Ruby: regenc.h Source File

00001 #ifndef ONIGURUMA_REGENC_H
00002 #define ONIGURUMA_REGENC_H
00003 /**********************************************************************
00004   regenc.h -  Onigmo (Oniguruma-mod) (regular expression library)
00005 **********************************************************************/
00006 /*-
00007  * Copyright (c) 2002-2008  K.Kosako  <sndgk393 AT ybb DOT ne DOT jp>
00008  * Copyright (c) 2011       K.Takata  <kentkt AT csc DOT jp>
00009  * All rights reserved.
00010  *
00011  * Redistribution and use in source and binary forms, with or without
00012  * modification, are permitted provided that the following conditions
00013  * are met:
00014  * 1. Redistributions of source code must retain the above copyright
00015  *    notice, this list of conditions and the following disclaimer.
00016  * 2. Redistributions in binary form must reproduce the above copyright
00017  *    notice, this list of conditions and the following disclaimer in the
00018  *    documentation and/or other materials provided with the distribution.
00019  *
00020  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
00021  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
00022  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
00023  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
00024  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
00025  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
00026  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
00027  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
00028  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
00029  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
00030  * SUCH DAMAGE.
00031  */
00032 #ifndef REGINT_H
00033 #ifndef RUBY_EXTERN
00034 #include "ruby/config.h"
00035 #include "ruby/defines.h"
00036 #endif
00037 #ifdef ONIG_ESCAPE_UCHAR_COLLISION
00038 #undef ONIG_ESCAPE_UCHAR_COLLISION
00039 #endif
00040 #endif
00041 #include "ruby/oniguruma.h"
00042 
00043 RUBY_SYMBOL_EXPORT_BEGIN
00044 
00045 typedef struct {
00046   OnigCodePoint from;
00047   OnigCodePoint to;
00048 } OnigPairCaseFoldCodes;
00049 
00050 
00051 #ifndef NULL
00052 #define NULL   ((void* )0)
00053 #endif
00054 
00055 #ifndef TRUE
00056 #define TRUE    1
00057 #endif
00058 
00059 #ifndef FALSE
00060 #define FALSE   0
00061 #endif
00062 
00063 #ifndef ARG_UNUSED
00064 #if defined(__GNUC__)
00065 #  define ARG_UNUSED  __attribute__ ((unused))
00066 #else
00067 #  define ARG_UNUSED
00068 #endif
00069 #endif
00070 
00071 #define ONIG_IS_NULL(p)                    (((void*)(p)) == (void*)0)
00072 #define ONIG_IS_NOT_NULL(p)                (((void*)(p)) != (void*)0)
00073 #define ONIG_CHECK_NULL_RETURN(p)          if (ONIG_IS_NULL(p)) return NULL
00074 #define ONIG_CHECK_NULL_RETURN_VAL(p,val)  if (ONIG_IS_NULL(p)) return (val)
00075 
00076 #define enclen(enc,p,e) ((enc->max_enc_len == enc->min_enc_len) ? enc->min_enc_len : ONIGENC_MBC_ENC_LEN(enc,p,e))
00077 
00078 /* character types bit flag */
