--- embedaddon/pcre/pcre_internal.h 2012/02/21 23:05:51 1.1 +++ embedaddon/pcre/pcre_internal.h 2013/07/22 08:25:55 1.1.1.4 @@ -7,7 +7,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2011 University of Cambridge + Copyright (c) 1997-2013 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -40,7 +40,8 @@ POSSIBILITY OF SUCH DAMAGE. /* This header contains definitions that are shared between the different modules, but which are not relevant to the exported API. This includes some -functions whose names all begin with "_pcre_". */ +functions whose names all begin with "_pcre_", "_pcre16_" or "_pcre32_" +depending on the PRIV macro. */ #ifndef PCRE_INTERNAL_H #define PCRE_INTERNAL_H @@ -51,20 +52,40 @@ functions whose names all begin with "_pcre_". */ #define PCRE_DEBUG #endif -/* We do not support both EBCDIC and UTF-8 at the same time. The "configure" -script prevents both being selected, but not everybody uses "configure". */ +/* PCRE is compiled as an 8 bit library if it is not requested otherwise. */ -#if defined EBCDIC && defined SUPPORT_UTF8 -#error The use of both EBCDIC and SUPPORT_UTF8 is not supported. +#if !defined COMPILE_PCRE16 && !defined COMPILE_PCRE32 +#define COMPILE_PCRE8 #endif -/* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The +/* If SUPPORT_UCP is defined, SUPPORT_UTF must also be defined. The "configure" script ensures this, but not everybody uses "configure". */ -#if defined SUPPORT_UCP && !defined SUPPORT_UTF8 +#if defined SUPPORT_UCP && !(defined SUPPORT_UTF) +#define SUPPORT_UTF 1 +#endif + +/* We define SUPPORT_UTF if SUPPORT_UTF8 is enabled for compatibility +reasons with existing code. */ + +#if defined SUPPORT_UTF8 && !(defined SUPPORT_UTF) +#define SUPPORT_UTF 1 +#endif + +/* Fixme: SUPPORT_UTF8 should be eventually disappear from the code. +Until then we define it if SUPPORT_UTF is defined. */ + +#if defined SUPPORT_UTF && !(defined SUPPORT_UTF8) #define SUPPORT_UTF8 1 #endif +/* We do not support both EBCDIC and UTF-8/16/32 at the same time. The "configure" +script prevents both being selected, but not everybody uses "configure". */ + +#if defined EBCDIC && defined SUPPORT_UTF +#error The use of both EBCDIC and SUPPORT_UTF is not supported. +#endif + /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef inline, and there are *still* stupid compilers about that don't like indented pre-processor statements, or at least there were when I first wrote this. After @@ -91,6 +112,12 @@ setjmp and stdarg are used is when NO_RECURSE is set. #include #include +/* Valgrind (memcheck) support */ + +#ifdef SUPPORT_VALGRIND +#include +#endif + /* When compiling a DLL for Windows, the exported symbols have to be declared using some MS magic. I found some useful information on this web page: http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the @@ -158,30 +185,40 @@ set, we ensure here that it has no effect. */ #define PCRE_CALL_CONVENTION #endif -/* We need to have types that specify unsigned 16-bit and 32-bit integers. We +/* We need to have types that specify unsigned 8, 16 and 32-bit integers. We cannot determine these outside the compilation (e.g. by running a program as part of "configure") because PCRE is often cross-compiled for use on other systems. Instead we make use of the maximum sizes that are available at preprocessor time in standard C environments. */ +typedef unsigned char pcre_uint8; + #if USHRT_MAX == 65535 - typedef unsigned short pcre_uint16; - typedef short pcre_int16; +typedef unsigned short pcre_uint16; +typedef short pcre_int16; +#define PCRE_UINT16_MAX USHRT_MAX +#define PCRE_INT16_MAX SHRT_MAX #elif UINT_MAX == 65535 - typedef unsigned int pcre_uint16; - typedef int pcre_int16; +typedef unsigned int pcre_uint16; +typedef int pcre_int16; +#define PCRE_UINT16_MAX UINT_MAX +#define PCRE_INT16_MAX INT_MAX #else - #error Cannot determine a type for 16-bit unsigned integers +#error Cannot determine a type for 16-bit integers #endif -#if UINT_MAX == 4294967295 - typedef unsigned int pcre_uint32; - typedef int pcre_int32; -#elif ULONG_MAX == 4294967295 - typedef unsigned long int pcre_uint32; - typedef long int pcre_int32; +#if UINT_MAX == 4294967295U +typedef unsigned int pcre_uint32; +typedef int pcre_int32; +#define PCRE_UINT32_MAX UINT_MAX +#define PCRE_INT32_MAX INT_MAX +#elif ULONG_MAX == 4294967295UL +typedef unsigned long int pcre_uint32; +typedef long int pcre_int32; +#define PCRE_UINT32_MAX ULONG_MAX +#define PCRE_INT32_MAX LONG_MAX #else - #error Cannot determine a type for 32-bit unsigned integers +#error Cannot determine a type for 32-bit integers #endif /* When checking for integer overflow in pcre_compile(), we need to handle @@ -192,9 +229,9 @@ stdint.h is available, include it; it may define INT64 have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set by "configure". */ -#if HAVE_STDINT_H +#if defined HAVE_STDINT_H #include -#elif HAVE_INTTYPES_H +#elif defined HAVE_INTTYPES_H #include #endif @@ -206,13 +243,54 @@ by "configure". */ /* All