA, 1, 2}, // 24 hour - {LOW_K, UDATPG_HOUR_FIELD, DT_NUMERIC + 11*DT_DELTA, 1, 2}, // 24 hour - {LOW_H, UDATPG_HOUR_FIELD, DT_NUMERIC, 1, 2}, // 12 hour - {CAP_K, UDATPG_HOUR_FIELD, DT_NUMERIC + DT_DELTA, 1, 2}, // 12 hour - // The C code has had versions of the following 3, keep & update. Should not need these, but... - // Without these, certain tests using e.g. staticGetSkeleton fail because j/J in patterns - // get skipped instead of mapped to the right hour chars, for example in - // DateFormatTest::TestPatternFromSkeleton - // IntlTestDateTimePatternGeneratorAPI:: testStaticGetSkeleton - // DateIntervalFormatTest::testTicket11985 - // Need to investigate better handling of jJC replacement e.g. in staticGetSkeleton. - {CAP_J, UDATPG_HOUR_FIELD, DT_NUMERIC + 5*DT_DELTA, 1, 2}, // 12/24 hour no AM/PM - {LOW_J, UDATPG_HOUR_FIELD, DT_NUMERIC + 6*DT_DELTA, 1, 6}, // 12/24 hour - {CAP_C, UDATPG_HOUR_FIELD, DT_NUMERIC + 7*DT_DELTA, 1, 6}, // 12/24 hour with preferred dayPeriods for 12 - - {LOW_M, UDATPG_MINUTE_FIELD, DT_NUMERIC, 1, 2}, - - {LOW_S, UDATPG_SECOND_FIELD, DT_NUMERIC, 1, 2}, - {CAP_A, UDATPG_SECOND_FIELD, DT_NUMERIC + DT_DELTA, 1, 1000}, - - {CAP_S, UDATPG_FRACTIONAL_SECOND_FIELD, DT_NUMERIC, 1, 1000}, - - {LOW_V, UDATPG_ZONE_FIELD, DT_SHORT - 2*DT_DELTA, 1, 0}, - {LOW_V, UDATPG_ZONE_FIELD, DT_LONG - 2*DT_DELTA, 4, 0}, - {LOW_Z, UDATPG_ZONE_FIELD, DT_SHORT, 1, 3}, - {LOW_Z, UDATPG_ZONE_FIELD, DT_LONG, 4, 0}, - {CAP_Z, UDATPG_ZONE_FIELD, DT_NARROW - DT_DELTA, 1, 3}, - {CAP_Z, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 4, 0}, - {CAP_Z, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 5, 0}, - {CAP_O, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 1, 0}, - {CAP_O, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 4, 0}, - {CAP_V, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 1, 0}, - {CAP_V, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 2, 0}, - {CAP_V, UDATPG_ZONE_FIELD, DT_LONG-1 - DT_DELTA, 3, 0}, - {CAP_V, UDATPG_ZONE_FIELD, DT_LONG-2 - DT_DELTA, 4, 0}, - {CAP_X, UDATPG_ZONE_FIELD, DT_NARROW - DT_DELTA, 1, 0}, - {CAP_X, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 2, 0}, - {CAP_X, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 4, 0}, - {LOW_X, UDATPG_ZONE_FIELD, DT_NARROW - DT_DELTA, 1, 0}, - {LOW_X, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 2, 0}, - {LOW_X, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 4, 0}, - - {0, UDATPG_FIELD_COUNT, 0, 0, 0} , // last row of dtTypes[] - }; - -static const char* const CLDR_FIELD_APPEND[] = { - "Era", "Year", "Quarter", "Month", "Week", "*", "Day-Of-Week", - "*", "*", "Day", "*", // The UDATPG_x_FIELD constants and these fields have a different order than in ICU4J - "Hour", "Minute", "Second", "*", "Timezone" -}; - -static const char* const CLDR_FIELD_NAME[UDATPG_FIELD_COUNT] = { - "era", "year", "quarter", "month", "week", "weekOfMonth", "weekday", - "dayOfYear", "weekdayOfMonth", "day", "dayperiod", // The UDATPG_x_FIELD constants and these fields have a different order than in ICU4J - "hour", "minute", "second", "*", "zone" -}; - -static const char* const CLDR_FIELD_WIDTH[] = { // [UDATPG_WIDTH_COUNT] - "", "-short", "-narrow" -}; - -// TODO(ticket:13619): remove when definition uncommented in dtptngen.h. -static const int32_t UDATPG_WIDTH_COUNT = UDATPG_NARROW + 1; -static constexpr UDateTimePGDisplayWidth UDATPG_WIDTH_APPENDITEM = UDATPG_WIDE; -static constexpr int32_t UDATPG_FIELD_KEY_MAX = 24; // max length of CLDR field tag (type + width) - -// For appendItems -static const UChar UDATPG_ItemFormat[]= {0x7B, 0x30, 0x7D, 0x20, 0x251C, 0x7B, 0x32, 0x7D, 0x3A, - 0x20, 0x7B, 0x31, 0x7D, 0x2524, 0}; // {0} \u251C{2}: {1}\u2524 - -//static const UChar repeatedPatterns[6]={CAP_G, CAP_E, LOW_Z, LOW_V, CAP_Q, 0}; // "GEzvQ" - -static const char DT_DateTimePatternsTag[]="DateTimePatterns"; -static const char DT_DateTimeCalendarTag[]="calendar"; -static const char DT_DateTimeGregorianTag[]="gregorian"; -static const char DT_DateTimeAppendItemsTag[]="appendItems"; -static const char DT_DateTimeFieldsTag[]="fields"; -static const char DT_DateTimeAvailableFormatsTag[]="availableFormats"; -//static const UnicodeString repeatedPattern=UnicodeString(repeatedPatterns); - -UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DateTimePatternGenerator) -UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DTSkeletonEnumeration) -UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DTRedundantEnumeration) - -DateTimePatternGenerator* U_EXPORT2 -DateTimePatternGenerator::createInstance(UErrorCode& status) { - return createInstance(Locale::getDefault(), status); -} - -DateTimePatternGenerator* U_EXPORT2 -DateTimePatternGenerator::createInstance(const Locale& locale, UErrorCode& status) { - if (U_FAILURE(status)) { - return nullptr; - } - LocalPointer result( - new DateTimePatternGenerator(locale, status), status); - return U_SUCCESS(status) ? result.orphan() : nullptr; -} - -DateTimePatternGenerator* U_EXPORT2 -DateTimePatternGenerator::createEmptyInstance(UErrorCode& status) { - if (U_FAILURE(status)) { - return nullptr; - } - LocalPointer result( - new DateTimePatternGenerator(status), status); - return U_SUCCESS(status) ? result.orphan() : nullptr; -} - -DateTimePatternGenerator::DateTimePatternGenerator(UErrorCode &status) : - skipMatcher(nullptr), - fAvailableFormatKeyHash(nullptr), - internalErrorCode(U_ZERO_ERROR) -{ - fp = new FormatParser(); - dtMatcher = new DateTimeMatcher(); - distanceInfo = new DistanceInfo(); - patternMap = new PatternMap(); - if (fp == nullptr || dtMatcher == nullptr || distanceInfo == nullptr || patternMap == nullptr) { - internalErrorCode = status = U_MEMORY_ALLOCATION_ERROR; - } -} - -DateTimePatternGenerator::DateTimePatternGenerator(const Locale& locale, UErrorCode &status) : - skipMatcher(nullptr), - fAvailableFormatKeyHash(nullptr), - internalErrorCode(U_ZERO_ERROR) -{ - fp = new FormatParser(); - dtMatcher = new DateTimeMatcher(); - distanceInfo = new DistanceInfo(); - patternMap = new PatternMap(); - if (fp == nullptr || dtMatcher == nullptr || distanceInfo == nullptr || patternMap == nullptr) { - internalErrorCode = status = U_MEMORY_ALLOCATION_ERROR; - } - else { - initData(locale, status); - } -} - -DateTimePatternGenerator::DateTimePatternGenerator(const DateTimePatternGenerator& other) : - UObject(), - skipMatcher(nullptr), - fAvailableFormatKeyHash(nullptr), - internalErrorCode(U_ZERO_ERROR) -{ - fp = new FormatParser(); - dtMatcher = new DateTimeMatcher(); - distanceInfo = new DistanceInfo(); - patternMap = new PatternMap(); - if (fp == nullptr || dtMatcher == nullptr || distanceInfo == nullptr || patternMap == nullptr) { - internalErrorCode = U_MEMORY_ALLOCATION_ERROR; - } - *this=other; -} - -DateTimePatternGenerator& -DateTimePatternGenerator::operator=(const DateTimePatternGenerator& other) { - // reflexive case - if (&other == this) { - return *this; - } - internalErrorCode = other.internalErrorCode; - pLocale = other.pLocale; - fDefaultHourFormatChar = other.fDefaultHourFormatChar; - *fp = *(other.fp); - dtMatcher->copyFrom(other.dtMatcher->skeleton); - *distanceInfo = *(other.distanceInfo); - dateTimeFormat = other.dateTimeFormat; - decimal = other.decimal; - // NUL-terminate for the C API. - dateTimeFormat.getTerminatedBuffer(); - decimal.getTerminatedBuffer(); - delete skipMatcher; - if ( other.skipMatcher == nullptr ) { - skipMatcher = nullptr; - } - else { - skipMatcher = new DateTimeMatcher(*other.skipMatcher); - if (skipMatcher == nullptr) - { - internalErrorCode = U_MEMORY_ALLOCATION_ERROR; - return *this; - } - } - for (int32_t i=0; i< UDATPG_FIELD_COUNT; ++i ) { - appendItemFormats[i] = other.appendItemFormats[i]; - appendItemFormats[i].getTerminatedBuffer(); // NUL-terminate for the C API. - for (int32_t j=0; j< UDATPG_WIDTH_COUNT; ++j ) { - fieldDisplayNames[i][j] = other.fieldDisplayNames[i][j]; - fieldDisplayNames[i][j].getTerminatedBuffer(); // NUL-terminate for the C API. - } - } - patternMap->copyFrom(*other.patternMap, internalErrorCode); - copyHashtable(other.fAvailableFormatKeyHash, internalErrorCode); - return *this; -} - - -UBool -DateTimePatternGenerator::operator==(const DateTimePatternGenerator& other) const { - if (this == &other) { - return TRUE; - } - if ((pLocale==other.pLocale) && (patternMap->equals(*other.patternMap)) && - (dateTimeFormat==other.dateTimeFormat) && (decimal==other.decimal)) { - for ( int32_t i=0 ; i list; - int32_t length; - if (value.getType() == URES_STRING) { - if (list.allocateInsteadAndReset(2) == nullptr) { - errorCode = U_MEMORY_ALLOCATION_ERROR; - return; - } - list[0] = getHourFormatFromUnicodeString(value.getUnicodeString(errorCode)); - length = 1; - } - else { - ResourceArray allowedFormats = value.getArray(errorCode); - length = allowedFormats.getSize(); - if (list.allocateInsteadAndReset(length + 1) == nullptr) { - errorCode = U_MEMORY_ALLOCATION_ERROR; - return; - } - for (int32_t k = 0; k < length; ++k) { - allowedFormats.getValue(k, value); - list[k] = getHourFormatFromUnicodeString(value.getUnicodeString(errorCode)); - } - } - list[length] = ALLOWED_HOUR_FORMAT_UNKNOWN; - uhash_put(localeToAllowedHourFormatsMap, - const_cast(regionOrLocale), list.orphan(), &errorCode); - if (U_FAILURE(errorCode)) { return; } - } - } - } - } - - AllowedHourFormat getHourFormatFromUnicodeString(const UnicodeString &s) { - if (s.length() == 1) { - if (s[0] == LOW_H) { return ALLOWED_HOUR_FORMAT_h; } - if (s[0] == CAP_H) { return ALLOWED_HOUR_FORMAT_H; } - } else if (s.length() == 2) { - if (s[0] == LOW_H && s[1] == LOW_B) { return ALLOWED_HOUR_FORMAT_hb; } - if (s[0] == CAP_H && s[1] == LOW_B) { return ALLOWED_HOUR_FORMAT_Hb; } - if (s[0] == LOW_H && s[1] == CAP_B) { return ALLOWED_HOUR_FORMAT_hB; } - if (s[0] == CAP_H && s[1] == CAP_B) { return ALLOWED_HOUR_FORMAT_HB; } - } - - return ALLOWED_HOUR_FORMAT_UNKNOWN; - } -}; - -} // namespace - -AllowedHourFormatsSink::~AllowedHourFormatsSink() {} - -U_CFUNC void U_CALLCONV DateTimePatternGenerator::loadAllowedHourFormatsData(UErrorCode &status) { - if (U_FAILURE(status)) { return; } - localeToAllowedHourFormatsMap = uhash_open( - uhash_hashChars, uhash_compareChars, nullptr, &status); - if (U_FAILURE(status)) { return; } - - uhash_setValueDeleter(localeToAllowedHourFormatsMap, deleteAllowedHourFormats); - ucln_i18n_registerCleanup(UCLN_I18N_ALLOWED_HOUR_FORMATS, allowedHourFormatsCleanup); - - LocalUResourceBundlePointer rb(ures_openDirect(nullptr, "supplementalData", &status)); - if (U_FAILURE(status)) { return; } - - AllowedHourFormatsSink sink; - // TODO: Currently in the enumeration each table allocates a new array. - // Try to reduce the number of memory allocations. Consider storing a - // UVector32 with the concatenation of all of the sub-arrays, put the start index - // into the hashmap, store 6 single-value sub-arrays right at the beginning of the - // vector (at index enum*2) for easy data sharing, copy sub-arrays into runtime - // object. Remember to clean up the vector, too. - ures_getAllItemsWithFallback(rb.getAlias(), "timeData", sink, status); -} - -void DateTimePatternGenerator::getAllowedHourFormats(const Locale &locale, UErrorCode &status) { - if (U_FAILURE(status)) { return; } - const char *localeID = locale.getName(); - char