00079 #define BIT_CTYPE_NEWLINE  (1<< ONIGENC_CTYPE_NEWLINE)
00080 #define BIT_CTYPE_ALPHA    (1<< ONIGENC_CTYPE_ALPHA)
00081 #define BIT_CTYPE_BLANK    (1<< ONIGENC_CTYPE_BLANK)
00082 #define BIT_CTYPE_CNTRL    (1<< ONIGENC_CTYPE_CNTRL)
00083 #define BIT_CTYPE_DIGIT    (1<< ONIGENC_CTYPE_DIGIT)
00084 #define BIT_CTYPE_GRAPH    (1<< ONIGENC_CTYPE_GRAPH)
00085 #define BIT_CTYPE_LOWER    (1<< ONIGENC_CTYPE_LOWER)
00086 #define BIT_CTYPE_PRINT    (1<< ONIGENC_CTYPE_PRINT)
00087 #define BIT_CTYPE_PUNCT    (1<< ONIGENC_CTYPE_PUNCT)
00088 #define BIT_CTYPE_SPACE    (1<< ONIGENC_CTYPE_SPACE)
00089 #define BIT_CTYPE_UPPER    (1<< ONIGENC_CTYPE_UPPER)
00090 #define BIT_CTYPE_XDIGIT   (1<< ONIGENC_CTYPE_XDIGIT)
00091 #define BIT_CTYPE_WORD     (1<< ONIGENC_CTYPE_WORD)
00092 #define BIT_CTYPE_ALNUM    (1<< ONIGENC_CTYPE_ALNUM)
00093 #define BIT_CTYPE_ASCII    (1<< ONIGENC_CTYPE_ASCII)
00094 
00095 #define CTYPE_TO_BIT(ctype)  (1<<(ctype))
00096 #define CTYPE_IS_WORD_GRAPH_PRINT(ctype) \
00097   ((ctype) == ONIGENC_CTYPE_WORD || (ctype) == ONIGENC_CTYPE_GRAPH ||\
00098    (ctype) == ONIGENC_CTYPE_PRINT)
00099 
00100 
00101 typedef struct {
00102   const UChar *name;
00103   int       ctype;
00104   short int len;
00105 } PosixBracketEntryType;
00106 
00107 #define PosixBracketEntryInit(name, ctype) {(const UChar *)name, ctype, (short int)(sizeof(name) - 1)}
00108 
00109 #define USE_CRNL_AS_LINE_TERMINATOR
00110 #define USE_UNICODE_PROPERTIES
00111 /* #define USE_UNICODE_CASE_FOLD_TURKISH_AZERI */
00112 /* #define USE_UNICODE_ALL_LINE_TERMINATORS */  /* see Unicode.org UTS #18 */
00113 
00114 
00115 #define ONIG_ENCODING_INIT_DEFAULT           ONIG_ENCODING_ASCII
00116 
00117 /* for encoding system implementation (internal) */
00118 ONIG_EXTERN int onigenc_ascii_apply_all_case_fold P_((OnigCaseFoldType flag, OnigApplyAllCaseFoldFunc f, void* arg, OnigEncoding enc));
00119 ONIG_EXTERN int onigenc_ascii_get_case_fold_codes_by_str P_((OnigCaseFoldType flag, const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[], OnigEncoding enc));
00120 ONIG_EXTERN int onigenc_apply_all_case_fold_with_map P_((int map_size, const OnigPairCaseFoldCodes map[], int ess_tsett_flag, OnigCaseFoldType flag, OnigApplyAllCaseFoldFunc f, void* arg));
00121 ONIG_EXTERN int onigenc_get_case_fold_codes_by_str_with_map P_((int map_size, const OnigPairCaseFoldCodes map[], int ess_tsett_flag, OnigCaseFoldType flag, const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[]));
00122 ONIG_EXTERN int onigenc_not_support_get_ctype_code_range P_((OnigCtype ctype, OnigCodePoint* sb_out, const OnigCodePoint* ranges[], OnigEncoding enc));
00123 ONIG_EXTERN int onigenc_is_mbc_newline_0x0a P_((const UChar* p, const UChar* end, OnigEncoding enc));
00124 
00125 
00126 /* methods for single byte encoding */
00127 ONIG_EXTERN int onigenc_ascii_mbc_case_fold P_((OnigCaseFoldType flag, const UChar** p, const UChar* end, UChar* lower, OnigEncoding enc));
00128 ONIG_EXTERN int onigenc_single_byte_mbc_enc_len P_((const UChar* p, const UChar* e, OnigEncoding enc));