character handling must be done as unsigned characters. Otherwise there are problems with top-bit-set characters and functions such as isspace(). -However, we leave the interface to the outside world as char *, because that -should make things easier for callers. We define a short type for unsigned char -to save lots of typing. I tried "uchar", but it causes problems on Digital -Unix, where it is defined in sys/types, so use "uschar" instead. */ +However, we leave the interface to the outside world as char * or short *, +because that should make things easier for callers. This character type is +called pcre_uchar. -typedef unsigned char uschar; +The IN_UCHARS macro multiply its argument with the byte size of the current +pcre_uchar type. Useful for memcpy and such operations, whose require the +byte size of their input/output buffers. +The MAX_255 macro checks whether its pcre_uchar input is less than 256. + +The TABLE_GET macro is designed for accessing elements of tables whose contain +exactly 256 items. When the character is able to contain more than 256 +items, some check is needed before accessing these tables. +*/ + +#if defined COMPILE_PCRE8 + +typedef unsigned char pcre_uchar; +#define IN_UCHARS(x) (x) +#define MAX_255(c) 1 +#define TABLE_GET(c, table, default) ((table)[c]) + +#elif defined COMPILE_PCRE16 + +#if USHRT_MAX != 65535 +/* This is a warning message. Change PCRE_UCHAR16 to a 16 bit data type in +pcre.h(.in) and disable (comment out) this message. */ +#error Warning: PCRE_UCHAR16 is not a 16 bit data type. +#endif + +typedef pcre_uint16 pcre_uchar; +#define UCHAR_SHIFT (1) +#define IN_UCHARS(x) ((x) << UCHAR_SHIFT) +#define MAX_255(c) ((c) <= 255u) +#define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default)) + +#elif defined COMPILE_PCRE32 + +typedef pcre_uint32 pcre_uchar; +#define UCHAR_SHIFT (2) +#define IN_UCHARS(x) ((x) << UCHAR_SHIFT) +#define MAX_255(c) ((c) <= 255u) +#define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default)) + +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE[8|16|32] */ + /* This is an unsigned int value that no character can ever have. UTF-8 characters only go up to 0x7fffffff (though Unicode doesn't go beyond 0x0010ffff). */ @@ -234,12 +312,12 @@ start/end of string field names are. */ #define IS_NEWLINE(p) \ ((NLBLOCK->nltype != NLTYPE_FIXED)? \ ((p) < NLBLOCK->PSEND && \ - _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\ - utf8)) \ + PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, \ + &(NLBLOCK->nllen), utf)) \ : \ ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \ - (p)[0] == NLBLOCK->nl[0] && \ - (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \ + RAWUCHARTEST(p) == NLBLOCK->nl[0] && \ + (NLBLOCK->nllen == 1 || RAWUCHARTEST(p+1) == NLBLOCK->nl[1]) \ ) \ ) @@ -248,12 +326,12 @@ start/end of string field names are. */ #define WAS_NEWLINE(p) \ ((NLBLOCK->nltype != NLTYPE_FIXED)? \ ((p) > NLBLOCK->PSSTART && \ - _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \ - &(NLBLOCK->nllen), utf8)) \ + PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \ + &(NLBLOCK->nllen), utf)) \ : \ ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \ - (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \ - (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \ + RAWUCHARTEST(p - NLBLOCK->nllen) == NLBLOCK->nl[0] && \ + (NLBLOCK->nllen == 1 || RAWUCHARTEST(p - NLBLOCK->nllen + 1) == NLBLOCK->nl[1]) \ ) \ ) @@ -267,21 +345,22 @@ used for the external interface and appears in pcre.h, must begin with PCRE_. */ #ifdef CUSTOM_SUBJECT_PTR -#define PCRE_SPTR CUSTOM_SUBJECT_PTR -#define USPTR CUSTOM_SUBJECT_PTR +#define PCRE_PUCHAR CUSTOM_SUBJECT_PTR #else -#define PCRE_SPTR const char * -#define USPTR const unsigned char * +#define PCRE_PUCHAR const pcre_uchar * #endif - - /* Include the public PCRE header and the definitions of UCP character property values. */ #include "pcre.h" #include "ucp.h" +#ifdef COMPILE_PCRE32 +/* Assert that the public PCRE_UCHAR32 is a 32-bit type */ +typedef int __assert_pcre_uchar32_size[sizeof(PCRE_UCHAR32) == 4 ? 1 : -1]; +#endif + /* When compiling for use with the Virtual Pascal compiler, these functions need to have their names changed. PCRE must be compiled with the -DVPCOMPAT option on the command line. */ @@ -343,6 +422,8 @@ The macros are controlled by the value of LINK_SIZE. T the config.h file, but can be overridden by using -D on the command line. This is automated on Unix systems via the "configure" command. */ +#if defined COMPILE_PCRE8 + #if LINK_SIZE == 2 #define PUT(a,n,d) \ @@ -379,14 +460,69 @@ is automated on Unix systems via the "configure" comma #define GET(a,n) \ (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3]) -#define MAX_PATTERN_SIZE (1 << 30) /* Keep it positive */ +/* Keep it positive */ +#define MAX_PATTERN_SIZE (1 << 30) +#else +#error LINK_SIZE must be either 2, 3, or 4 +#endif +#elif defined COMPILE_PCRE16 + +#if LINK_SIZE == 2 + +/* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */ +#undef LINK_SIZE +#define LINK_SIZE 1 + +#define PUT(a,n,d) \ + (a[n] = (d)) + +#define GET(a,n) \ + (a[n]) + +#define MAX_PATTERN_SIZE (1 << 16) + +#elif LINK_SIZE == 3 || LINK_SIZE == 4 + +/* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */ +#undef LINK_SIZE +#define LINK_SIZE 2 + +#define PUT(a,n,d) \ + (a[n] = (d) >> 16), \ + (a[(n)+1] = (d) & 65535) + +#define GET(a,n) \ + (((a)[n] << 16) | (a)[(n)+1]) + +/* Keep it positive */ +#define MAX_PATTERN_SIZE (1 << 30) + #else #error LINK_SIZE must be either 2, 3, or 4 #endif +#elif defined COMPILE_PCRE32 +/* Only supported LINK_SIZE is 4 */ +/* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */ +#undef LINK_SIZE +#define LINK_SIZE 1 + +#define PUT(a,n,d) \ + (a[n] = (d)) + +#define GET(a,n) \ + (a[n]) + +/* Keep it positive */ +#define MAX_PATTERN_SIZE (1 << 30) + +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE[8|16|32] */ + /* Convenience macro defined in terms of the others */ #define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE @@ -396,39 +532,89 @@ is automated on Unix systems via the "configure" comma offsets changes. There are used for repeat counts and for other things such as capturing parenthesis numbers in back references. */ +#if defined COMPILE_PCRE8 + +#define IMM2_SIZE 2 + #define PUT2(a,n,d) \ a[n] = (d) >> 8; \ a[(n)+1] = (d) & 255 +/* For reasons that I do not understand, the expression in this GET2 macro is +treated by gcc as a signed expression, even when a is declared as unsigned. It +seems that any kind of arithmetic results in a signed value. */ + #define GET2(a,n) \ - (((a)[n] << 8) | (a)[(n)+1]) + (unsigned int)(((a)[n] << 8) | (a)[(n)+1]) -#define PUT2INC(a,n,d) PUT2(a,n,d), a += 2 +#elif defined COMPILE_PCRE16 +#define IMM2_SIZE 1 -/* When UTF-8 encoding is being used, a character is no longer just a single +#define PUT2(a,n,d) \ + a[n] = d + +#define GET2(a,n) \ + a[n] + +#elif defined COMPILE_PCRE32 + +#define IMM2_SIZE 1 + +#define PUT2(a,n,d) \ + a[n] = d + +#define GET2(a,n) \ + a[n] + +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE[8|16|32] */ + +#define PUT2INC(a,n,d) PUT2(a,n,d), a += IMM2_SIZE + +/* The maximum length of a MARK name is currently one data unit; it may be +changed in future to be a fixed number of bytes or to depend on LINK_SIZE. */ + +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 +#define MAX_MARK ((1u << 16) - 1) +#else +#define MAX_MARK ((1u << 8) - 1) +#endif + +/* When UTF encoding is being used, a character is no longer just a single byte. The macros for character handling generate simple sequences when used in -byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is -not used when UTF-8 is not supported, so it is not defined, and BACKCHAR should -never be called in byte mode. To make sure they can never even appear when -UTF-8 support is omitted, we don't even define them. */ +character-mode, and more complicated ones for UTF characters. GETCHARLENTEST +and other macros are not used when UTF is not supported, so they are not +defined. To make sure they can never even appear when UTF support is omitted, +we don't even define them. */ -#ifndef SUPPORT_UTF8 +#ifndef SUPPORT_UTF + +/* #define MAX_VALUE_FOR_SINGLE_CHAR */ +/* #define HAS_EXTRALEN(c) */ +/* #define GET_EXTRALEN(c) */ +/* #define NOT_FIRSTCHAR(c) */ #define GETCHAR(c, eptr) c = *eptr; #define GETCHARTEST(c, eptr) c = *eptr; #define GETCHARINC(c, eptr) c = *eptr++; #define GETCHARINCTEST(c, eptr) c = *eptr++; #define GETCHARLEN(c, eptr, len) c = *eptr; +#define RAWUCHAR(eptr) (*(eptr)) +#define RAWUCHARINC(eptr) (*(eptr)++) +#define RAWUCHARTEST(eptr) (*(eptr)) +#define RAWUCHARINCTEST(eptr) (*(eptr)++) /* #define GETCHARLENTEST(c, eptr, len) */ /* #define BACKCHAR(eptr) */ +/* #define FORWARDCHAR(eptr) */ +/* #define ACROSSCHAR(condition, eptr, action) */ -#else /* SUPPORT_UTF8 */ +#else /* SUPPORT_UTF */ -/* These macros were originally written in the form of loops that used data -from the tables whose names start with _pcre_utf8_table. They were rewritten by -a user so as not to use loops, because in some environments this gives a -significant performance advantage, and it seems never to do any harm. */ +/* Tests whether the code point needs extra characters to decode. */ +#define HASUTF8EXTRALEN(c) ((c) >= 0xc0) + /* Base macro to pick up the remaining bytes of a UTF-8 character, not advancing the pointer. */ @@ -451,20 +637,6 @@ advancing the pointer. */ ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \ } -/* Get the next UTF-8 character, not advancing the pointer. This is called when -we know we are in UTF-8 mode. */ - -#define GETCHAR(c, eptr) \ - c = *eptr; \ - if (c >= 0xc0) GETUTF8(c, eptr); - -/* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the -pointer. */ - -#define GETCHARTEST(c, eptr) \ - c = *eptr; \ - if (utf8 && c >= 0xc0) GETUTF8(c, eptr); - /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing the pointer. */ @@ -499,6 +671,45 @@ the pointer. */ } \ } +#if defined