maxLocaleID[ULOC_FULLNAME_CAPACITY]; - int32_t length = uloc_addLikelySubtags(localeID, maxLocaleID, ULOC_FULLNAME_CAPACITY, &status); - if (U_FAILURE(status)) { - return; - } else if (length == ULOC_FULLNAME_CAPACITY) { // no room for NUL - status = U_BUFFER_OVERFLOW_ERROR; - return; - } - Locale maxLocale = Locale(maxLocaleID); - - const char *country = maxLocale.getCountry(); - if (*country == '\0') { country = "001"; } - const char *language = maxLocale.getLanguage(); - - CharString langCountry; - langCountry.append(language, static_cast(uprv_strlen(language)), status); - langCountry.append('_', status); - langCountry.append(country, static_cast(uprv_strlen(country)), status); - - int32_t *allowedFormats; - allowedFormats = (int32_t *)uhash_get(localeToAllowedHourFormatsMap, langCountry.data()); - if (allowedFormats == nullptr) { - allowedFormats = (int32_t *)uhash_get(localeToAllowedHourFormatsMap, const_cast(country)); - } - - if (allowedFormats != nullptr) { // Lookup is successful - for (int32_t i = 0; i < UPRV_LENGTHOF(fAllowedHourFormats); ++i) { - fAllowedHourFormats[i] = allowedFormats[i]; - if (allowedFormats[i] == ALLOWED_HOUR_FORMAT_UNKNOWN) { - break; - } - } - } else { // Lookup failed, twice - fAllowedHourFormats[0] = ALLOWED_HOUR_FORMAT_H; - fAllowedHourFormats[1] = ALLOWED_HOUR_FORMAT_UNKNOWN; - } -} - -UnicodeString -DateTimePatternGenerator::getSkeleton(const UnicodeString& pattern, UErrorCode& -/*status*/) { - FormatParser fp2; - DateTimeMatcher matcher; - PtnSkeleton localSkeleton; - matcher.set(pattern, &fp2, localSkeleton); - return localSkeleton.getSkeleton(); -} - -UnicodeString -DateTimePatternGenerator::staticGetSkeleton( - const UnicodeString& pattern, UErrorCode& /*status*/) { - FormatParser fp; - DateTimeMatcher matcher; - PtnSkeleton localSkeleton; - matcher.set(pattern, &fp, localSkeleton); - return localSkeleton.getSkeleton(); -} - -UnicodeString -DateTimePatternGenerator::getBaseSkeleton(const UnicodeString& pattern, UErrorCode& /*status*/) { - FormatParser fp2; - DateTimeMatcher matcher; - PtnSkeleton localSkeleton; - matcher.set(pattern, &fp2, localSkeleton); - return localSkeleton.getBaseSkeleton(); -} - -UnicodeString -DateTimePatternGenerator::staticGetBaseSkeleton( - const UnicodeString& pattern, UErrorCode& /*status*/) { - FormatParser fp; - DateTimeMatcher matcher; - PtnSkeleton localSkeleton; - matcher.set(pattern, &fp, localSkeleton); - return localSkeleton.getBaseSkeleton(); -} - -void -DateTimePatternGenerator::addICUPatterns(const Locale& locale, UErrorCode& status) { - if (U_FAILURE(status)) { return; } - UnicodeString dfPattern; - UnicodeString conflictingString; - DateFormat* df; - - // Load with ICU patterns - for (int32_t i=DateFormat::kFull; i<=DateFormat::kShort; i++) { - DateFormat::EStyle style = (DateFormat::EStyle)i; - df = DateFormat::createDateInstance(style, locale); - SimpleDateFormat* sdf; - if (df != nullptr && (sdf = dynamic_cast(df)) != nullptr) { - sdf->toPattern(dfPattern); - addPattern(dfPattern, FALSE, conflictingString, status); - } - // TODO Maybe we should return an error when the date format isn't simple. - delete df; - if (U_FAILURE(status)) { return; } - - df = DateFormat::createTimeInstance(style, locale); - if (df != nullptr && (sdf = dynamic_cast(df)) != nullptr) { - sdf->toPattern(dfPattern); - addPattern(dfPattern, FALSE, conflictingString, status); - - // TODO: C++ and Java are inconsistent (see #12568). - // C++ uses MEDIUM, but Java uses SHORT. - if ( i==DateFormat::kShort && !dfPattern.isEmpty() ) { - consumeShortTimePattern(dfPattern, status); - } - } - // TODO Maybe we should return an error when the date format isn't simple. - delete df; - if (U_FAILURE(status)) { return; } - } -} - -void -DateTimePatternGenerator::hackTimes(const UnicodeString& hackPattern, UErrorCode& status) { - UnicodeString conflictingString; - - fp->set(hackPattern); - UnicodeString mmss; - UBool gotMm=FALSE; - for (int32_t i=0; iitemNumber; ++i) { - UnicodeString field = fp->items[i]; - if ( fp->isQuoteLiteral(field) ) { - if ( gotMm ) { - UnicodeString quoteLiteral; - fp->getQuoteLiteral(quoteLiteral, &i); - mmss += quoteLiteral; - } - } - else { - if (fp->isPatternSeparator(field) && gotMm) { - mmss+=field; - } - else { - UChar ch=field.charAt(0); - if (ch==LOW_M) { - gotMm=TRUE; - mmss+=field; - } - else { - if (ch==LOW_S) { - if (!gotMm) { - break; - } - mmss+= field; - addPattern(mmss, FALSE, conflictingString, status); - break; - } - else { - if (gotMm || ch==LOW_Z || ch==CAP_Z || ch==LOW_V || ch==CAP_V) { - break; - } - } - } - } - } - } -} - -#define ULOC_LOCALE_IDENTIFIER_CAPACITY (ULOC_FULLNAME_CAPACITY + 1 + ULOC_KEYWORD_AND_VALUES_CAPACITY) - -static const UChar hourFormatChars[] = { CAP_H, LOW_H, CAP_K, LOW_K, 0 }; // HhKk, the hour format characters - -void -DateTimePatternGenerator::getCalendarTypeToUse(const Locale& locale, CharString& destination, UErrorCode& err) { - destination.clear().append(DT_DateTimeGregorianTag, -1, err); // initial default - if ( U_SUCCESS(err) ) { - UErrorCode localStatus = U_ZERO_ERROR; - char localeWithCalendarKey[ULOC_LOCALE_IDENTIFIER_CAPACITY]; - // obtain a locale that always has the calendar key value that should be used - ures_getFunctionalEquivalent( - localeWithCalendarKey, - ULOC_LOCALE_IDENTIFIER_CAPACITY, - nullptr, - "calendar", - "calendar", - locale.getName(), - nullptr, - FALSE, - &localStatus); - localeWithCalendarKey[ULOC_LOCALE_IDENTIFIER_CAPACITY-1] = 0; // ensure null termination - // now get the calendar key value from that locale - char calendarType[ULOC_KEYWORDS_CAPACITY]; - int32_t calendarTypeLen = uloc_getKeywordValue( - localeWithCalendarKey, - "calendar", - calendarType, - ULOC_KEYWORDS_CAPACITY, - &localStatus); - // If the input locale was invalid, don't fail with missing resource error, instead - // continue with default of Gregorian. - if (U_FAILURE(localStatus) && localStatus != U_MISSING_RESOURCE_ERROR) { - err = localStatus; - return; - } - if (calendarTypeLen < ULOC_KEYWORDS_CAPACITY) { - destination.clear().append(calendarType, -1, err); - if (U_FAILURE(err)) { return; } - } - } -} - -void -DateTimePatternGenerator::consumeShortTimePattern(const UnicodeString& shortTimePattern, - UErrorCode& status) { - if (U_FAILURE(status)) { return; } - // set fDefaultHourFormatChar to the hour format character from this pattern - int32_t tfIdx, tfLen = shortTimePattern.length(); - UBool ignoreChars = FALSE; - for (tfIdx = 0; tfIdx < tfLen; tfIdx++) { - UChar tfChar = shortTimePattern.charAt(tfIdx); - if ( tfChar == SINGLE_QUOTE ) { - ignoreChars = !ignoreChars; // toggle (handle quoted literals & '' for single quote) - } else if ( !ignoreChars && u_strchr(hourFormatChars, tfChar) != nullptr ) { - fDefaultHourFormatChar = tfChar; - break; - } - } - - // HACK for hh:ss - hackTimes(shortTimePattern, status); -} - -struct DateTimePatternGenerator::AppendItemFormatsSink : public ResourceSink { - - // Destination for data, modified via setters. - DateTimePatternGenerator& dtpg; - - AppendItemFormatsSink(DateTimePatternGenerator& _dtpg) : dtpg(_dtpg) {} - virtual ~AppendItemFormatsSink(); - - virtual void put(const char *key, ResourceValue &value, UBool /*noFallback*/, - UErrorCode &errorCode) { - ResourceTable itemsTable = value.getTable(errorCode); - if (U_FAILURE(errorCode)) { return; } - for (int32_t i = 0; itemsTable.getKeyAndValue(i, key, value); ++i) { - UDateTimePatternField field = dtpg.getAppendFormatNumber(key); - if (field == UDATPG_FIELD_COUNT) { continue; } - const UnicodeString& valueStr = value.getUnicodeString(errorCode); - if (dtpg.getAppendItemFormat(field).isEmpty() && !valueStr.isEmpty()) { - dtpg.setAppendItemFormat(field, valueStr); - } - } - } - - void fillInMissing() { - UnicodeString defaultItemFormat(TRUE, UDATPG_ItemFormat, UPRV_LENGTHOF(UDATPG_ItemFormat)-1); // Read-only alias. - for (int32_t i = 0; i < UDATPG_FIELD_COUNT; i++) { - UDateTimePatternField field = (UDateTimePatternField)i; - if (dtpg.getAppendItemFormat(field).isEmpty()) { - dtpg.setAppendItemFormat(field, defaultItemFormat); - } - } - } -}; - -struct DateTimePatternGenerator::AppendItemNamesSink : public ResourceSink { - - // Destination for data, modified via setters. - DateTimePatternGenerator& dtpg; - - AppendItemNamesSink(DateTimePatternGenerator& _dtpg) : dtpg(_dtpg) {} - virtual ~AppendItemNamesSink(); - - virtual void put(const char *key, ResourceValue &value, UBool /*noFallback*/, - UErrorCode &errorCode) { - ResourceTable itemsTable = value.getTable(errorCode); - if (U_FAILURE(errorCode)) { return; } - for (int32_t i = 0; itemsTable.getKeyAndValue(i, key, value); ++i) { - UDateTimePGDisplayWidth width; - UDateTimePatternField field = dtpg.getFieldAndWidthIndices(key, &width); - if (field == UDATPG_FIELD_COUNT) { continue; } - ResourceTable detailsTable = value.getTable(errorCode); - if (U_FAILURE(errorCode)) { return; } - for (int32_t j = 0; detailsTable.getKeyAndValue(j, key, value); ++j) { - if (uprv_strcmp(key, "dn") != 0) { continue; } - const UnicodeString& valueStr = value.getUnicodeString(errorCode); - if (dtpg.getFieldDisplayName(field,width).isEmpty() && !valueStr.isEmpty()) { - dtpg.setFieldDisplayName(field,width,valueStr); - } - break; - } - } - } - - void fillInMissing() { - for (int32_t i = 0; i < UDATPG_FIELD_COUNT; i++) { - UnicodeString& valueStr = dtpg.getMutableFieldDisplayName((UDateTimePatternField)i, UDATPG_WIDE); - if (valueStr.isEmpty()) { - valueStr = CAP_F; - U_ASSERT(i < 20); - if (i < 10) { - // F0, F1, ..., F9 - valueStr += (UChar)(i+0x30); - } else { - // F10, F11, ... - valueStr += (UChar)0x31; - valueStr += (UChar)(i-10 + 0x30); - } - // NUL-terminate for the C API. - valueStr.getTerminatedBuffer(); - } - for (int32_t j = 1; j < UDATPG_WIDTH_COUNT; j++) { - UnicodeString& valueStr2 = dtpg.getMutableFieldDisplayName((UDateTimePatternField)i, (UDateTimePGDisplayWidth)j); - if (valueStr2.isEmpty()) { - valueStr2 = dtpg.getFieldDisplayName((UDateTimePatternField)i, (UDateTimePGDisplayWidth)(j-1)); - } - } - } - } -}; - -struct DateTimePatternGenerator::AvailableFormatsSink : public ResourceSink { - - // Destination for data, modified via setters. - DateTimePatternGenerator& dtpg; - - // Temporary variable, required for calling addPatternWithSkeleton. - UnicodeString conflictingPattern; - - AvailableFormatsSink(DateTimePatternGenerator& _dtpg) : dtpg(_dtpg) {} - virtual ~AvailableFormatsSink(); - - virtual void put(const char *key, ResourceValue &value, UBool isRoot, - UErrorCode &errorCode) { - ResourceTable itemsTable = value.getTable(errorCode); - if (U_FAILURE(errorCode)) { return; } - for (int32_t i = 0; itemsTable.getKeyAndValue(i, key, value); ++i) { - const UnicodeString formatKey(key, -1, US_INV); - if (!dtpg.isAvailableFormatSet(formatKey) ) { - dtpg.setAvailableFormat(formatKey, errorCode); - // Add pattern with its associated skeleton. Override any duplicate - // derived from std patterns, but not a previous availableFormats entry: - const UnicodeString& formatValue = value.getUnicodeString(errorCode); - conflictingPattern.remove(); - dtpg.addPatternWithSkeleton(formatValue, &formatKey, !isRoot, conflictingPattern, errorCode); - } - } - } -}; - -// Virtual destructors must be defined out of