00129 ONIG_EXTERN OnigCodePoint onigenc_single_byte_mbc_to_code P_((const UChar* p, const UChar* end, OnigEncoding enc));
00130 ONIG_EXTERN int onigenc_single_byte_code_to_mbclen P_((OnigCodePoint code, OnigEncoding enc));
00131 ONIG_EXTERN int onigenc_single_byte_code_to_mbc P_((OnigCodePoint code, UChar *buf, OnigEncoding enc));
00132 ONIG_EXTERN UChar* onigenc_single_byte_left_adjust_char_head P_((const UChar* start, const UChar* s, const OnigUChar* end, OnigEncoding enc));
00133 ONIG_EXTERN int onigenc_always_true_is_allowed_reverse_match P_((const UChar* s, const UChar* end, OnigEncoding enc));
00134 ONIG_EXTERN int onigenc_always_false_is_allowed_reverse_match P_((const UChar* s, const UChar* end, OnigEncoding enc));
00135 ONIG_EXTERN int onigenc_ascii_is_code_ctype P_((OnigCodePoint code, unsigned int ctype, OnigEncoding enc));
00136 
00137 /* methods for multi byte encoding */
00138 ONIG_EXTERN OnigCodePoint onigenc_mbn_mbc_to_code P_((OnigEncoding enc, const UChar* p, const UChar* end));
00139 ONIG_EXTERN int onigenc_mbn_mbc_case_fold P_((OnigEncoding enc, OnigCaseFoldType flag, const UChar** p, const UChar* end, UChar* lower));
00140 ONIG_EXTERN int onigenc_mb2_code_to_mbclen P_((OnigCodePoint code, OnigEncoding enc));
00141 ONIG_EXTERN int onigenc_mb2_code_to_mbc P_((OnigEncoding enc, OnigCodePoint code, UChar *buf));
00142 ONIG_EXTERN int onigenc_minimum_property_name_to_ctype P_((OnigEncoding enc, UChar* p, UChar* end));
00143 ONIG_EXTERN int onigenc_unicode_property_name_to_ctype P_((OnigEncoding enc, UChar* p, UChar* end));
00144 ONIG_EXTERN int onigenc_mb2_is_code_ctype P_((OnigEncoding enc, OnigCodePoint code, unsigned int ctype));
00145 ONIG_EXTERN int onigenc_mb4_code_to_mbclen P_((OnigCodePoint code, OnigEncoding enc));
00146 ONIG_EXTERN int onigenc_mb4_code_to_mbc P_((OnigEncoding enc, OnigCodePoint code, UChar *buf));
00147 ONIG_EXTERN int onigenc_mb4_is_code_ctype P_((OnigEncoding enc, OnigCodePoint code, unsigned int ctype));
00148 
00149 
00150 /* in enc/unicode.c */
00151 ONIG_EXTERN int onigenc_unicode_is_code_ctype P_((OnigCodePoint code, unsigned int ctype, OnigEncoding enc));
00152 ONIG_EXTERN int onigenc_utf16_32_get_ctype_code_range P_((OnigCtype ctype, OnigCodePoint *sb_out, const OnigCodePoint* ranges[], OnigEncoding enc));
00153 ONIG_EXTERN int onigenc_unicode_ctype_code_range P_((int ctype, const OnigCodePoint* ranges[]));
00154 ONIG_EXTERN int onigenc_unicode_get_case_fold_codes_by_str P_((OnigEncoding enc, OnigCaseFoldType flag, const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[]));
00155 ONIG_EXTERN int onigenc_unicode_mbc_case_fold P_((OnigEncoding enc, OnigCaseFoldType flag, const UChar** pp, const UChar* end, UChar* fold));
00156 ONIG_EXTERN int onigenc_unicode_apply_all_case_fold P_((OnigCaseFoldType flag, OnigApplyAllCaseFoldFunc f, void* arg, OnigEncoding enc));
00157 
00158 
00159 #define UTF16_IS_SURROGATE_FIRST(c)    (((c) & 0xfc) == 0xd8)
00160 #define UTF16_IS_SURROGATE_SECOND(c)   (((c) & 0xfc) == 0xdc)
00161 