COMPILE_PCRE8 + +/* These macros were originally written in the form of loops that used data +from the tables whose names start with PRIV(utf8_table). They were rewritten by +a user so as not to use loops, because in some environments this gives a +significant performance advantage, and it seems never to do any harm. */ + +/* Tells the biggest code point which can be encoded as a single character. */ + +#define MAX_VALUE_FOR_SINGLE_CHAR 127 + +/* Tests whether the code point needs extra characters to decode. */ + +#define HAS_EXTRALEN(c) ((c) >= 0xc0) + +/* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE. +Otherwise it has an undefined behaviour. */ + +#define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f]) + +/* Returns TRUE, if the given character is not the first character +of a UTF sequence. */ + +#define NOT_FIRSTCHAR(c) (((c) & 0xc0) == 0x80) + +/* Get the next UTF-8 character, not advancing the pointer. This is called when +we know we are in UTF-8 mode. */ + +#define GETCHAR(c, eptr) \ + c = *eptr; \ + if (c >= 0xc0) GETUTF8(c, eptr); + +/* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the +pointer. */ + +#define GETCHARTEST(c, eptr) \ + c = *eptr; \ + if (utf && c >= 0xc0) GETUTF8(c, eptr); + /* Get the next UTF-8 character, advancing the pointer. This is called when we know we are in UTF-8 mode. */ @@ -511,7 +722,7 @@ This is called when we don't know if we are in UTF-8 m #define GETCHARINCTEST(c, eptr) \ c = *eptr++; \ - if (utf8 && c >= 0xc0) GETUTF8INC(c, eptr); + if (utf && c >= 0xc0) GETUTF8INC(c, eptr); /* Base macro to pick up the remaining bytes of a UTF-8 character, not advancing the pointer, incrementing the length. */ @@ -563,39 +774,391 @@ do not know if we are in UTF-8 mode. */ #define GETCHARLENTEST(c, eptr, len) \ c = *eptr; \ - if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len); + if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len); +/* Returns the next uchar, not advancing the pointer. This is called when +we know we are in UTF mode. */ + +#define RAWUCHAR(eptr) \ + (*(eptr)) + +/* Returns the next uchar, advancing the pointer. This is called when +we know we are in UTF mode. */ + +#define RAWUCHARINC(eptr) \ + (*((eptr)++)) + +/* Returns the next uchar, testing for UTF mode, and not advancing the +pointer. */ + +#define RAWUCHARTEST(eptr) \ + (*(eptr)) + +/* Returns the next uchar, testing for UTF mode, advancing the +pointer. */ + +#define RAWUCHARINCTEST(eptr) \ + (*((eptr)++)) + /* If the pointer is not at the start of a character, move it back until it is. This is called only in UTF-8 mode - we don't put a test within the macro because almost all calls are already within a block of UTF-8 only code. */ #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr-- -#endif /* SUPPORT_UTF8 */ +/* Same as above, just in the other direction. */ +#define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++ +/* Same as above, but it allows a fully customizable form. */ +#define ACROSSCHAR(condition, eptr, action) \ + while((condition) && ((eptr) & 0xc0) == 0x80) action -/* In case there is no definition of offsetof() provided - though any proper -Standard C system should have one. */ +#elif defined COMPILE_PCRE16 -#ifndef offsetof -#define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field)) +/* Tells the biggest code point which can be encoded as a single character. */ + +#define MAX_VALUE_FOR_SINGLE_CHAR 65535 + +/* Tests whether the code point needs extra characters to decode. */ + +#define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800) + +/* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE. +Otherwise it has an undefined behaviour. */ + +#define GET_EXTRALEN(c) 1 + +/* Returns TRUE, if the given character is not the first character +of a UTF sequence. */ + +#define NOT_FIRSTCHAR(c) (((c) & 0xfc00) == 0xdc00) + +/* Base macro to pick up the low surrogate of a UTF-16 character, not +advancing the pointer. */ + +#define GETUTF16(c, eptr) \ + { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; } + +/* Get the next UTF-16 character, not advancing the pointer. This is called when +we know we are in UTF-16 mode. */ + +#define GETCHAR(c, eptr) \ + c = *eptr; \ + if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr); + +/* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the +pointer. */ + +#define GETCHARTEST(c, eptr) \ + c = *eptr; \ + if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr); + +/* Base macro to pick up the low surrogate of a UTF-16 character, advancing +the pointer. */ + +#define GETUTF16INC(c, eptr) \ + { c = (((c & 0x3ff) << 10) | (*eptr++ & 0x3ff)) + 0x10000; } + +/* Get the next UTF-16 character, advancing the pointer. This is called when we +know we are in UTF-16 mode. */ + +#define GETCHARINC(c, eptr) \ + c = *eptr++; \ + if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr); + +/* Get the next character, testing for UTF-16 mode, and advancing the pointer. +This is called when we don't know if we are in UTF-16 