line. -DateTimePatternGenerator::AppendItemFormatsSink::~AppendItemFormatsSink() {} -DateTimePatternGenerator::AppendItemNamesSink::~AppendItemNamesSink() {} -DateTimePatternGenerator::AvailableFormatsSink::~AvailableFormatsSink() {} - -void -DateTimePatternGenerator::addCLDRData(const Locale& locale, UErrorCode& errorCode) { - if (U_FAILURE(errorCode)) { return; } - UnicodeString rbPattern, value, field; - CharString path; - - LocalUResourceBundlePointer rb(ures_open(nullptr, locale.getName(), &errorCode)); - if (U_FAILURE(errorCode)) { return; } - - CharString calendarTypeToUse; // to be filled in with the type to use, if all goes well - getCalendarTypeToUse(locale, calendarTypeToUse, errorCode); - if (U_FAILURE(errorCode)) { return; } - - // Local err to ignore resource not found exceptions - UErrorCode err = U_ZERO_ERROR; - - // Load append item formats. - AppendItemFormatsSink appendItemFormatsSink(*this); - path.clear() - .append(DT_DateTimeCalendarTag, errorCode) - .append('/', errorCode) - .append(calendarTypeToUse, errorCode) - .append('/', errorCode) - .append(DT_DateTimeAppendItemsTag, errorCode); // i.e., calendar/xxx/appendItems - if (U_FAILURE(errorCode)) { return; } - ures_getAllItemsWithFallback(rb.getAlias(), path.data(), appendItemFormatsSink, err); - appendItemFormatsSink.fillInMissing(); - - // Load CLDR item names. - err = U_ZERO_ERROR; - AppendItemNamesSink appendItemNamesSink(*this); - ures_getAllItemsWithFallback(rb.getAlias(), DT_DateTimeFieldsTag, appendItemNamesSink, err); - appendItemNamesSink.fillInMissing(); - - // Load the available formats from CLDR. - err = U_ZERO_ERROR; - initHashtable(errorCode); - if (U_FAILURE(errorCode)) { return; } - AvailableFormatsSink availableFormatsSink(*this); - path.clear() - .append(DT_DateTimeCalendarTag, errorCode) - .append('/', errorCode) - .append(calendarTypeToUse, errorCode) - .append('/', errorCode) - .append(DT_DateTimeAvailableFormatsTag, errorCode); // i.e., calendar/xxx/availableFormats - if (U_FAILURE(errorCode)) { return; } - ures_getAllItemsWithFallback(rb.getAlias(), path.data(), availableFormatsSink, err); -} - -void -DateTimePatternGenerator::initHashtable(UErrorCode& err) { - if (U_FAILURE(err)) { return; } - if (fAvailableFormatKeyHash!=nullptr) { - return; - } - LocalPointer hash(new Hashtable(FALSE, err), err); - if (U_SUCCESS(err)) { - fAvailableFormatKeyHash = hash.orphan(); - } -} - -void -DateTimePatternGenerator::setAppendItemFormat(UDateTimePatternField field, const UnicodeString& value) { - appendItemFormats[field] = value; - // NUL-terminate for the C API. - appendItemFormats[field].getTerminatedBuffer(); -} - -const UnicodeString& -DateTimePatternGenerator::getAppendItemFormat(UDateTimePatternField field) const { - return appendItemFormats[field]; -} - -void -DateTimePatternGenerator::setAppendItemName(UDateTimePatternField field, const UnicodeString& value) { - setFieldDisplayName(field, UDATPG_WIDTH_APPENDITEM, value); -} - -const UnicodeString& -DateTimePatternGenerator::getAppendItemName(UDateTimePatternField field) const { - return fieldDisplayNames[field][UDATPG_WIDTH_APPENDITEM]; -} - -void -DateTimePatternGenerator::setFieldDisplayName(UDateTimePatternField field, UDateTimePGDisplayWidth width, const UnicodeString& value) { - fieldDisplayNames[field][width] = value; - // NUL-terminate for the C API. - fieldDisplayNames[field][width].getTerminatedBuffer(); -} - -UnicodeString -DateTimePatternGenerator::getFieldDisplayName(UDateTimePatternField field, UDateTimePGDisplayWidth width) const { - return fieldDisplayNames[field][width]; -} - -UnicodeString& -DateTimePatternGenerator::getMutableFieldDisplayName(UDateTimePatternField field, UDateTimePGDisplayWidth width) { - return fieldDisplayNames[field][width]; -} - -void -DateTimePatternGenerator::getAppendName(UDateTimePatternField field, UnicodeString& value) { - value = SINGLE_QUOTE; - value += fieldDisplayNames[field][UDATPG_WIDTH_APPENDITEM]; - value += SINGLE_QUOTE; -} - -UnicodeString -DateTimePatternGenerator::getBestPattern(const UnicodeString& patternForm, UErrorCode& status) { - return getBestPattern(patternForm, UDATPG_MATCH_NO_OPTIONS, status); -} - -UnicodeString -DateTimePatternGenerator::getBestPattern(const UnicodeString& patternForm, UDateTimePatternMatchOptions options, UErrorCode& status) { - if (U_FAILURE(status)) { - return UnicodeString(); - } - if (U_FAILURE(internalErrorCode)) { - status = internalErrorCode; - return UnicodeString(); - } - const UnicodeString *bestPattern = nullptr; - UnicodeString dtFormat; - UnicodeString resultPattern; - int32_t flags = kDTPGNoFlags; - - int32_t dateMask=(1<set(patternFormMapped, fp); - const PtnSkeleton* specifiedSkeleton = nullptr; - bestPattern=getBestRaw(*dtMatcher, -1, distanceInfo, status, &specifiedSkeleton); - if (U_FAILURE(status)) { - return UnicodeString(); - } - - if ( distanceInfo->missingFieldMask==0 && distanceInfo->extraFieldMask==0 ) { - resultPattern = adjustFieldTypes(*bestPattern, specifiedSkeleton, flags, options); - - return resultPattern; - } - int32_t neededFields = dtMatcher->getFieldMask(); - UnicodeString datePattern=getBestAppending(neededFields & dateMask, flags, status, options); - UnicodeString timePattern=getBestAppending(neededFields & timeMask, flags, status, options); - if (U_FAILURE(status)) { - return UnicodeString(); - } - if (datePattern.length()==0) { - if (timePattern.length()==0) { - resultPattern.remove(); - } - else { - return timePattern; - } - } - if (timePattern.length()==0) { - return datePattern; - } - resultPattern.remove(); - status = U_ZERO_ERROR; - dtFormat=getDateTimeFormat(); - SimpleFormatter(dtFormat, 2, 2, status).format(timePattern, datePattern, resultPattern, status); - return resultPattern; -} - -/* - * Map a skeleton that may have metacharacters jJC to one without, by replacing - * the metacharacters with locale-appropriate fields of h/H/k/K and of a/b/B - * (depends on fDefaultHourFormatChar and fAllowedHourFormats being set, which in - * turn depends on initData having been run). This method also updates the flags - * as necessary. Returns the updated skeleton. - */ -UnicodeString -DateTimePatternGenerator::mapSkeletonMetacharacters(const UnicodeString& patternForm, int32_t* flags, UErrorCode& status) { - UnicodeString patternFormMapped; - patternFormMapped.remove(); - UBool inQuoted = FALSE; - int32_t patPos, patLen = patternForm.length(); - for (patPos = 0; patPos < patLen; patPos++) { - UChar patChr = patternForm.charAt(patPos); - if (patChr == SINGLE_QUOTE) { - inQuoted = !inQuoted; - } else if (!inQuoted) { - // Handle special mappings for 'j' and 'C' in which fields lengths - // 1,3,5 => hour field length 1 - // 2,4,6 => hour field length 2 - // 1,2 => abbreviated dayPeriod (field length 1..3) - // 3,4 => long dayPeriod (field length 4) - // 5,6 => narrow dayPeriod (field length 5) - if (patChr == LOW_J || patChr == CAP_C) { - int32_t extraLen = 0; // 1 less than total field length - while (patPos+1 < patLen && patternForm.charAt(patPos+1)==patChr) { - extraLen++; - patPos++; - } - int32_t hourLen = 1 + (extraLen & 1); - int32_t dayPeriodLen = (extraLen < 2)? 1: 3 + (extraLen >> 1); - UChar hourChar = LOW_H; - UChar dayPeriodChar = LOW_A; - if (patChr == LOW_J) { - hourChar = fDefaultHourFormatChar; - } else { - AllowedHourFormat preferred; - if (fAllowedHourFormats[0] != ALLOWED_HOUR_FORMAT_UNKNOWN) { - preferred = (AllowedHourFormat)fAllowedHourFormats[0]; - } else { - status = U_INVALID_FORMAT_ERROR; - return UnicodeString(); - } - if (preferred == ALLOWED_HOUR_FORMAT_H || preferred == ALLOWED_HOUR_FORMAT_HB || preferred == ALLOWED_HOUR_FORMAT_Hb) { - hourChar = CAP_H; - } - // in #13183 just add b/B to skeleton, no longer need to set special flags - if (preferred == ALLOWED_HOUR_FORMAT_HB || preferred == ALLOWED_HOUR_FORMAT_hB) { - dayPeriodChar = CAP_B; - } else if (preferred == ALLOWED_HOUR_FORMAT_Hb || preferred == ALLOWED_HOUR_FORMAT_hb) { - dayPeriodChar = LOW_B; - } - } - if (hourChar==CAP_H || hourChar==LOW_K) { - dayPeriodLen = 0; - } - while (dayPeriodLen-- > 0) { - patternFormMapped.append(dayPeriodChar); - } - while (hourLen-- > 0) { - patternFormMapped.append(hourChar); - } - } else if (patChr == CAP_J) { - // Get pattern for skeleton with H, then replace H or k - // with fDefaultHourFormatChar (if different) - patternFormMapped.append(CAP_H); - *flags |= kDTPGSkeletonUsesCapJ; - } else { - patternFormMapped.append(patChr); - } - } - } - return patternFormMapped; -} - -UnicodeString -DateTimePatternGenerator::replaceFieldTypes(const UnicodeString& pattern, - const UnicodeString& skeleton, - UErrorCode& status) { - return replaceFieldTypes(pattern, skeleton, UDATPG_MATCH_NO_OPTIONS, status); -} - -UnicodeString -DateTimePatternGenerator::replaceFieldTypes(const UnicodeString& pattern, - const UnicodeString& skeleton, - UDateTimePatternMatchOptions options, - UErrorCode& status) { - if (U_FAILURE(status)) { - return UnicodeString(); - } - if (U_FAILURE(internalErrorCode)) { - status = internalErrorCode; - return UnicodeString(); - } - dtMatcher->set(skeleton, fp); - UnicodeString result = adjustFieldTypes(pattern, nullptr, kDTPGNoFlags, options); - return result; -} - -void -DateTimePatternGenerator::setDecimal(const UnicodeString& newDecimal) { - this->decimal = newDecimal; - // NUL-terminate for the C API. - this->decimal.getTerminatedBuffer(); -} - -const UnicodeString& -DateTimePatternGenerator::getDecimal() const { - return decimal; -} - -void -DateTimePatternGenerator::addCanonicalItems(UErrorCode& status) { - if (U_FAILURE(status)) { return; } - UnicodeString conflictingPattern; - - for (int32_t i=0; i 0) { - addPattern(UnicodeString(Canonical_Items[i]), FALSE, conflictingPattern, status); - } - if (U_FAILURE(status)) { return; } - } -} - -void -DateTimePatternGenerator::setDateTimeFormat(const UnicodeString& dtFormat) { - dateTimeFormat = dtFormat; - // NUL-terminate for the C API. - dateTimeFormat.getTerminatedBuffer(); -} - -const UnicodeString& -DateTimePatternGenerator::getDateTimeFormat() const { - return dateTimeFormat; -} - -void -DateTimePatternGenerator::setDateTimeFromCalendar(const Locale& locale, UErrorCode& status) { - if (U_FAILURE(status)) { return; } - - const UChar *resStr; - int32_t resStrLen = 0; - - LocalPointer fCalendar(Calendar::createInstance(locale, status), status); - if (U_FAILURE(status)) { return; } - - LocalUResourceBundlePointer calData(ures_open(nullptr, locale.getBaseName(), &status)); - if (U_FAILURE(status)) { return; } - ures_getByKey(calData.getAlias(), DT_DateTimeCalendarTag, calData.getAlias(), &status); - if (U_FAILURE(status)) { return; } - - LocalUResourceBundlePointer dateTimePatterns; - if (fCalendar->getType() != nullptr && *fCalendar->getType() != '\0' - && uprv_strcmp(fCalendar->getType(), DT_DateTimeGregorianTag) != 0) { - dateTimePatterns.adoptInstead(ures_getByKeyWithFallback(calData.getAlias(), fCalendar->getType(), - nullptr, &status)); - ures_getByKeyWithFallback(dateTimePatterns.getAlias(), DT_DateTimePatternsTag, - dateTimePatterns.getAlias(), &status); - } - - if (dateTimePatterns.isNull() || status == U_MISSING_RESOURCE_ERROR) { - status = U_ZERO_ERROR; - dateTimePatterns.adoptInstead(ures_getByKeyWithFallback(calData.getAlias(), DT_DateTimeGregorianTag, - dateTimePatterns.orphan(), &status)); - ures_getByKeyWithFallback(dateTimePatterns.getAlias(), DT_DateTimePatternsTag, - dateTimePatterns.getAlias(), &status); - } - if (U_FAILURE(status)) { return; } - - if (ures_getSize(dateTimePatterns.getAlias()) <= DateFormat::kDateTime) - { - status = U_INVALID_FORMAT_ERROR; - return; - } - resStr = ures_getStringByIndex(dateTimePatterns.getAlias(), (int32_t)DateFormat::kDateTime, &resStrLen, &status); - setDateTimeFormat(UnicodeString(TRUE, resStr, resStrLen)); -} - -void -DateTimePatternGenerator::setDecimalSymbols(const Locale& locale, UErrorCode& status) { - DecimalFormatSymbols dfs = DecimalFormatSymbols(locale, status); - if(U_SUCCESS(status)) { - decimal = dfs.getSymbol(DecimalFormatSymbols::kDecimalSeparatorSymbol); - // NUL-terminate for the C API. - decimal.getTerminatedBuffer(); - } -} - -UDateTimePatternConflict -DateTimePatternGenerator::addPattern( - const UnicodeString& pattern, - UBool override, - UnicodeString &conflictingPattern, - UErrorCode& status) -{ - if (U_FAILURE(internalErrorCode)) { - status = internalErrorCode; - return UDATPG_NO_CONFLICT; - } - - return addPatternWithSkeleton(pattern, nullptr, override, conflictingPattern, status); -} - -// For DateTimePatternGenerator::addPatternWithSkeleton - -// If skeletonToUse is specified, then an availableFormats entry is being added. In this case: -// 1. We pass that skeleton to matcher.set instead of having it derive a skeleton from the pattern. -// 2. If the new entry's skeleton or basePattern does match an existing entry but that entry also had a skeleton specified -// (i.e. it was also from availableFormats), then the new entry does not override it regardless of the value of the override -// parameter. This prevents later availableFormats entries from a parent locale overriding earlier ones from the actual -// specified locale. However, availableFormats entries *should* override entries with matching skeleton whose skeleton was -// derived (i.e. entries derived from the standard date/time patters for the specified locale). -// 3. When adding the pattern (patternMap->add), we set a new boolean to indicate that the added entry had a -// specified skeleton (which sets a new field in the PtnElem in the PatternMap). -UDateTimePatternConflict -DateTimePatternGenerator::addPatternWithSkeleton( - const UnicodeString& pattern, - const UnicodeString* skeletonToUse, - UBool override, - UnicodeString& conflictingPattern, - UErrorCode& status) -{ - if (U_FAILURE(internalErrorCode)) { - status = internalErrorCode; - return UDATPG_NO_CONFLICT; - } - - UnicodeString basePattern; - PtnSkeleton skeleton; - UDateTimePatternConflict conflictingStatus = UDATPG_NO_CONFLICT; - - DateTimeMatcher matcher; - if ( skeletonToUse == nullptr ) { - matcher.set(pattern, fp, skeleton); - matcher.getBasePattern(basePattern); - } else { - matcher.set(*skeletonToUse, fp, skeleton); // no longer trims skeleton fields to max len 3, per #7930 - matcher.getBasePattern(basePattern); // or perhaps instead: basePattern = *skeletonToUse; - } - // We only care about base conflicts - and replacing the pattern associated with a base - if: - // 1. the conflicting previous base pattern did *not* have an explicit skeleton; in that case the previous - // base + pattern combination was derived from either (a) a canonical item, (b) a standard format, or - // (c) a pattern specified programmatically with a previous call to addPattern (which would only happen - // if we are getting here from a subsequent call to addPattern). - // 2. a skeleton is specified for the current pattern, but override=false; in that case we are checking - // availableFormats items from root, which should not override any previous entry with the same base. - UBool entryHadSpecifiedSkeleton; - const UnicodeString *duplicatePattern = patternMap->getPatternFromBasePattern(basePattern, entryHadSpecifiedSkeleton); - if (duplicatePattern != nullptr && (!entryHadSpecifiedSkeleton || (skeletonToUse != nullptr && !override))) { - conflictingStatus = UDATPG_BASE_CONFLICT; - conflictingPattern = *duplicatePattern; - if (!override) { - return conflictingStatus; - } - } - // The only time we get here with override=true and skeletonToUse!=null is when adding availableFormats - // items from CLDR data. In that case, we don't want an item from a parent locale to replace an item with - // same skeleton from the specified locale, so skip the current item if skeletonWasSpecified is true for - // the previously-specified conflicting item. - const PtnSkeleton* entrySpecifiedSkeleton = nullptr; - duplicatePattern = patternMap->getPatternFromSkeleton(skeleton, &entrySpecifiedSkeleton); - if (duplicatePattern != nullptr ) { - conflictingStatus = UDATPG_CONFLICT; - conflictingPattern = *duplicatePattern; - if (!override || (skeletonToUse != nullptr && entrySpecifiedSkeleton != nullptr)) { - return conflictingStatus; - } - } - patternMap->add(basePattern, skeleton, pattern, skeletonToUse != nullptr, status); - if(U_FAILURE(status)) { - return conflictingStatus; - } - - return UDATPG_NO_CONFLICT; -} - - -UDateTimePatternField -DateTimePatternGenerator::getAppendFormatNumber(const char* field) const { - for (int32_t i=0; i0; --i) { - if (uprv_strcmp(CLDR_FIELD_WIDTH[i], hyphenPtr)==0) { - *widthP=(UDateTimePGDisplayWidth)i; - break; - } - } - *hyphenPtr = 0; // now delete width portion of key - } - for (int32_t i=0; igetPatternFromSkeleton(*trial.getSkeletonPtr(), &specifiedSkeleton); - missingFields->setTo(tempInfo); - if (distance==0) { - break; - } - } - } - - // If the best raw match had a specified skeleton and that skeleton was requested by the caller, - // then return it too. This generally happens when the caller needs to pass that skeleton - // through to adjustFieldTypes so the latter can do a better job. - if (bestPattern && specifiedSkeletonPtr) { - *specifiedSkeletonPtr = specifiedSkeleton; - } - return bestPattern; -} - -UnicodeString -DateTimePatternGenerator::adjustFieldTypes(const UnicodeString& pattern, - const PtnSkeleton* specifiedSkeleton, - int32_t flags, - UDateTimePatternMatchOptions options) { - UnicodeString newPattern; - fp->set(pattern); - for (int32_t i=0; i < fp->itemNumber; i++) { - UnicodeString field = fp->items[i]; - if ( fp->isQuoteLiteral(field) ) { - - UnicodeString quoteLiteral; - fp->getQuoteLiteral(quoteLiteral, &i); - newPattern += quoteLiteral; - } - else { - if (fp->isPatternSeparator(field)) { - newPattern+=field; - continue; - } - int32_t canonicalIndex = fp->getCanonicalIndex(field); - if (canonicalIndex < 0) { - newPattern+=field; - continue; // don't adjust - } - const dtTypeElem *row = &dtTypes[canonicalIndex]; - int32_t typeValue = row->field; - - // handle day periods - with #13183, no longer need special handling here, integrated with normal types - - if ((flags & kDTPGFixFractionalSeconds) != 0 && typeValue == UDATPG_SECOND_FIELD) { - field += decimal; - dtMatcher->skeleton.original.appendFieldTo(UDATPG_FRACTIONAL_SECOND_FIELD, field); - } else if (dtMatcher->skeleton.type[typeValue]!=0) { - // Here: - // - "reqField" is the field from the originally requested skeleton, with length - // "reqFieldLen". - // - "field" is the field from the found pattern. - // - // The adjusted field should consist of characters from the originally requested - // skeleton, except in the case of UDATPG_HOUR_FIELD or UDATPG_MONTH_FIELD or - // UDATPG_WEEKDAY_FIELD or UDATPG_YEAR_FIELD, in which case it should consist - // of characters from the found pattern. - // - // The length of the adjusted field (adjFieldLen) should match that in the originally - // requested skeleton, except that in the following cases the length of the adjusted field - // should match that in the found pattern (i.e. the length of this pattern field should - // not be adjusted): - // 1. typeValue is UDATPG_HOUR_FIELD/MINUTE/SECOND and the corresponding bit in options is - // not set (ticket #7180). Note, we may want to implement a similar change for other - // numeric fields (MM, dd, etc.) so the default behavior is to get locale preference for - // field length, but options bits can be used to override this. - // 2. There is a specified skeleton for the found pattern and one of the following is true: - // a) The length of the field in the skeleton (skelFieldLen) is equal to reqFieldLen. - // b) The pattern field is numeric and the skeleton field is not, or vice versa. - - UChar reqFieldChar = dtMatcher->skeleton.original.getFieldChar(typeValue); - int32_t reqFieldLen = dtMatcher->skeleton.original.getFieldLength(typeValue); - if (reqFieldChar == CAP_E && reqFieldLen < 3) - reqFieldLen = 3; // 1-3 for E are equivalent to 3 for c,e - int32_t adjFieldLen = reqFieldLen; - if ( (typeValue==UDATPG_HOUR_FIELD && (options & UDATPG_MATCH_HOUR_FIELD_LENGTH)==0) || - (typeValue==UDATPG_MINUTE_FIELD && (options & UDATPG_MATCH_MINUTE_FIELD_LENGTH)==0) || - (typeValue==UDATPG_SECOND_FIELD && (options & UDATPG_MATCH_SECOND_FIELD_LENGTH)==0) ) { - adjFieldLen = field.length(); - } else if (specifiedSkeleton) { - int32_t skelFieldLen = specifiedSkeleton->original.getFieldLength(typeValue); - UBool patFieldIsNumeric = (row->type > 0); - UBool skelFieldIsNumeric = (specifiedSkeleton->type[typeValue] > 0); - if (skelFieldLen == reqFieldLen || (patFieldIsNumeric && !skelFieldIsNumeric) || (skelFieldIsNumeric && !patFieldIsNumeric)) { - // don't adjust the field length in the found pattern - adjFieldLen = field.length(); - } - } - UChar c = (typeValue!= UDATPG_HOUR_FIELD - && typeValue!= UDATPG_MONTH_FIELD - && typeValue!= UDATPG_WEEKDAY_FIELD - && (typeValue!= UDATPG_YEAR_FIELD || reqFieldChar==CAP_Y)) - ? reqFieldChar - : field.charAt(0); - if (typeValue == UDATPG_HOUR_FIELD && (flags & kDTPGSkeletonUsesCapJ) != 0) { - c = fDefaultHourFormatChar; - } - field.remove(); - for (int32_t j=adjFieldLen; j>0; --j) { - field += c; - } - } - newPattern+=field; - } - } - return newPattern; -} - -UnicodeString -DateTimePatternGenerator::getBestAppending(int32_t missingFields, int32_t flags, UErrorCode &status, UDateTimePatternMatchOptions options) { - if (U_FAILURE(status)) { - return UnicodeString(); - } - UnicodeString resultPattern, tempPattern; - const UnicodeString* tempPatternPtr; - int32_t lastMissingFieldMask=0; - if (missingFields!=0) { - resultPattern=UnicodeString(); - const PtnSkeleton* specifiedSkeleton=nullptr; - tempPatternPtr = getBestRaw(*dtMatcher, missingFields, distanceInfo, status, &specifiedSkeleton); - if (U_FAILURE(status)) { - return UnicodeString(); - } - tempPattern = *tempPatternPtr; - resultPattern = adjustFieldTypes(tempPattern, specifiedSkeleton, flags, options); - if ( distanceInfo->missingFieldMask==0 ) { - return resultPattern; - } - while (distanceInfo->missingFieldMask!