00162 #define ONIGENC_ISO_8859_1_TO_LOWER_CASE(c) \
00163   OnigEncISO_8859_1_ToLowerCaseTable[c]
00164 #define ONIGENC_ISO_8859_1_TO_UPPER_CASE(c) \
00165   OnigEncISO_8859_1_ToUpperCaseTable[c]
00166 
00167 ONIG_EXTERN const UChar OnigEncISO_8859_1_ToLowerCaseTable[];
00168 ONIG_EXTERN const UChar OnigEncISO_8859_1_ToUpperCaseTable[];
00169 
00170 ONIG_EXTERN int
00171 onigenc_with_ascii_strncmp P_((OnigEncoding enc, const UChar* p, const UChar* end, const UChar* sascii /* ascii */, int n));
00172 ONIG_EXTERN int
00173 onigenc_with_ascii_strnicmp P_((OnigEncoding enc, const UChar* p, const UChar* end, const UChar* sascii /* ascii */, int n));
00174 ONIG_EXTERN UChar*
00175 onigenc_step P_((OnigEncoding enc, const UChar* p, const UChar* end, int n));
00176 
00177 /* defined in regexec.c, but used in enc/xxx.c */
00178 extern int  onig_is_in_code_range P_((const UChar* p, OnigCodePoint code));
00179 
00180 ONIG_EXTERN OnigEncoding  OnigEncDefaultCharEncoding;
00181 ONIG_EXTERN const UChar  OnigEncAsciiToLowerCaseTable[];
00182 ONIG_EXTERN const UChar  OnigEncAsciiToUpperCaseTable[];
00183 ONIG_EXTERN const unsigned short OnigEncAsciiCtypeTable[];
00184 
00185 #define ONIGENC_IS_ASCII_CODE(code)  ((code) < 0x80)
00186 #define ONIGENC_ASCII_CODE_TO_LOWER_CASE(c) OnigEncAsciiToLowerCaseTable[c]
00187 #define ONIGENC_ASCII_CODE_TO_UPPER_CASE(c) OnigEncAsciiToUpperCaseTable[c]
00188 #define ONIGENC_IS_ASCII_CODE_CTYPE(code,ctype) \
00189   ((OnigEncAsciiCtypeTable[code] & CTYPE_TO_BIT(ctype)) != 0)
00190 #define ONIGENC_IS_ASCII_CODE_CASE_AMBIG(code) \
00191  (ONIGENC_IS_ASCII_CODE_CTYPE(code, ONIGENC_CTYPE_UPPER) ||\
00192   ONIGENC_IS_ASCII_CODE_CTYPE(code, ONIGENC_CTYPE_LOWER))
00193 
00194 /* Check if the code is in the range. (from <= code && code <= to) */
00195 #define ONIGENC_IS_IN_RANGE(code, from, to) \
00196   ((OnigCodePoint )((code) - (from)) <= (OnigCodePoint )((to) - (from)))
00197 
00198 
00199 #ifdef ONIG_ENC_REGISTER
00200 extern int ONIG_ENC_REGISTER(const char *, OnigEncodingType*);
00201 #define OnigEncodingName(n) encoding_##n
00202 #define OnigEncodingDeclare(n) static OnigEncodingType OnigEncodingName(n)
00203 #define OnigEncodingDefine(f,n)                      \
00204     OnigEncodingDeclare(n);                          \
00205     void Init_##f(void) {                            \
00206         ONIG_ENC_REGISTER(OnigEncodingName(n).name,  \
00207                           &OnigEncodingName(n));     \
00208     }                                                \
00209     OnigEncodingDeclare(n)
00210 #else
00211 #define OnigEncodingName(n) OnigEncoding##n
00212 #define OnigEncodingDeclare(n) OnigEncodingType OnigEncodingName(n)
00213 #define OnigEncodingDefine(f,n) OnigEncodingDeclare(n)
00214 #endif
00215 
00216 /* macros for define replica encoding and encoding alias */
00217 #define ENC_REPLICATE(name, orig)
00218 #define ENC_ALIAS(name, orig)
00219 #define ENC_DUMMY(name)
00220 
00221 RUBY_SYMBOL_EXPORT_END
00222 
00223 #endif /* ONIGURUMA_REGENC_H */
00224