mode. */ + +#define GETCHARINCTEST(c, eptr) \ + c = *eptr++; \ + if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr); + +/* Base macro to pick up the low surrogate of a UTF-16 character, not +advancing the pointer, incrementing the length. */ + +#define GETUTF16LEN(c, eptr, len) \ + { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; } + +/* Get the next UTF-16 character, not advancing the pointer, incrementing +length if there is a low surrogate. This is called when we know we are in +UTF-16 mode. */ + +#define GETCHARLEN(c, eptr, len) \ + c = *eptr; \ + if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len); + +/* Get the next UTF-816character, testing for UTF-16 mode, not advancing the +pointer, incrementing length if there is a low surrogate. This is called when +we do not know if we are in UTF-16 mode. */ + +#define GETCHARLENTEST(c, eptr, len) \ + c = *eptr; \ + if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len); + +/* Returns the next uchar, not advancing the pointer. This is called when +we know we are in UTF mode. */ + +#define RAWUCHAR(eptr) \ + (*(eptr)) + +/* Returns the next uchar, advancing the pointer. This is called when +we know we are in UTF mode. */ + +#define RAWUCHARINC(eptr) \ + (*((eptr)++)) + +/* Returns the next uchar, testing for UTF mode, and not advancing the +pointer. */ + +#define RAWUCHARTEST(eptr) \ + (*(eptr)) + +/* Returns the next uchar, testing for UTF mode, advancing the +pointer. */ + +#define RAWUCHARINCTEST(eptr) \ + (*((eptr)++)) + +/* If the pointer is not at the start of a character, move it back until +it is. This is called only in UTF-16 mode - we don't put a test within the +macro because almost all calls are already within a block of UTF-16 only +code. */ + +#define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr-- + +/* Same as above, just in the other direction. */ +#define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++ + +/* Same as above, but it allows a fully customizable form. */ +#define ACROSSCHAR(condition, eptr, action) \ + if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action + +#elif defined COMPILE_PCRE32 + +/* These are trivial for the 32-bit library, since all UTF-32 characters fit +into one pcre_uchar unit. */ +#define MAX_VALUE_FOR_SINGLE_CHAR (0x10ffffu) +#define HAS_EXTRALEN(c) (0) +#define GET_EXTRALEN(c) (0) +#define NOT_FIRSTCHAR(c) (0) + +/* Get the next UTF-32 character, not advancing the pointer. This is called when +we know we are in UTF-32 mode. */ + +#define GETCHAR(c, eptr) \ + c = *(eptr); + +/* Get the next UTF-32 character, testing for UTF-32 mode, and not advancing the +pointer. */ + +#define GETCHARTEST(c, eptr) \ + c = *(eptr); + +/* Get the next UTF-32 character, advancing the pointer. This is called when we +know we are in UTF-32 mode. */ + +#define GETCHARINC(c, eptr) \ + c = *((eptr)++); + +/* Get the next character, testing for UTF-32 mode, and advancing the pointer. +This is called when we don't know if we are in UTF-32 mode. */ + +#define GETCHARINCTEST(c, eptr) \ + c = *((eptr)++); + +/* Get the next UTF-32 character, not advancing the pointer, not incrementing +length (since all UTF-32 is of length 1). This is called when we know we are in +UTF-32 mode. */ + +#define GETCHARLEN(c, eptr, len) \ + GETCHAR(c, eptr) + +/* Get the next UTF-32character, testing for UTF-32 mode, not advancing the +pointer, not incrementing the length (since all UTF-32 is of length 1). +This is called when we do not know if we are in UTF-32 mode. */ + +#define GETCHARLENTEST(c, eptr, len) \ + GETCHARTEST(c, eptr) + +/* Returns the next uchar, not advancing the pointer. This is called when +we know we are in UTF mode. */ + +#define RAWUCHAR(eptr) \ + (*(eptr)) + +/* Returns the next uchar, advancing the pointer. This is called when +we know we are in UTF mode. */ + +#define RAWUCHARINC(eptr) \ + (*((eptr)++)) + +/* Returns the next uchar, testing for UTF mode, and not advancing the +pointer. */ + +#define RAWUCHARTEST(eptr) \ + (*(eptr)) + +/* Returns the next uchar, testing for UTF mode, advancing the +pointer. */ + +#define RAWUCHARINCTEST(eptr) \ + (*((eptr)++)) + +/* If the pointer is not at the start of a character, move it back until +it is. This is called only in UTF-32 mode - we don't put a test within the +macro because almost all calls are already within a block of UTF-32 only +code. +These are all no-ops since all UTF-32 characters fit into one pcre_uchar. */ + +#define BACKCHAR(eptr) do { } while (0) + +/* Same as above, just in the other direction. */ +#define FORWARDCHAR(eptr) do { } while (0) + +/* Same as above, but it allows a fully customizable form. */ +#define ACROSSCHAR(condition, eptr, action) do { } while (0) + +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE[8|16|32] */ + +#endif /* SUPPORT_UTF */ + +/* Tests for Unicode horizontal and vertical whitespace characters must check a +number of different values. Using a switch statement for this generates the +fastest code (no loop, no memory