=0) { // precondition: EVERY single field must work! - if ( lastMissingFieldMask == distanceInfo->missingFieldMask ) { - break; // cannot find the proper missing field - } - if (((distanceInfo->missingFieldMask & UDATPG_SECOND_AND_FRACTIONAL_MASK)==UDATPG_FRACTIONAL_MASK) && - ((missingFields & UDATPG_SECOND_AND_FRACTIONAL_MASK) == UDATPG_SECOND_AND_FRACTIONAL_MASK)) { - resultPattern = adjustFieldTypes(resultPattern, specifiedSkeleton, flags | kDTPGFixFractionalSeconds, options); - distanceInfo->missingFieldMask &= ~UDATPG_FRACTIONAL_MASK; - continue; - } - int32_t startingMask = distanceInfo->missingFieldMask; - tempPatternPtr = getBestRaw(*dtMatcher, distanceInfo->missingFieldMask, distanceInfo, status, &specifiedSkeleton); - if (U_FAILURE(status)) { - return UnicodeString(); - } - tempPattern = *tempPatternPtr; - tempPattern = adjustFieldTypes(tempPattern, specifiedSkeleton, flags, options); - int32_t foundMask=startingMask& ~distanceInfo->missingFieldMask; - int32_t topField=getTopBitNumber(foundMask); - - if (appendItemFormats[topField].length() != 0) { - UnicodeString appendName; - getAppendName((UDateTimePatternField)topField, appendName); - const UnicodeString *values[3] = { - &resultPattern, - &tempPattern, - &appendName - }; - SimpleFormatter(appendItemFormats[topField], 2, 3, status). - formatAndReplace(values, 3, resultPattern, nullptr, 0, status); - } - lastMissingFieldMask = distanceInfo->missingFieldMask; - } - } - return resultPattern; -} - -int32_t -DateTimePatternGenerator::getTopBitNumber(int32_t foundMask) const { - if ( foundMask==0 ) { - return 0; - } - int32_t i=0; - while (foundMask!=0) { - foundMask >>=1; - ++i; - } - if (i-1 >UDATPG_ZONE_FIELD) { - return UDATPG_ZONE_FIELD; - } - else - return i-1; -} - -void -DateTimePatternGenerator::setAvailableFormat(const UnicodeString &key, UErrorCode& err) -{ - fAvailableFormatKeyHash->puti(key, 1, err); -} - -UBool -DateTimePatternGenerator::isAvailableFormatSet(const UnicodeString &key) const { - return (UBool)(fAvailableFormatKeyHash->geti(key) == 1); -} - -void -DateTimePatternGenerator::copyHashtable(Hashtable *other, UErrorCode &status) { - if (other == nullptr || U_FAILURE(status)) { - return; - } - if (fAvailableFormatKeyHash != nullptr) { - delete fAvailableFormatKeyHash; - fAvailableFormatKeyHash = nullptr; - } - initHashtable(status); - if(U_FAILURE(status)){ - return; - } - int32_t pos = UHASH_FIRST; - const UHashElement* elem = nullptr; - // walk through the hash table and create a deep clone - while((elem = other->nextElement(pos))!= nullptr){ - const UHashTok otherKeyTok = elem->key; - UnicodeString* otherKey = (UnicodeString*)otherKeyTok.pointer; - fAvailableFormatKeyHash->puti(*otherKey, 1, status); - if(U_FAILURE(status)){ - return; - } - } -} - -StringEnumeration* -DateTimePatternGenerator::getSkeletons(UErrorCode& status) const { - if (U_FAILURE(status)) { - return nullptr; - } - if (U_FAILURE(internalErrorCode)) { - status = internalErrorCode; - return nullptr; - } - LocalPointer skeletonEnumerator( - new DTSkeletonEnumeration(*patternMap, DT_SKELETON, status), status); - - return U_SUCCESS(status) ? skeletonEnumerator.orphan() : nullptr; -} - -const UnicodeString& -DateTimePatternGenerator::getPatternForSkeleton(const UnicodeString& skeleton) const { - PtnElem *curElem; - - if (skeleton.length() ==0) { - return emptyString; - } - curElem = patternMap->getHeader(skeleton.charAt(0)); - while ( curElem != nullptr ) { - if ( curElem->skeleton->getSkeleton()==skeleton ) { - return curElem->pattern; - } - curElem = curElem->next.getAlias(); - } - return emptyString; -} - -StringEnumeration* -DateTimePatternGenerator::getBaseSkeletons(UErrorCode& status) const { - if (U_FAILURE(status)) { - return nullptr; - } - if (U_FAILURE(internalErrorCode)) { - status = internalErrorCode; - return nullptr; - } - LocalPointer baseSkeletonEnumerator( - new DTSkeletonEnumeration(*patternMap, DT_BASESKELETON, status), status); - - return U_SUCCESS(status) ? baseSkeletonEnumerator.orphan() : nullptr; -} - -StringEnumeration* -DateTimePatternGenerator::getRedundants(UErrorCode& status) { - if (U_FAILURE(status)) { return nullptr; } - if (U_FAILURE(internalErrorCode)) { - status = internalErrorCode; - return nullptr; - } - LocalPointer output(new DTRedundantEnumeration(), status); - if (U_FAILURE(status)) { return nullptr; } - const UnicodeString *pattern; - PatternMapIterator it(status); - if (U_FAILURE(status)) { return nullptr; } - - for (it.set(*patternMap); it.hasNext(); ) { - DateTimeMatcher current = it.next(); - pattern = patternMap->getPatternFromSkeleton(*(it.getSkeleton())); - if ( isCanonicalItem(*pattern) ) { - continue; - } - if ( skipMatcher == nullptr ) { - skipMatcher = new DateTimeMatcher(current); - if (skipMatcher == nullptr) { - status = U_MEMORY_ALLOCATION_ERROR; - return nullptr; - } - } - else { - *skipMatcher = current; - } - UnicodeString trial = getBestPattern(current.getPattern(), status); - if (U_FAILURE(status)) { return nullptr; } - if (trial == *pattern) { - ((DTRedundantEnumeration *)output.getAlias())->add(*pattern, status); - if (U_FAILURE(status)) { return nullptr; } - } - if (current.equals(skipMatcher)) { - continue; - } - } - return output.orphan(); -} - -UBool -DateTimePatternGenerator::isCanonicalItem(const UnicodeString& item) const { - if ( item.length() != 1 ) { - return FALSE; - } - for (int32_t i=0; iisDupAllowed = other.isDupAllowed; - for (int32_t bootIndex = 0; bootIndex < MAX_PATTERN_ENTRIES; ++bootIndex) { - PtnElem *curElem, *otherElem, *prevElem=nullptr; - otherElem = other.boot[bootIndex]; - while (otherElem != nullptr) { - LocalPointer newElem(new PtnElem(otherElem->basePattern, otherElem->pattern), status); - if (U_FAILURE(status)) { - return; // out of memory - } - newElem->skeleton.adoptInsteadAndCheckErrorCode(new PtnSkeleton(*(otherElem->skeleton)), status); - if (U_FAILURE(status)) { - return; // out of memory - } - newElem->skeletonWasSpecified = otherElem->skeletonWasSpecified; - - // Release ownership from the LocalPointer of the PtnElem object. - // The PtnElem will now be owned by either the boot (for the first entry in the linked-list) - // or owned by the previous PtnElem object in the linked-list. - curElem = newElem.orphan(); - - if (this->boot[bootIndex] == nullptr) { - this->boot[bootIndex] = curElem; - } else { - if (prevElem != nullptr) { - prevElem->next.adoptInstead(curElem); - } else { - U_ASSERT(false); - } - } - prevElem = curElem; - otherElem = otherElem->next.getAlias(); - } - - } -} - -PtnElem* -PatternMap::getHeader(UChar baseChar) const { - PtnElem* curElem; - - if ( (baseChar >= CAP_A) && (baseChar <= CAP_Z) ) { - curElem = boot[baseChar-CAP_A]; - } - else { - if ( (baseChar >=LOW_A) && (baseChar <= LOW_Z) ) { - curElem = boot[26+baseChar-LOW_A]; - } - else { - return nullptr; - } - } - return curElem; -} - -PatternMap::~PatternMap() { - for (int32_t i=0; i < MAX_PATTERN_ENTRIES; ++i ) { - if (boot[i] != nullptr ) { - delete boot[i]; - boot[i] = nullptr; - } - } -} // PatternMap destructor - -void -PatternMap::add(const UnicodeString& basePattern, - const PtnSkeleton& skeleton, - const UnicodeString& value,// mapped pattern value - UBool skeletonWasSpecified, - UErrorCode &status) { - UChar baseChar = basePattern.charAt(0); - PtnElem *curElem, *baseElem; - status = U_ZERO_ERROR; - - // the baseChar must be A-Z or a-z - if ((baseChar >= CAP_A) && (baseChar <= CAP_Z)) { - baseElem = boot[baseChar-CAP_A]; - } - else { - if ((baseChar >=LOW_A) && (baseChar <= LOW_Z)) { - baseElem = boot[26+baseChar-LOW_A]; - } - else { - status = U_ILLEGAL_CHARACTER; - return; - } - } - - if (baseElem == nullptr) { - LocalPointer newElem(new PtnElem(basePattern, value), status); - if (U_FAILURE(status)) { - return; // out of memory - } - newElem->skeleton.adoptInsteadAndCheckErrorCode(new PtnSkeleton(skeleton), status); - if (U_FAILURE(status)) { - return; // out of memory - } - newElem->skeletonWasSpecified = skeletonWasSpecified; - if (baseChar >= LOW_A) { - boot[26 + (baseChar - LOW_A)] = newElem.orphan(); // the boot array now owns the PtnElem. - } - else { - boot[baseChar - CAP_A] = newElem.orphan(); // the boot array now owns the PtnElem. - } - } - if ( baseElem != nullptr ) { - curElem = getDuplicateElem(basePattern, skeleton, baseElem); - - if (curElem == nullptr) { - // add new element to the list. - curElem = baseElem; - while( curElem -> next != nullptr ) - { - curElem = curElem->next.getAlias(); - } - - LocalPointer newElem(new PtnElem(basePattern, value), status); - if (U_FAILURE(status)) { - return; // out of memory - } - newElem->skeleton.adoptInsteadAndCheckErrorCode(new PtnSkeleton(skeleton), status); - if (U_FAILURE(status)) { - return; // out of memory - } - newElem->skeletonWasSpecified = skeletonWasSpecified; - curElem->next.adoptInstead(newElem.orphan()); - curElem = curElem->next.getAlias(); - } - else { - // Pattern exists in the list already. - if ( !isDupAllowed ) { - return; - } - // Overwrite the value. - curElem->pattern = value; - // It was a bug that we were not doing the following previously, - // though that bug hid other problems by making things partly work. - curElem->skeletonWasSpecified = skeletonWasSpecified; - } - } -} // PatternMap::add - -// Find the pattern from the given basePattern string. -const UnicodeString * -PatternMap::getPatternFromBasePattern(const UnicodeString& basePattern, UBool& skeletonWasSpecified) const { // key to search for - PtnElem *curElem; - - if ((curElem=getHeader(basePattern.charAt(0)))==nullptr) { - return nullptr; // no match - } - - do { - if ( basePattern.compare(curElem->basePattern)==0 ) { - skeletonWasSpecified = curElem->skeletonWasSpecified; - return &(curElem->pattern); - } - curElem = curElem->next.getAlias(); - } while (curElem != nullptr); - - return nullptr; -} // PatternMap::getFromBasePattern - - -// Find the pattern from the given skeleton. -// At least when this is called from getBestRaw & addPattern (in which case specifiedSkeletonPtr is non-NULL), -// the comparison should be based on skeleton.original (which is unique and tied to the distance measurement in bestRaw) -// and not skeleton.baseOriginal (which is not unique); otherwise we may pick a different skeleton than the one with the -// optimum distance value in getBestRaw. When this is called from public getRedundants (specifiedSkeletonPtr is NULL), -// for now it will continue to compare based on baseOriginal so as not to change the behavior unnecessarily. -const UnicodeString * -PatternMap::getPatternFromSkeleton(const PtnSkeleton& skeleton, const PtnSkeleton** specifiedSkeletonPtr) const { // key to search for - PtnElem *curElem; - - if (specifiedSkeletonPtr) { - *specifiedSkeletonPtr = nullptr; - } - - // find boot entry - UChar baseChar = skeleton.getFirstChar(); - if ((curElem=getHeader(baseChar))==nullptr) { - return nullptr; // no match - } - - do { - UBool equal; - if (specifiedSkeletonPtr != nullptr) { // called from DateTimePatternGenerator::getBestRaw or addPattern, use original - equal = curElem->skeleton->original == skeleton.original; - } else { // called from DateTimePatternGenerator::getRedundants, use baseOriginal - equal = curElem->skeleton->baseOriginal == skeleton.baseOriginal; - } - if (equal) { - if (specifiedSkeletonPtr && curElem->skeletonWasSpecified) { - *specifiedSkeletonPtr = curElem->skeleton.getAlias(); - } - return &(curElem->pattern); - } - curElem = curElem->next.getAlias(); - } while (curElem != nullptr); - - return nullptr; -} - -UBool -PatternMap::equals(const PatternMap& other) const { - if ( this==&other ) { - return TRUE; - } - for (int32_t bootIndex = 0; bootIndex < MAX_PATTERN_ENTRIES; ++bootIndex) { - if (boot[bootIndex] == other.boot[bootIndex]) { - continue; - } - if ((boot[bootIndex] == nullptr) || (other.boot[bootIndex] == nullptr)) { - return FALSE; - } - PtnElem *otherElem = other.boot[bootIndex]; - PtnElem *myElem = boot[bootIndex]; - while ((otherElem != nullptr) || (myElem != nullptr)) { - if ( myElem == otherElem ) { - break; - } - if ((otherElem == nullptr) || (myElem == nullptr)) { - return FALSE; - } - if ( (myElem->basePattern != otherElem->basePattern) || - (myElem->pattern != otherElem->pattern) ) { - return FALSE; - } - if ((myElem->skeleton.getAlias() != otherElem->skeleton.getAlias()) && - !myElem->skeleton->equals(*(otherElem->skeleton))) { - return FALSE; - } - myElem = myElem->next.getAlias(); - otherElem = otherElem->next.getAlias(); - } - } - return TRUE; -} - -// find any key existing in the mapping table already. -// return TRUE if there is an existing key, otherwise return FALSE. -PtnElem* -PatternMap::getDuplicateElem( - const UnicodeString &basePattern, - const PtnSkeleton &skeleton, - PtnElem *baseElem) { - PtnElem *curElem; - - if ( baseElem == nullptr ) { - return nullptr; - } - else { - curElem = baseElem; - } - do { - if ( basePattern.compare(curElem->basePattern)==0 ) { - UBool isEqual = TRUE; - for (int32_t i = 0; i < UDATPG_FIELD_COUNT; ++i) { - if (curElem->skeleton->type[i] != skeleton.type[i] ) { - isEqual = FALSE; - break; - } - } - if (isEqual) { - return curElem; - } - } - curElem = curElem->next.getAlias(); - } while( curElem != nullptr ); - - // end of the list - return nullptr; - -} // PatternMap::getDuplicateElem - -DateTimeMatcher::DateTimeMatcher(void) { -} - -DateTimeMatcher::~DateTimeMatcher() {} - -DateTimeMatcher::DateTimeMatcher(const DateTimeMatcher& other) { - copyFrom(other.skeleton); -} - - -void -DateTimeMatcher::set(const UnicodeString& pattern, FormatParser* fp) { - PtnSkeleton localSkeleton; - return set(pattern, fp, localSkeleton); -} - -void -DateTimeMatcher::set(const UnicodeString& pattern, FormatParser* fp, PtnSkeleton& skeletonResult) { - int32_t i; - for (i=0; iset(pattern); - for (i=0; i < fp->itemNumber; i++) { - const UnicodeString& value = fp->items[i]; - // don't skip 'a' anymore, dayPeriod handled specially below - - if ( fp->isQuoteLiteral(value) ) { - UnicodeString quoteLiteral; - fp->getQuoteLiteral(quoteLiteral, &i); - continue; - } - int32_t canonicalIndex = fp->getCanonicalIndex(value); - if (canonicalIndex < 0) { - continue; - } - const dtTypeElem *row = &dtTypes[canonicalIndex]; - int32_t field = row->field; - skeletonResult.original.populate(field, value); - UChar repeatChar = row->patternChar; - int32_t repeatCount = row->minLen; - skeletonResult.baseOriginal.populate(field, repeatChar, repeatCount); - int16_t subField = row->type; - if (row->type > 0) { - U_ASSERT(value.length() < INT16_MAX); - subField += static_cast(value.length()); - } - skeletonResult.type[field] = subField; - } - // #13183, handle special behavior for day period characters (a, b, B) - if (!skeletonResult.original.isFieldEmpty(UDATPG_HOUR_FIELD)) { - if (skeletonResult.original.getFieldChar(UDATPG_HOUR_FIELD)==LOW_H || skeletonResult.original.getFieldChar(UDATPG_HOUR_FIELD)==CAP_K) { - // We have a skeleton with 12-hour-cycle format - if (skeletonResult.original.isFieldEmpty(UDATPG_DAYPERIOD_FIELD)) { - // But we do not have a day period in the skeleton; add the default DAYPERIOD (currently "a") - for (i = 0; dtTypes[i].patternChar != 0; i++) { - if ( dtTypes[i].field == UDATPG_DAYPERIOD_FIELD ) { - // first entry for UDATPG_DAYPERIOD_FIELD - skeletonResult.original.populate(UDATPG_DAYPERIOD_FIELD, dtTypes[i].patternChar, dtTypes[i].minLen); - skeletonResult.baseOriginal.populate(UDATPG_DAYPERIOD_FIELD, dtTypes[i].patternChar, dtTypes[i].minLen); - skeletonResult.type[UDATPG_DAYPERIOD_FIELD] = dtTypes[i].type; - skeletonResult.addedDefaultDayPeriod = TRUE; - break; - } - } - } - } else { - // Skeleton has 24-hour-cycle hour format and has dayPeriod, delete dayPeriod (i.e. ignore it) - skeletonResult.original.clearField(UDATPG_DAYPERIOD_FIELD); - skeletonResult.baseOriginal.clearField(UDATPG_DAYPERIOD_FIELD); - skeletonResult.type[UDATPG_DAYPERIOD_FIELD] = NONE; - } - } - copyFrom(skeletonResult); -} - -void -DateTimeMatcher::getBasePattern(UnicodeString &result ) { - result.remove(); // Reset the result first. - skeleton.baseOriginal.appendTo(result); -} - -UnicodeString -DateTimeMatcher::getPattern() { - UnicodeString result; - return skeleton.original.appendTo(result); -} - -int32_t -DateTimeMatcher::getDistance(const DateTimeMatcher& other, int32_t includeMask, DistanceInfo& distanceInfo) const { - int32_t result = 0; - distanceInfo.clear(); - for (int32_t i=0; iskeleton.original; -} - -int32_t -DateTimeMatcher::getFieldMask() const { - int32_t result = 0; - - for (int32_t i=0; i= pattern.length()) { - return DONE; - } - // check the current char is between A-Z or a-z - do { - UChar c=pattern.charAt(curLoc); - if ( (c>=CAP_A && c<=CAP_Z) || (c>=LOW_A && c<=LOW_Z) ) { - curLoc++; - } - else { - startPos = curLoc; - *len=1; - return ADD_TOKEN; - } - - if ( pattern.charAt(curLoc)!= pattern.charAt(startPos) ) { - break; // not the same token - } - } while(curLoc <= pattern.length()); - *len = curLoc-startPos; - return ADD_TOKEN; -} - -void -FormatParser::set(const UnicodeString& pattern) { - int32_t startPos = 0; - TokenStatus result = START; - int32_t len = 0; - itemNumber = 0; - - do { - result = setTokens( pattern, startPos, &len ); - if ( result == ADD_TOKEN ) - { - items[itemNumber++] = UnicodeString(pattern, startPos, len ); - startPos += len; - } - else { - break; - } - } while (result==ADD_TOKEN && itemNumber < MAX_DT_TOKEN); -} - -int32_t -FormatParser::getCanonicalIndex(const UnicodeString& s, UBool strict) { - int32_t len = s.length(); - if (len == 0) { - return -1; - } - UChar ch = s.charAt(0); - - // Verify that all are the same character. - for (int32_t l = 1; l < len; l++) { - if (ch != s.charAt(l)) { - return -1; - } - } - int32_t i = 0; - int32_t bestRow = -1; - while (dtTypes[i].patternChar != 0x0000) { - if ( dtTypes[i].patternChar != ch ) { - ++i; - continue; - } - bestRow = i; - if (dtTypes[i].patternChar != dtTypes[i+1].patternChar) { - return i; - } - if (dtTypes[i+1].minLen <= len) { - ++i; - continue; - } - return i; - } - return strict ? -1 : bestRow; -} - -UBool -FormatParser::isQuoteLiteral(const UnicodeString& s) { - return (UBool)(s.charAt(0) == SINGLE_QUOTE); -} - -// This function assumes the current itemIndex points to the quote literal. -// Please call isQuoteLiteral prior to this function. -void -FormatParser::getQuoteLiteral(UnicodeString& quote, int32_t *itemIndex) { - int32_t i = *itemIndex; - - quote.remove(); - if (items[i].charAt(0)==SINGLE_QUOTE) { - quote += items[i]; - ++i; - } - while ( i < itemNumber ) { - if ( items[i].charAt(0)==SINGLE_QUOTE ) { - if ( (i+1patternMap=&newPatternMap; -} - -PtnSkeleton* -PatternMapIterator::getSkeleton() const { - if ( nodePtr == nullptr ) { - return nullptr; - } - else { - return nodePtr->skeleton.getAlias(); - } -} - -UBool -PatternMapIterator::hasNext() const { - int32_t headIndex = bootIndex; - PtnElem *curPtr = nodePtr; - - if (patternMap==nullptr) { - return FALSE; - } - while ( headIndex < MAX_PATTERN_ENTRIES ) { - if ( curPtr != nullptr ) { - if ( curPtr->next != nullptr ) { - return TRUE; - } - else { - headIndex++; - curPtr=nullptr; - continue; - } - } - else { - if ( patternMap->boot[headIndex] != nullptr ) { - return TRUE; - } - else { - headIndex++; - continue; - } - } - } - return FALSE; -} - -DateTimeMatcher& -PatternMapIterator::next() { - while ( bootIndex < MAX_PATTERN_ENTRIES ) { - if ( nodePtr != nullptr ) { - if ( nodePtr->next != nullptr ) { - nodePtr = nodePtr->next.getAlias(); - break; - } - else { - bootIndex++; - nodePtr=nullptr; - continue; - } - } - else { - if ( patternMap->boot[bootIndex] != nullptr ) { - nodePtr = patternMap->boot[bootIndex]; - break; - } - else { - bootIndex++; - continue; - } - } - } - if (nodePtr!=nullptr) { - matcher->copyFrom(*nodePtr->skeleton); - } - else { - matcher->copyFrom(); - } - return *matcher; -} - - -SkeletonFields::SkeletonFields() { - // Set initial values to zero - clear(); -} - -void SkeletonFields::clear() { - uprv_memset(chars, 0, sizeof(chars)); - uprv_memset(lengths, 0, sizeof(lengths)); -} - -void SkeletonFields::copyFrom(const SkeletonFields& other) { - uprv_memcpy(chars, other.chars, sizeof(chars)); - uprv_memcpy(lengths, other.lengths, sizeof(lengths)); -} - -void SkeletonFields::clearField(int32_t field) { - chars[field] = 0; - lengths[field] = 0; -} - -UChar SkeletonFields::getFieldChar(int32_t field) const { - return chars[field]; -} - -int32_t SkeletonFields::getFieldLength(int32_t field) const { - return lengths[field]; -} - -void SkeletonFields::populate(int32_t field, const UnicodeString& value) { - populate(field, value.charAt(0), value.length()); -} - -void SkeletonFields::populate(int32_t field, UChar ch, int32_t length) { - chars[field] = (int8_t) ch; - lengths[field] = (int8_t) length; -} - -UBool SkeletonFields::isFieldEmpty(int32_t field) const { - return lengths[field] == 0; -} - -UnicodeString& SkeletonFields::appendTo(UnicodeString& string) const { - for (int32_t i = 0; i < UDATPG_FIELD_COUNT; ++i) { - appendFieldTo(i, string); - } - return string; -} - -UnicodeString& SkeletonFields::appendFieldTo(int32_t field, UnicodeString& string) const { - UChar ch(chars[field]); - int32_t length = (int32_t) lengths[field]; - - for (int32_t i=0; i= 0) { - // for backward compatibility: if DateTimeMatcher.set added a single 'a' that - // was not in the provided skeleton, remove it here before returning skeleton. - result.remove(pos, 1); - } - return result; -} - -UnicodeString -PtnSkeleton::getBaseSkeleton() const { - UnicodeString result; - result = baseOriginal.appendTo(result); - int32_t pos; - if (addedDefaultDayPeriod && (pos = result.indexOf(LOW_A)) >= 0) { - // for backward compatibility: if DateTimeMatcher.set added a single 'a' that - // was not in the provided skeleton, remove it here before returning skeleton. - result.remove(pos, 1); - } - return result; -} - -UChar -PtnSkeleton::getFirstChar() const { - return baseOriginal.getFirstChar(); -} - -PtnSkeleton::~PtnSkeleton() { -} - -PtnElem::PtnElem(const UnicodeString &basePat, const UnicodeString &pat) : - basePattern(basePat), skeleton(nullptr), pattern(pat), next(nullptr) -{ -} - -PtnElem::~PtnElem() { -} - -DTSkeletonEnumeration::DTSkeletonEnumeration(PatternMap& patternMap, dtStrEnum type, UErrorCode& status) : fSkeletons(nullptr) { - PtnElem *curElem; - PtnSkeleton *curSkeleton; - UnicodeString s; - int32_t bootIndex; - - pos=0; - fSkeletons.adoptInsteadAndCheckErrorCode(new UVector(status), status); - if (U_FAILURE(status)) { - return; - } - - for (bootIndex=0; bootIndexbasePattern; - break; - case DT_PATTERN: - s=curElem->pattern; - break; - case DT_SKELETON: - curSkeleton=curElem->skeleton.getAlias(); - s=curSkeleton->getSkeleton(); - break; - } - if ( !isCanonicalItem(s) ) { - LocalPointer newElem(new UnicodeString(s), status); - if (U_FAILURE(status)) { - return; - } - fSkeletons->addElement(newElem.getAlias(), status); - if (U_FAILURE(status)) { - fSkeletons.adoptInstead(nullptr); - return; - } - newElem.orphan(); // fSkeletons vector now owns the UnicodeString. - } - curElem = curElem->next.getAlias(); - } - } - if ((bootIndex==MAX_PATTERN_ENTRIES) && (curElem!