/[pcre]/code/trunk/pcre_internal.h
ViewVC logotype

Diff of /code/trunk/pcre_internal.h

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 341 by ph10, Sat Apr 19 16:41:04 2008 UTC revision 364 by ph10, Fri Jul 11 14:53:41 2008 UTC
# Line 132  PCRE_EXP_DATA_DEFN only if they are not Line 132  PCRE_EXP_DATA_DEFN only if they are not
132  #  endif  #  endif
133  #endif  #endif
134    
135    /* When compiling with the MSVC compiler, it is sometimes necessary to include
136    a "calling convention" before exported function names. (This is secondhand
137    information; I know nothing about MSVC myself). For example, something like
138    
139      void __cdecl function(....)
140    
141    might be needed. In order so make this easy, all the exported functions have
142    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
143    set, we ensure here that it has no effect. */
144    
145    #ifndef PCRE_CALL_CONVENTION
146    #define PCRE_CALL_CONVENTION
147    #endif
148    
149  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
150  cannot determine these outside the compilation (e.g. by running a program as  cannot determine these outside the compilation (e.g. by running a program as
151  part of "configure") because PCRE is often cross-compiled for use on other  part of "configure") because PCRE is often cross-compiled for use on other
# Line 140  preprocessor time in standard C environm Line 154  preprocessor time in standard C environm
154    
155  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
156    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
157      typedef short pcre_int16;
158  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
159    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
160      typedef int pcre_int16;
161  #else  #else
162    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
163  #endif  #endif
164    
165  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
166    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
167      typedef int pcre_int32;
168  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
169    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
170      typedef long int pcre_int32;
171  #else  #else
172    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
173  #endif  #endif
# Line 363  never be called in byte mode. To make su Line 381  never be called in byte mode. To make su
381  support is omitted, we don't even define it. */  support is omitted, we don't even define it. */
382    
383  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
 #define NEXTCHAR(p) p++;  
384  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
385  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
386  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
# Line 373  support is omitted, we don't even define Line 390  support is omitted, we don't even define
390    
391  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
392    
 /* Advance a character pointer one byte in non-UTF-8 mode and by one character  
 in UTF-8 mode. */  
   
 #define NEXTCHAR(p) \  
   p++; \  
   if (utf8) { while((*p & 0xc0) == 0x80) p++; }  
   
393  /* Get the next UTF-8 character, not advancing the pointer. This is called when  /* Get the next UTF-8 character, not advancing the pointer. This is called when
394  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
395    
# Line 605  contain UTF-8 characters with values gre Line 615  contain UTF-8 characters with values gre
615  value such as \n. They must have non-zero values, as check_escape() returns  value such as \n. They must have non-zero values, as check_escape() returns
616  their negation. Also, they must appear in the same order as in the opcode  their negation. Also, they must appear in the same order as in the opcode
617  definitions below, up to ESC_z. There's a dummy for OP_ANY because it  definitions below, up to ESC_z. There's a dummy for OP_ANY because it
618  corresponds to "." rather than an escape sequence, and another for OP_ALLANY  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
619  (which is used for [^] in JavaScript compatibility mode).  (which is used for [^] in JavaScript compatibility mode).
620    
621  The final escape must be ESC_REF as subsequent values are used for  The final escape must be ESC_REF as subsequent values are used for
# Line 616  put in between that don't consume a char Line 626  put in between that don't consume a char
626  */  */
627    
628  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
629         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
630         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
631         ESC_REF };         ESC_REF };
632    
633    
# Line 780  enum { Line 790  enum {
790    
791    OP_FAIL,           /* 109 */    OP_FAIL,           /* 109 */
792    OP_ACCEPT,         /* 110 */    OP_ACCEPT,         /* 110 */
793    
794    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
795    
796    OP_SKIPZERO        /* 111 */    OP_SKIPZERO        /* 111 */
797  };  };
798    
# Line 809  for debugging. The macro is referenced o Line 819  for debugging. The macro is referenced o
819    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
820    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
821    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
822    "Skip zero"    "Skip zero"
823    
824    
825  /* This macro defines the length of fixed length operations in the compiled  /* This macro defines the length of fixed length operations in the compiled
# Line 1015  typedef struct match_data { Line 1025  typedef struct match_data {
1025    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1026    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1027    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1028    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1029    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1030    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1031    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
# Line 1126  extern BOOL         _pcre_is_newline(con Line 1136  extern BOOL         _pcre_is_newline(con
1136  extern int          _pcre_ord2utf8(int, uschar *);  extern int          _pcre_ord2utf8(int, uschar *);
1137  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1138                        const pcre_study_data *, pcre_study_data *);                        const pcre_study_data *, pcre_study_data *);
 extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  
 extern unsigned int _pcre_ucp_othercase(const unsigned int);  
1139  extern int          _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_valid_utf8(const uschar *, int);
1140  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1141                        int *, BOOL);                        int *, BOOL);
1142  extern BOOL         _pcre_xclass(int, const uschar *);  extern BOOL         _pcre_xclass(int, const uschar *);
1143    
1144    
1145    /* Unicode character database (UCD) */
1146    
1147    typedef struct {
1148      uschar script;
1149      uschar chartype;
1150      pcre_int32 other_case;
1151    } ucd_record;
1152    
1153    extern const ucd_record  _pcre_ucd_records[];
1154    extern const uschar      _pcre_ucd_stage1[];
1155    extern const pcre_uint16 _pcre_ucd_stage2[];
1156    extern const int         _pcre_ucp_gentype[];
1157    
1158    
1159    /* UCD access macros */
1160    
1161    #define UCD_BLOCK_SIZE 128
1162    #define GET_UCD(ch) (_pcre_ucd_records + \
1163            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1164            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1165    
1166    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1167    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1168    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1169    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1170    
1171  #endif  #endif
1172    
1173  /* End of pcre_internal.h */  /* End of pcre_internal.h */

Legend:
Removed from v.341  
changed lines
  Added in v.364

  ViewVC Help
Powered by ViewVC 1.1.5