access), and there are several places in the +interpreter code where this happens. In order to ensure that all the case lists +remain in step, we use macros so that there is only one place where the lists +are defined. + +These values are also required as lists in pcre_compile.c when processing \h, +\H, \v and \V in a character class. The lists are defined in pcre_tables.c, but +macros that define the values are here so that all the definitions are +together. The lists must be in ascending character order, terminated by +NOTACHAR (which is 0xffffffff). + +Any changes should ensure that the various macros are kept in step with each +other. NOTE: The values also appear in pcre_jit_compile.c. */ + +/* ------ ASCII/Unicode environments ------ */ + +#ifndef EBCDIC + +#define HSPACE_LIST \ + CHAR_HT, CHAR_SPACE, 0xa0, \ + 0x1680, 0x180e, 0x2000, 0x2001, 0x2002, 0x2003, 0x2004, 0x2005, \ + 0x2006, 0x2007, 0x2008, 0x2009, 0x200A, 0x202f, 0x205f, 0x3000, \ + NOTACHAR + +#define HSPACE_MULTIBYTE_CASES \ + case 0x1680: /* OGHAM SPACE MARK */ \ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ \ + case 0x2000: /* EN QUAD */ \ + case 0x2001: /* EM QUAD */ \ + case 0x2002: /* EN SPACE */ \ + case 0x2003: /* EM SPACE */ \ + case 0x2004: /* THREE-PER-EM SPACE */ \ + case 0x2005: /* FOUR-PER-EM SPACE */ \ + case 0x2006: /* SIX-PER-EM SPACE */ \ + case 0x2007: /* FIGURE SPACE */ \ + case 0x2008: /* PUNCTUATION SPACE */ \ + case 0x2009: /* THIN SPACE */ \ + case 0x200A: /* HAIR SPACE */ \ + case 0x202f: /* NARROW NO-BREAK SPACE */ \ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ \ + case 0x3000 /* IDEOGRAPHIC SPACE */ + +#define HSPACE_BYTE_CASES \ + case CHAR_HT: \ + case CHAR_SPACE: \ + case 0xa0 /* NBSP */ + +#define HSPACE_CASES \ + HSPACE_BYTE_CASES: \ + HSPACE_MULTIBYTE_CASES + +#define VSPACE_LIST \ + CHAR_LF, CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, 0x2028, 0x2029, NOTACHAR + +#define VSPACE_MULTIBYTE_CASES \ + case 0x2028: /* LINE SEPARATOR */ \ + case 0x2029 /* PARAGRAPH SEPARATOR */ + +#define VSPACE_BYTE_CASES \ + case CHAR_LF: \ + case CHAR_VT: \ + case CHAR_FF: \ + case CHAR_CR: \ + case CHAR_NEL + +#define VSPACE_CASES \ + VSPACE_BYTE_CASES: \ + VSPACE_MULTIBYTE_CASES + +/* ------ EBCDIC environments ------ */ + +#else +#define HSPACE_LIST CHAR_HT, CHAR_SPACE + +#define HSPACE_BYTE_CASES \ + case CHAR_HT: \ + case CHAR_SPACE + +#define HSPACE_CASES HSPACE_BYTE_CASES + +#ifdef EBCDIC_NL25 +#define VSPACE_LIST \ + CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, CHAR_LF, NOTACHAR +#else +#define VSPACE_LIST \ + CHAR_VT, CHAR_FF, CHAR_CR, CHAR_LF, CHAR_NEL, NOTACHAR #endif +#define VSPACE_BYTE_CASES \ + case CHAR_LF: \ + case CHAR_VT: \ + case CHAR_FF: \ + case CHAR_CR: \ + case CHAR_NEL +#define VSPACE_CASES VSPACE_BYTE_CASES +#endif /* EBCDIC */ + +/* ------ End of whitespace macros ------ */ + + + /* Private flags containing information about the compiled regex. They used to -live at the top end of the options word, but that got almost full, so now they -are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as -the restrictions on partial matching have been lifted. It remains for backwards +live at the top end of the options word, but that got almost full, so they were +moved to a 16-bit flags word - which got almost full, so now they are in a +32-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as the +restrictions on partial matching have been lifted. It remains for backwards compatibility. */ -#define PCRE_NOPARTIAL 0x0001 /* can't use partial with this regex */ -#define PCRE_FIRSTSET 0x0002 /* first_byte is set */ -#define PCRE_REQCHSET 0x0004 /* req_byte is set */ -#define PCRE_STARTLINE 0x0008 /* start after \n for multiline */ -#define PCRE_JCHANGED 0x0010 /* j option used in regex */ -#define PCRE_HASCRORLF 0x0020 /* explicit \r or \n in pattern */ -#define PCRE_HASTHEN 0x0040 /* pattern contains (*THEN) */ +#define PCRE_MODE8 0x00000001 /* compiled in 8 bit mode */ +#define PCRE_MODE16 0x00000002 /* compiled in 16 bit mode */ +#define PCRE_MODE32 0x00000004 /* compiled in 32 bit mode */ +#define PCRE_FIRSTSET 0x00000010 /* first_char is set */ +#define PCRE_FCH_CASELESS 0x00000020 /* caseless first char */ +#define PCRE_REQCHSET 0x00000040 /* req_byte is set */ +#define PCRE_RCH_CASELESS 0x00000080 /* caseless requested char */ +#define PCRE_STARTLINE 0x00000100 /* start after \n for multiline */ +#define PCRE_NOPARTIAL 0x00000200 /* can't use partial with this regex */ +#define PCRE_JCHANGED 0x00000400 /* j option used in regex */ +#define PCRE_HASCRORLF 0x00000800 /* explicit \r or \n in pattern */ +#define PCRE_HASTHEN 0x00001000 /* pattern contains (*THEN) */ +#define PCRE_MLSET 0x00002000 /* match limit set by regex */ +#define PCRE_RLSET 0x00004000 /* recursion limit set by regex */ +#if defined COMPILE_PCRE8 +#define PCRE_MODE PCRE_MODE8 +#elif defined COMPILE_PCRE16 +#define PCRE_MODE PCRE_MODE16 +#elif defined COMPILE_PCRE32 +#define PCRE_MODE PCRE_MODE32 +#endif +#define