=nullptr) ) { - status = U_BUFFER_OVERFLOW_ERROR; - } -} - -const UnicodeString* -DTSkeletonEnumeration::snext(UErrorCode& status) { - if (U_SUCCESS(status) && fSkeletons.isValid() && pos < fSkeletons->size()) { - return (const UnicodeString*)fSkeletons->elementAt(pos++); - } - return nullptr; -} - -void -DTSkeletonEnumeration::reset(UErrorCode& /*status*/) { - pos=0; -} - -int32_t -DTSkeletonEnumeration::count(UErrorCode& /*status*/) const { - return (fSkeletons.isNull()) ? 0 : fSkeletons->size(); -} - -UBool -DTSkeletonEnumeration::isCanonicalItem(const UnicodeString& item) { - if ( item.length() != 1 ) { - return FALSE; - } - for (int32_t i=0; isize(); ++i) { - if ((s = (UnicodeString *)fSkeletons->elementAt(i)) != nullptr) { - delete s; - } - } - } -} - -DTRedundantEnumeration::DTRedundantEnumeration() : pos(0), fPatterns(nullptr) { -} - -void -DTRedundantEnumeration::add(const UnicodeString& pattern, UErrorCode& status) { - if (U_FAILURE(status)) { return; } - if (fPatterns.isNull()) { - fPatterns.adoptInsteadAndCheckErrorCode(new UVector(status), status); - if (U_FAILURE(status)) { - return; - } - } - LocalPointer newElem(new UnicodeString(pattern), status); - if (U_FAILURE(status)) { - return; - } - fPatterns->addElement(newElem.getAlias(), status); - if (U_FAILURE(status)) { - fPatterns.adoptInstead(nullptr); - return; - } - newElem.orphan(); // fPatterns now owns the string. -} - -const UnicodeString* -DTRedundantEnumeration::snext(UErrorCode& status) { - if (U_SUCCESS(status) && fPatterns.isValid() && pos < fPatterns->size()) { - return (const UnicodeString*)fPatterns->elementAt(pos++); - } - return nullptr; -} - -void -DTRedundantEnumeration::reset(UErrorCode& /*status*/) { - pos=0; -} - -int32_t -DTRedundantEnumeration::count(UErrorCode& /*status*/) const { - return (fPatterns.isNull()) ? 0 : fPatterns->size(); -} - -UBool -DTRedundantEnumeration::isCanonicalItem(const UnicodeString& item) const { - if ( item.length() != 1 ) { - return FALSE; - } - for (int32_t i=0; isize(); ++i) { - if ((s = (UnicodeString *)fPatterns->elementAt(i)) != nullptr) { - delete s; - } - } - } -} - -U_NAMESPACE_END - - -#endif /* #if !UCONFIG_NO_FORMATTING */ - -//eof diff --git a/tools/icu/patches/63/source/tools/toolutil/pkg_genc.cpp b/tools/icu/patches/63/source/tools/toolutil/pkg_genc.cpp deleted file mode 100644 index f23995bb21..0000000000 --- a/tools/icu/patches/63/source/tools/toolutil/pkg_genc.cpp +++ /dev/null @@ -1,1221 +0,0 @@ -// © 2016 and later: Unicode, Inc. and others. -// License & terms of use: http://www.unicode.org/copyright.html -/****************************************************************************** - * Copyright (C) 2009-2016, International Business Machines - * Corporation and others. All Rights Reserved. - ******************************************************************************* - */ -#include "unicode/utypes.h" - -#if U_PLATFORM_HAS_WIN32_API -# define VC_EXTRALEAN -# define WIN32_LEAN_AND_MEAN -# define NOUSER -# define NOSERVICE -# define NOIME -# define NOMCX -#include -#include -# ifdef __GNUC__ -# define WINDOWS_WITH_GNUC -# endif -#endif - -#if U_PLATFORM_IS_LINUX_BASED && U_HAVE_ELF_H -# define U_ELF -#endif - -#ifdef U_ELF -# include -# if defined(ELFCLASS64) -# define U_ELF64 -# endif - /* Old elf.h headers may not have EM_X86_64, or have EM_X8664 instead. */ -# ifndef EM_X86_64 -# define EM_X86_64 62 -# endif -# define ICU_ENTRY_OFFSET 0 -#endif - -#include -#include -#include "unicode/putil.h" -#include "cmemory.h" -#include "cstring.h" -#include "filestrm.h" -#include "toolutil.h" -#include "unicode/uclean.h" -#include "uoptions.h" -#include "pkg_genc.h" -#include "filetools.h" - -#define MAX_COLUMN ((uint32_t)(0xFFFFFFFFU)) - -#define HEX_0X 0 /* 0x1234 */ -#define HEX_0H 1 /* 01234h */ - -/* prototypes --------------------------------------------------------------- */ -static void -getOutFilename(const char *inFilename, const char *destdir, char *outFilename, char *entryName, const char *newSuffix, const char *optFilename); - -static uint32_t -write8(FileStream *out, uint8_t byte, uint32_t column); - -static uint32_t -write32(FileStream *out, uint32_t byte, uint32_t column); - -#if U_PLATFORM == U_PF_OS400 -static uint32_t -write8str(FileStream *out, uint8_t byte, uint32_t column); -#endif -/* -------------------------------------------------------------------------- */ - -/* -Creating Template Files for New Platforms - -Let the cc compiler help you get started. -Compile this program - const unsigned int x[5] = {1, 2, 0xdeadbeef, 0xffffffff, 16}; -with the -S option to produce assembly output. - -For example, this will generate array.s: -gcc -S array.c - -This will produce a .s file that may look like this: - - .file "array.c" - .version "01.01" -gcc2_compiled.: - .globl x - .section .rodata - .align 4 - .type x,@object - .size x,20 -x: - .long 1 - .long 2 - .long -559038737 - .long -1 - .long 16 - .ident "GCC: (GNU) 2.96 20000731 (Red Hat Linux 7.1 2.96-85)" - -which gives a starting point that will compile, and can be transformed -to become the template, generally with some consulting of as docs and -some experimentation. - -If you want ICU to automatically use this assembly, you should -specify "GENCCODE_ASSEMBLY=-a name" in the specific config/mh-* file, -where the name is the compiler or platform that you used in this -assemblyHeader data structure. -*/ -static const struct AssemblyType { - const char *name; - const char *header; - const char *beginLine; - const char *footer; - int8_t hexType; /* HEX_0X or HEX_0h */ -} assemblyHeader[] = { - /* For gcc assemblers, the meaning of .align changes depending on the */ - /* hardware, so we use .balign 16 which always means 16 bytes. */ - /* https://sourceware.org/binutils/docs/as/Pseudo-Ops.html */ - {"gcc", - ".globl %s\n" - "\t.section .note.GNU-stack,\"\",%%progbits\n" - "\t.section .rodata\n" - "\t.balign 16\n" - "#ifdef U_HIDE_DATA_SYMBOL\n" - "\t.hidden %s\n" - "#endif\n" - "\t.type %s,%%object\n" - "%s:\n\n", - - ".long ",".size %s, .-%s\n",HEX_0X - }, - {"gcc-darwin", - /*"\t.section __TEXT,__text,regular,pure_instructions\n" - "\t.section __TEXT,__picsymbolstub1,symbol_stubs,pure_instructions,32\n"*/ - ".globl _%s\n" - "#ifdef U_HIDE_DATA_SYMBOL\n" - "\t.private_extern _%s\n" - "#endif\n" - "\t.data\n" - "\t.const\n" - "\t.balign 16\n" - "_%s:\n\n", - - ".long ","",HEX_0X - }, - {"gcc-cygwin", - ".globl _%s\n" - "\t.section .rodata\n" - "\t.balign 16\n" - "_%s:\n\n", - - ".long ","",HEX_0X - }, - {"gcc-mingw64", - ".globl %s\n" - "\t.section .rodata\n" - "\t.balign 16\n" - "%s:\n\n", - - ".long ","",HEX_0X - }, -/* 16 bytes alignment. */ -/* http://docs.oracle.com/cd/E19641-01/802-1947/802-1947.pdf */ - {"sun", - "\t.section \".rodata\"\n" - "\t.align 16\n" - ".globl %s\n" - "%s:\n", - - ".word ","",HEX_0X - }, -/* 16 bytes alignment for sun-x86. */ -/* http://docs.oracle.com/cd/E19963-01/html/821-1608/eoiyg.html */ - {"sun-x86", - "Drodata.rodata:\n" - "\t.type Drodata.rodata,@object\n" - "\t.size Drodata.rodata,0\n" - "\t.globl %s\n" - "\t.align 16\n" - "%s:\n", - - ".4byte ","",HEX_0X - }, -/* 1<<4 bit alignment for aix. */ -/* http://pic.dhe.ibm.com/infocenter/aix/v6r1/index.jsp?topic=%2Fcom.ibm.aix.aixassem%2Fdoc%2Falangref%2Fidalangref_csect_pseudoop.htm */ - {"xlc", - ".globl %s{RO}\n" - "\t.toc\n" - "%s:\n" - "\t.csect %s{RO}, 4\n", - - ".long ","",HEX_0X - }, - {"aCC-ia64", - "\t.file \"%s.s\"\n" - "\t.type %s,@object\n" - "\t.global %s\n" - "\t.secalias .abe$0.rodata, \".rodata\"\n" - "\t.section .abe$0.rodata = \"a\", \"progbits\"\n" - "\t.align 16\n" - "%s::\t", - - "data4 ","",HEX_0X - }, - {"aCC-parisc", - "\t.SPACE $TEXT$\n" - "\t.SUBSPA $LIT$\n" - "%s\n" - "\t.EXPORT %s\n" - "\t.ALIGN 16\n", - - ".WORD ","",HEX_0X - }, -/* align 16 bytes */ -/* http://msdn.microsoft.com/en-us/library/dwa9fwef.aspx */ - { "masm", - "\tTITLE %s\n" - "; generated by genccode\n" - ".386\n" - ".model flat\n" - "\tPUBLIC _%s\n" - "ICUDATA_%s\tSEGMENT READONLY PARA PUBLIC FLAT 'DATA'\n" - "\tALIGN 16\n" - "_%s\tLABEL DWORD\n", - "\tDWORD ","\nICUDATA_%s\tENDS\n\tEND\n",HEX_0H - } -}; - -static int32_t assemblyHeaderIndex = -1; -static int32_t hexType = HEX_0X; - -U_CAPI UBool U_EXPORT2 -checkAssemblyHeaderName(const char* optAssembly) { - int32_t idx; - assemblyHeaderIndex = -1; - for (idx = 0; idx < UPRV_LENGTHOF(assemblyHeader); idx++) { - if (uprv_strcmp(optAssembly, assemblyHeader[idx].name) == 0) { - assemblyHeaderIndex = idx; - hexType = assemblyHeader[idx].hexType; /* set the hex type */ - return TRUE; - } - } - - return FALSE; -} - - -U_CAPI void U_EXPORT2 -printAssemblyHeadersToStdErr(void) { - int32_t idx; - fprintf(stderr, "%s", assemblyHeader[0].name); - for (idx = 1; idx < UPRV_LENGTHOF(assemblyHeader); idx++) { - fprintf(stderr, ", %s", assemblyHeader[idx].name); - } - fprintf(stderr, - ")\n"); -} - -U_CAPI void U_EXPORT2 -writeAssemblyCode(const char *filename, const char *destdir, const char *optEntryPoint, const char *optFilename, char *outFilePath) { - uint32_t column = MAX_COLUMN; - char entry[64]; - uint32_t buffer[1024]; - char *bufferStr = (char *)buffer; - FileStream *in, *out; - size_t i, length; - - in=T_FileStream_open(filename, "rb"); - if(in==NULL) { - fprintf(stderr, "genccode: unable to open input file %s\n", filename); - exit(U_FILE_ACCESS_ERROR); - } - - getOutFilename(filename, destdir, bufferStr, entry, ".S", optFilename); - out=T_FileStream_open(bufferStr, "w"); - if(out==NULL) { - fprintf(stderr, "genccode: unable to open output file %s\n", bufferStr); - exit(U_FILE_ACCESS_ERROR); - } - - if (outFilePath != NULL) { - uprv_strcpy(outFilePath, bufferStr); - } - -#if defined (WINDOWS_WITH_GNUC) && U_PLATFORM != U_PF_CYGWIN - /* Need to fix the file separator character when using MinGW. */ - swapFileSepChar(outFilePath, U_FILE_SEP_CHAR, '/'); -#endif - - if(optEntryPoint != NULL) { - uprv_strcpy(entry, optEntryPoint); - uprv_strcat(entry, "_dat"); - } - - /* turn dashes or dots in the entry name into underscores */ - length=uprv_strlen(entry); - for(i=0; i= 0 ; i--) -#endif - { - uint8_t value = ptrIdx[i]; - if (value || seenNonZero) { - *(s++)=hexToStr[value>>4]; - *(s++)=hexToStr[value&0xF]; - seenNonZero = 1; - } - } - if(hexType==HEX_0H) { - *(s++)='h'; - } - } - - *(s++)=0; - T_FileStream_writeLine(out, bitFieldStr); - return column; -} - -static uint32_t -write8(FileStream *out, uint8_t byte, uint32_t column) { - char s[4]; - int i=0; - - /* convert the byte value to a string */ - if(byte>=100) { - s[i++]=(char)('0'+byte/100); - byte%=100; - } - if(i>0 || byte>=10) { - s[i++]=(char)('0'+byte/10); - byte%=10; - } - s[i++]=(char)('0'+byte); - s[i]=0; - - /* write the value, possibly with comma and newline */ - if(column==MAX_COLUMN) { - /* first byte */ - column=1; - } else if(column<16) { - T_FileStream_writeLine(out, ","); - ++column; - } else { - T_FileStream_writeLine(out, ",\n"); - column=1; - } - T_FileStream_writeLine(out, s); - return column; -} - -#if U_PLATFORM == U_PF_OS400 -static uint32_t -write8str(FileStream *out, uint8_t byte, uint32_t column) { - char s[8]; - - if (byte > 7) - sprintf(s, "\\x%X", byte); - else - sprintf(s, "\\%X", byte); - - /* write the value, possibly with comma and newline */ - if(column==MAX_COLUMN) { - /* first byte */ - column=1; - T_FileStream_writeLine(out, "\""); - } else if(column<24) { - ++column; - } else { - T_FileStream_writeLine(out, "\"\n\""); - column=1; - } - T_FileStream_writeLine(out, s); - return column; -} -#endif - -static void -getOutFilename(const char *inFilename, const char *destdir, char *outFilename, char *entryName, const char *newSuffix, const char *optFilename) { - const char *basename=findBasename(inFilename), *suffix=uprv_strrchr(basename, '.'); - - /* copy path */ - if(destdir!