diff options
Diffstat (limited to 'lib/uniwbrk.h')
-rw-r--r-- | lib/uniwbrk.h | 103 |
1 files changed, 0 insertions, 103 deletions
diff --git a/lib/uniwbrk.h b/lib/uniwbrk.h deleted file mode 100644 index 416f47ea..00000000 --- a/lib/uniwbrk.h +++ /dev/null @@ -1,103 +0,0 @@ -/* DO NOT EDIT! GENERATED AUTOMATICALLY! */ -/* Word breaks in Unicode strings. - Copyright (C) 2001-2003, 2005-2018 Free Software Foundation, Inc. - Written by Bruno Haible <bruno@clisp.org>, 2009. - - This program is free software: you can redistribute it and/or - modify it under the terms of either: - - * the GNU Lesser General Public License as published by the Free - Software Foundation; either version 3 of the License, or (at your - option) any later version. - - or - - * the GNU General Public License as published by the Free - Software Foundation; either version 2 of the License, or (at your - option) any later version. - - or both in parallel, as here. - This program is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - Lesser General Public License for more details. - - You should have received a copy of the GNU Lesser General Public License - along with this program. If not, see <https://www.gnu.org/licenses/>. */ - -#ifndef _UNIWBRK_H -#define _UNIWBRK_H - -/* Get size_t. */ -#include <stddef.h> - -#include "unitypes.h" - - -#ifdef __cplusplus -extern "C" { -#endif - -/* ========================================================================= */ - -/* Property defined in Unicode Standard Annex #29, section "Word Boundaries" - <http://www.unicode.org/reports/tr29/#Word_Boundaries> */ - -/* Possible values of the Word_Break property. - This enumeration may be extended in the future. */ -enum -{ - WBP_OTHER = 0, - WBP_CR = 11, - WBP_LF = 12, - WBP_NEWLINE = 10, - WBP_EXTEND = 8, - WBP_FORMAT = 9, - WBP_KATAKANA = 1, - WBP_ALETTER = 2, - WBP_MIDNUMLET = 3, - WBP_MIDLETTER = 4, - WBP_MIDNUM = 5, - WBP_NUMERIC = 6, - WBP_EXTENDNUMLET = 7, - WBP_RI = 13, - WBP_DQ = 14, - WBP_SQ = 15, - WBP_HL = 16, - WBP_ZWJ = 17, - WBP_EB = 18, - WBP_EM = 19, - WBP_GAZ = 20, - WBP_EBG = 21 -}; - -/* Return the Word_Break property of a Unicode character. */ -extern int - uc_wordbreak_property (ucs4_t uc) - _UC_ATTRIBUTE_CONST; - -/* ========================================================================= */ - -/* Word breaks. */ - -/* Determine the word break points in S, and store the result at p[0..n-1]. - p[i] = 1 means that there is a word boundary between s[i-1] and s[i]. - p[i] = 0 means that s[i-1] and s[i] must not be separated. - */ -extern void - u8_wordbreaks (const uint8_t *s, size_t n, char *p); -extern void - u16_wordbreaks (const uint16_t *s, size_t n, char *p); -extern void - u32_wordbreaks (const uint32_t *s, size_t n, char *p); -extern void - ulc_wordbreaks (const char *s, size_t n, char *p); - -/* ========================================================================= */ - -#ifdef __cplusplus -} -#endif - - -#endif /* _UNIWBRK_H */ |