PCRE_MODE_MASK (PCRE_MODE8 | PCRE_MODE16 | PCRE_MODE32) + /* Flags for the "extra" block produced by pcre_study(). */ #define PCRE_STUDY_MAPPED 0x0001 /* a map of starting chars exists */ @@ -612,7 +1175,7 @@ time, run time, or study time, respectively. */ PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \ PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \ PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \ - PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE) + PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE|PCRE_NEVER_UTF) #define PUBLIC_EXEC_OPTIONS \ (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \ @@ -626,29 +1189,27 @@ time, run time, or study time, respectively. */ PCRE_NO_START_OPTIMIZE) #define PUBLIC_STUDY_OPTIONS \ - PCRE_STUDY_JIT_COMPILE + (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \ + PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE|PCRE_STUDY_EXTRA_NEEDED) -/* Magic number to provide a small check against being handed junk. Also used -to detect whether a pattern was compiled on a host of different endianness. */ +#define PUBLIC_JIT_EXEC_OPTIONS \ + (PCRE_NO_UTF8_CHECK|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|\ + PCRE_NOTEMPTY_ATSTART|PCRE_PARTIAL_SOFT|PCRE_PARTIAL_HARD) +/* Magic number to provide a small check against being handed junk. */ + #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */ -/* Negative values for the firstchar and reqchar variables */ +/* This variable is used to detect a loaded regular expression +in different endianness. */ -#define REQ_UNSET (-2) -#define REQ_NONE (-1) +#define REVERSED_MAGIC_NUMBER 0x45524350UL /* 'ERCP' */ /* The maximum remaining length of subject we are prepared to search for a req_byte match. */ #define REQ_BYTE_MAX 1000 -/* Flags added to firstbyte or reqbyte; a "non-literal" item is either a -variable-length repeat, or a anything other than literal characters. */ - -#define REQ_CASELESS 0x0100 /* indicates caselessness */ -#define REQ_VARY 0x0200 /* reqbyte followed non-literal item */ - /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in environments where these macros are defined elsewhere. Unfortunately, there is no way to do the same for the typedef. */ @@ -677,25 +1238,74 @@ for) in a minority area (EBCDIC platforms), this is no application that did need both could compile two versions of the library, using macros to give the functions distinct names. */ -#ifndef SUPPORT_UTF8 +#ifndef SUPPORT_UTF /* UTF-8 support is not enabled; use the platform-dependent character literals -so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */ +so that PCRE works in both ASCII and EBCDIC environments, but only in non-UTF +mode. Newline characters are problematic in EBCDIC. Though it has CR and LF +characters, a common practice has been to use its NL (0x15) character as the +line terminator in C-like processing environments. However, sometimes the LF +(0x25) character is used instead, according to this Unicode document: +http://unicode.org/standard/reports/tr13/tr13-5.html + +PCRE defaults EBCDIC NL to 0x15, but has a build-time option to select 0x25 +instead. Whichever is *not* chosen is defined as NEL. + +In both ASCII and EBCDIC environments, CHAR_NL and CHAR_LF are synonyms for the +same code point. */ + +#ifdef EBCDIC + +#ifndef EBCDIC_NL25 +#define CHAR_NL '\x15' +#define CHAR_NEL '\x25' +#define STR_NL "\x15" +#define STR_NEL "\x25" +#else +#define CHAR_NL '\x25' +#define CHAR_NEL '\x15' +#define STR_NL "\x25" +#define STR_NEL "\x15" +#endif + +#define CHAR_LF CHAR_NL +#define STR_LF STR_NL + +#define CHAR_ESC '\047' +#define CHAR_DEL '\007' +#define STR_ESC "\047" +#define STR_DEL "\007" + +#else /* Not EBCDIC */ + +/* In ASCII/Unicode, linefeed is '\n' and we equate this to NL for +compatibility. NEL is the Unicode newline character; make sure it is +a positive value. */ + +#define CHAR_LF '\n' +#define CHAR_NL CHAR_LF +#define CHAR_NEL ((unsigned char)'\x85') +#define CHAR_ESC '\033' +#define CHAR_DEL '\177' + +#define STR_LF "\n" +#define STR_NL STR_LF +#define STR_NEL "\x85" +#define STR_ESC "\033" +#define STR_DEL "\177" + +#endif /* EBCDIC */ + +/* The remaining definitions work in both environments. */ + +#define CHAR_NULL '\0' #define CHAR_HT '\t' #define CHAR_VT '\v' #define CHAR_FF '\f' #define CHAR_CR '\r' -#define CHAR_NL '\n' #define CHAR_BS '\b' #define CHAR_BEL '\a' -#ifdef EBCDIC -#define CHAR_ESC '\047' -#define CHAR_DEL '\007' -#else -#define CHAR_ESC '\033' -#define CHAR_DEL '\177' -#endif #define CHAR_SPACE ' ' #define CHAR_EXCLAMATION_MARK '!' @@ -797,16 +1407,8 @@ so that PCRE works on both ASCII and EBCDIC platforms, #define STR_VT "\v" #define STR_FF "\f" #define STR_CR "\r" -#define STR_NL "\n" #define STR_BS "\b" #define STR_BEL "\a" -#ifdef EBCDIC -#define STR_ESC "\047" -#define STR_DEL "\007" -#else -#define STR_ESC "\033" -#define STR_DEL "\177" -#endif #define STR_SPACE " " #define STR_EXCLAMATION_MARK "!" @@ -938,10 +1540,15 @@ so that PCRE works on both ASCII