=NULL && *destdir!=0) { - do { - *outFilename++=*destdir++; - } while(*destdir!=0); - if(*(outFilename-1)!=U_FILE_SEP_CHAR) { - *outFilename++=U_FILE_SEP_CHAR; - } - inFilename=basename; - } else { - while(inFilename macros are predefined by the MSVC compiler based - // on the target compilation architecture. - // https://docs.microsoft.com/cpp/preprocessor/predefined-macros - - // link.exe will link an IMAGE_FILE_MACHINE_UNKNOWN data-only .obj file - // no matter what architecture it is targeting (though other values are - // required to match). Unfortunately, the variable name decoration/mangling - // is slightly different on x86, which means we can't use the UNKNOWN type - // for all architectures though. -# if defined(_M_IX86) - *pCPU = IMAGE_FILE_MACHINE_I386; -# else - *pCPU = IMAGE_FILE_MACHINE_UNKNOWN; -# endif -# if defined(_M_IA64) || defined(_M_AMD64) || defined (_M_ARM64) - *pBits = 64; // Doesn't seem to be used for anything interesting though? -# elif defined(_M_IX86) || defined(_M_ARM) - *pBits = 32; -# else -# error "Unknown platform for CAN_GENERATE_OBJECTS." -# endif -#else -# error "Unknown platform for CAN_GENERATE_OBJECTS." -#endif - return; - } - - in=T_FileStream_open(filename, "rb"); - if(in==NULL) { - fprintf(stderr, "genccode: unable to open match-arch file %s\n", filename); - exit(U_FILE_ACCESS_ERROR); - } - length=T_FileStream_read(in, buffer.bytes, sizeof(buffer.bytes)); - -#ifdef U_ELF - if(length<(int32_t)sizeof(Elf32_Ehdr)) { - fprintf(stderr, "genccode: match-arch file %s is too short\n", filename); - exit(U_UNSUPPORTED_ERROR); - } - if( - buffer.header32.e_ident[0]!=ELFMAG0 || - buffer.header32.e_ident[1]!=ELFMAG1 || - buffer.header32.e_ident[2]!=ELFMAG2 || - buffer.header32.e_ident[3]!=ELFMAG3 || - buffer.header32.e_ident[EI_CLASS]ELFCLASS64 - ) { - fprintf(stderr, "genccode: match-arch file %s is not an ELF object file, or not supported\n", filename); - exit(U_UNSUPPORTED_ERROR); - } - - *pBits= buffer.header32.e_ident[EI_CLASS]==ELFCLASS32 ? 32 : 64; /* only 32 or 64: see check above */ -#ifdef U_ELF64 - if(*pBits!=32 && *pBits!=64) { - fprintf(stderr, "genccode: currently only supports 32-bit and 64-bit ELF format\n"); - exit(U_UNSUPPORTED_ERROR); - } -#else - if(*pBits!=32) { - fprintf(stderr, "genccode: built with elf.h missing 64-bit definitions\n"); - exit(U_UNSUPPORTED_ERROR); - } -#endif - - *pIsBigEndian=(UBool)(buffer.header32.e_ident[EI_DATA]==ELFDATA2MSB); - if(*pIsBigEndian!=U_IS_BIG_ENDIAN) { - fprintf(stderr, "genccode: currently only same-endianness ELF formats are supported\n"); - exit(U_UNSUPPORTED_ERROR); - } - /* TODO: Support byte swapping */ - - *pCPU=buffer.header32.e_machine; -#elif U_PLATFORM_HAS_WIN32_API - if(lengthMachine; - /* - * The number of bits is implicit with the Machine value. - * *pBits is ignored in the calling code, so this need not be precise. - */ - *pBits= *pCPU==IMAGE_FILE_MACHINE_I386 ? 32 : 64; - /* Windows always runs on little-endian CPUs. */ - *pIsBigEndian=FALSE; -#else -# error "Unknown platform for CAN_GENERATE_OBJECTS." -#endif - - T_FileStream_close(in); -} - -U_CAPI void U_EXPORT2 -writeObjectCode(const char *filename, const char *destdir, const char *optEntryPoint, const char *optMatchArch, const char *optFilename, char *outFilePath) { - /* common variables */ - char buffer[4096], entry[96]={ 0 }; - FileStream *in, *out; - const char *newSuffix; - int32_t i, entryLength, length, size, entryOffset=0, entryLengthOffset=0; - - uint16_t cpu, bits; - UBool makeBigEndian; - - /* platform-specific variables and initialization code */ -#ifdef U_ELF - /* 32-bit Elf file header */ - static Elf32_Ehdr header32={ - { - /* e_ident[] */ - ELFMAG0, ELFMAG1, ELFMAG2, ELFMAG3, - ELFCLASS32, - U_IS_BIG_ENDIAN ? ELFDATA2MSB : ELFDATA2LSB, - EV_CURRENT /* EI_VERSION */ - }, - ET_REL, - EM_386, - EV_CURRENT, /* e_version */ - 0, /* e_entry */ - 0, /* e_phoff */ - (Elf32_Off)sizeof(Elf32_Ehdr), /* e_shoff */ - 0, /* e_flags */ - (Elf32_Half)sizeof(Elf32_Ehdr), /* eh_size */ - 0, /* e_phentsize */ - 0, /* e_phnum */ - (Elf32_Half)sizeof(Elf32_Shdr), /* e_shentsize */ - 5, /* e_shnum */ - 2 /* e_shstrndx */ - }; - - /* 32-bit Elf section header table */ - static Elf32_Shdr sectionHeaders32[5]={ - { /* SHN_UNDEF */ - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 - }, - { /* .symtab */ - 1, /* sh_name */ - SHT_SYMTAB, - 0, /* sh_flags */ - 0, /* sh_addr */ - (Elf32_Off)(sizeof(header32)+sizeof(sectionHeaders32)), /* sh_offset */ - (Elf32_Word)(2*sizeof(Elf32_Sym)), /* sh_size */ - 3, /* sh_link=sect hdr index of .strtab */ - 1, /* sh_info=One greater than the symbol table index of the last - * local symbol (with STB_LOCAL). */ - 4, /* sh_addralign */ - (Elf32_Word)(sizeof(Elf32_Sym)) /* sh_entsize */ - }, - { /* .shstrtab */ - 9, /* sh_name */ - SHT_STRTAB, - 0, /* sh_flags */ - 0, /* sh_addr */ - (Elf32_Off)(sizeof(header32)+sizeof(sectionHeaders32)+2*sizeof(Elf32_Sym)), /* sh_offset */ - 40, /* sh_size */ - 0, /* sh_link */ - 0, /* sh_info */ - 1, /* sh_addralign */ - 0 /* sh_entsize */ - }, - { /* .strtab */ - 19, /* sh_name */ - SHT_STRTAB, - 0, /* sh_flags */ - 0, /* sh_addr */ - (Elf32_Off)(sizeof(header32)+sizeof(sectionHeaders32)+2*sizeof(Elf32_Sym)+40), /* sh_offset */ - (Elf32_Word)sizeof(entry), /* sh_size */ - 0, /* sh_link */ - 0, /* sh_info */ - 1, /* sh_addralign */ - 0 /* sh_entsize */ - }, - { /* .rodata */ - 27, /* sh_name */ - SHT_PROGBITS, - SHF_ALLOC, /* sh_flags */ - 0, /* sh_addr */ - (Elf32_Off)(sizeof(header32)+sizeof(sectionHeaders32)+2*sizeof(Elf32_Sym)+40+sizeof(entry)), /* sh_offset */ - 0, /* sh_size */ - 0, /* sh_link */ - 0, /* sh_info */ - 16, /* sh_addralign */ - 0 /* sh_entsize */ - } - }; - - /* symbol table */ - static Elf32_Sym symbols32[2]={ - { /* STN_UNDEF */ - 0, 0, 0, 0, 0, 0 - }, - { /* data entry point */ - 1, /* st_name */ - 0, /* st_value */ - 0, /* st_size */ - ELF64_ST_INFO(STB_GLOBAL, STT_OBJECT), - 0, /* st_other */ - 4 /* st_shndx=index of related section table entry */ - } - }; - - /* section header string table, with decimal string offsets */ - static const char sectionStrings[40]= - /* 0 */ "\0" - /* 1 */ ".symtab\0" - /* 9 */ ".shstrtab\0" - /* 19 */ ".strtab\0" - /* 27 */ ".rodata\0" - /* 35 */ "\0\0\0\0"; /* contains terminating NUL */ - /* 40: padded to multiple of 8 bytes */ - - /* - * Use entry[] for the string table which will contain only the - * entry point name. - * entry[0] must be 0 (NUL) - * The entry point name can be up to 38 characters long (sizeof(entry)-2). - */ - - /* 16-align .rodata in the .o file, just in case */ - static const char padding[16]={ 0 }; - int32_t paddingSize; - -#ifdef U_ELF64 - /* 64-bit Elf file header */ - static Elf64_Ehdr header64={ - { - /* e_ident[] */ - ELFMAG0, ELFMAG1, ELFMAG2, ELFMAG3, - ELFCLASS64, - U_IS_BIG_ENDIAN ? ELFDATA2MSB : ELFDATA2LSB, - EV_CURRENT /* EI_VERSION */ - }, - ET_REL, - EM_X86_64, - EV_CURRENT, /* e_version */ - 0, /* e_entry */ - 0, /* e_phoff */ - (Elf64_Off)sizeof(Elf64_Ehdr), /* e_shoff */ - 0, /* e_flags */ - (Elf64_Half)sizeof(Elf64_Ehdr), /* eh_size */ - 0, /* e_phentsize */ - 0, /* e_phnum */ - (Elf64_Half)sizeof(Elf64_Shdr), /* e_shentsize */ - 5, /* e_shnum */ - 2 /* e_shstrndx */ - }; - - /* 64-bit Elf section header table */ - static Elf64_Shdr sectionHeaders64[5]={ - { /* SHN_UNDEF */ - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 - }, - { /* .symtab */ - 1, /* sh_name */ - SHT_SYMTAB, - 0, /* sh_flags */ - 0, /* sh_addr */ - (Elf64_Off)(sizeof(header64)+sizeof(sectionHeaders64)), /* sh_offset */ - (Elf64_Xword)(2*sizeof(Elf64_Sym)), /* sh_size */ - 3, /* sh_link=sect hdr index of .strtab */ - 1, /* sh_info=One greater than the symbol table index of the last - * local symbol (with STB_LOCAL). */ - 4, /* sh_addralign */ - (Elf64_Xword)(sizeof(Elf64_Sym)) /* sh_entsize */ - }, - { /* .shstrtab */ - 9, /* sh_name */ - SHT_STRTAB, - 0, /* sh_flags */ - 0, /* sh_addr */ - (Elf64_Off)(sizeof(header64)+sizeof(sectionHeaders64)+2*sizeof(Elf64_Sym)), /* sh_offset */ - 40, /* sh_size */ - 0, /* sh_link */ - 0, /* sh_info */ - 1, /* sh_addralign */ - 0 /* sh_entsize */ - }, - { /* .strtab */ - 19, /* sh_name */ - SHT_STRTAB, - 0, /* sh_flags */ - 0, /* sh_addr */ - (Elf64_Off)(sizeof(header64)+sizeof(sectionHeaders64)+2*sizeof(Elf64_Sym)+40), /* sh_offset */ - (Elf64_Xword)sizeof(entry), /* sh_size */ - 0, /* sh_link */ - 0, /* sh_info */ - 1, /* sh_addralign */ - 0 /* sh_entsize */ - }, - { /* .rodata */ - 27, /* sh_name */ - SHT_PROGBITS, - SHF_ALLOC, /* sh_flags */ - 0, /* sh_addr */ - (Elf64_Off)(sizeof(header64)+sizeof(sectionHeaders64)+2*sizeof(Elf64_Sym)+40+sizeof(entry)), /* sh_offset */ - 0, /* sh_size */ - 0, /* sh_link */ - 0, /* sh_info */ - 16, /* sh_addralign */ - 0 /* sh_entsize */ - } - }; - - /* - * 64-bit symbol table - * careful: different order of items compared with Elf32_sym! - */ - static Elf64_Sym symbols64[2]={ - { /* STN_UNDEF */ - 0, 0, 0, 0, 0, 0 - }, - { /* data entry point */ - 1, /* st_name */ - ELF64_ST_INFO(STB_GLOBAL, STT_OBJECT), - 0, /* st_other */ - 4, /* st_shndx=index of related section table entry */ - 0, /* st_value */ - 0 /* st_size */ - } - }; - -#endif /* U_ELF64 */ - - /* entry[] have a leading NUL */ - entryOffset=1; - - /* in the common code, count entryLength from after the NUL */ - entryLengthOffset=1; - - newSuffix=".o"; - -#elif U_PLATFORM_HAS_WIN32_API - struct { - IMAGE_FILE_HEADER fileHeader; - IMAGE_SECTION_HEADER sections[2]; - char linkerOptions[100]; - } objHeader; - IMAGE_SYMBOL symbols[1]; - struct { - DWORD sizeofLongNames; - char longNames[100]; - } symbolNames; - - /* - * entry sometimes have a leading '_' - * overwritten if entryOffset==0 depending on the target platform - * see check for cpu below - */ - entry[0]='_'; - - newSuffix=".obj"; -#else -# error "Unknown platform for CAN_GENERATE_OBJECTS." -#endif - - /* deal with options, files and the entry point name */ - getArchitecture(&cpu, &bits, &makeBigEndian, optMatchArch); - if (optMatchArch) - { - printf("genccode: --match-arch cpu=%hu bits=%hu big-endian=%d\n", cpu, bits, makeBigEndian); - } - else - { - printf("genccode: using architecture cpu=%hu bits=%hu big-endian=%d\n", cpu, bits, makeBigEndian); - } -#if U_PLATFORM_HAS_WIN32_API - if(cpu==IMAGE_FILE_MACHINE_I386) { - entryOffset=1; - } -#endif - - in=T_FileStream_open(filename, "rb"); - if(in==NULL) { - fprintf(stderr, "genccode: unable to open input file %s\n", filename); - exit(U_FILE_ACCESS_ERROR); - } - size=T_FileStream_size(in); - - getOutFilename(filename, destdir, buffer, entry+entryOffset, newSuffix, optFilename); - if (outFilePath != NULL) { - uprv_strcpy(outFilePath, buffer); - } - - if(optEntryPoint != NULL) { - uprv_strcpy(entry+entryOffset, optEntryPoint); - uprv_strcat(entry+entryOffset, "_dat"); - } - /* turn dashes in the entry name into underscores */ - entryLength=(int32_t)uprv_strlen(entry+entryLengthOffset); - for(i=0; i