and EBCDIC platforms, #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)" #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)" #define STRING_UTF8_RIGHTPAR "UTF8)" +#define STRING_UTF16_RIGHTPAR "UTF16)" +#define STRING_UTF32_RIGHTPAR "UTF32)" +#define STRING_UTF_RIGHTPAR "UTF)" #define STRING_UCP_RIGHTPAR "UCP)" #define STRING_NO_START_OPT_RIGHTPAR "NO_START_OPT)" +#define STRING_LIMIT_MATCH_EQ "LIMIT_MATCH=" +#define STRING_LIMIT_RECURSION_EQ "LIMIT_RECURSION=" -#else /* SUPPORT_UTF8 */ +#else /* SUPPORT_UTF */ /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode @@ -951,12 +1558,15 @@ only. */ #define CHAR_VT '\013' #define CHAR_FF '\014' #define CHAR_CR '\015' -#define CHAR_NL '\012' +#define CHAR_LF '\012' +#define CHAR_NL CHAR_LF +#define CHAR_NEL ((unsigned char)'\x85') #define CHAR_BS '\010' #define CHAR_BEL '\007' #define CHAR_ESC '\033' #define CHAR_DEL '\177' +#define CHAR_NULL '\0' #define CHAR_SPACE '\040' #define CHAR_EXCLAMATION_MARK '\041' #define CHAR_QUOTATION_MARK '\042' @@ -1193,10 +1803,15 @@ only. */ #define STRING_BSR_ANYCRLF_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS #define STRING_BSR_UNICODE_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS #define STRING_UTF8_RIGHTPAR STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS +#define STRING_UTF16_RIGHTPAR STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS +#define STRING_UTF32_RIGHTPAR STR_U STR_T STR_F STR_3 STR_2 STR_RIGHT_PARENTHESIS +#define STRING_UTF_RIGHTPAR STR_U STR_T STR_F STR_RIGHT_PARENTHESIS #define STRING_UCP_RIGHTPAR STR_U STR_C STR_P STR_RIGHT_PARENTHESIS #define STRING_NO_START_OPT_RIGHTPAR STR_N STR_O STR_UNDERSCORE STR_S STR_T STR_A STR_R STR_T STR_UNDERSCORE STR_O STR_P STR_T STR_RIGHT_PARENTHESIS +#define STRING_LIMIT_MATCH_EQ STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_M STR_A STR_T STR_C STR_H STR_EQUALS_SIGN +#define STRING_LIMIT_RECURSION_EQ STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_R STR_E STR_C STR_U STR_R STR_S STR_I STR_O STR_N STR_EQUALS_SIGN -#endif /* SUPPORT_UTF8 */ +#endif /* SUPPORT_UTF */ /* Escape items that are just an encoding of a particular data value. */ @@ -1209,7 +1824,7 @@ only. */ #endif #ifndef ESC_n -#define ESC_n CHAR_NL +#define ESC_n CHAR_LF #endif #ifndef ESC_r @@ -1234,9 +1849,11 @@ only. */ #define PT_SPACE 6 /* Perl space - Z plus 9,10,12,13 */ #define PT_PXSPACE 7 /* POSIX space - Z plus 9,10,11,12,13 */ #define PT_WORD 8 /* Word - L plus N plus underscore */ +#define PT_CLIST 9 /* Pseudo-property: match character list */ +#define PT_UCNC 10 /* Universal Character nameable character */ /* Flag bits and data types for the extended class (OP_XCLASS) for classes that -contain UTF-8 characters with values greater than 255. */ +contain characters with values greater than 255. */ #define XCL_NOT 0x01 /* Flag: this is a negative class */ #define XCL_MAP 0x02 /* Flag: a 32-byte map is present */ @@ -1249,19 +1866,19 @@ contain UTF-8 characters with values greater than 255. /* These are escaped items that aren't just an encoding of a particular data value such as \n. They must have non-zero values, as check_escape() returns -their negation. Also, they must appear in the same order as in the opcode +0 for a data character. Also, they must appear in the same order as in the opcode definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it corresponds to "." in DOTALL mode rather than an escape sequence. It is also -used for [^] in JavaScript compatibility mode, and for \C in non-utf8 mode. In +used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In non-DOTALL mode, "." behaves like \N. The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc. -when PCRE_UCP is set, when replacement of \d etc by \p sequences is required. +when PCRE_UCP is set and replacement of \d etc by \p sequences is required. They must be contiguous, and remain in order so that the replacements can be looked up from a table. -The final escape must be ESC_REF as subsequent values are used for -backreferences (\1, \2, \3, etc). There are two tests in the code for an escape +Negative numbers are used to encode a backreference (\1, \2, \3, etc.) in +check_escape(). There are two tests in the code for an escape greater than ESC_b and less than ESC_Z to detect the types that may be repeated. These are the types that consume characters. If any new escapes are put in between that don't consume a character, that code will have to change. @@ -1271,8 +1888,7 @@ enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, E ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k, - ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu, - ESC_REF }; + ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu }; /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to OP_EOD must correspond in order to the list of escapes immediately above. @@ -1298,7 +1914,7 @@ enum { OP_NOT_WORDCHAR, /* 10 \W */ OP_WORDCHAR, /* 11 \w */ - OP_ANY, /* 12 Match any character except newline */ + OP_ANY, /* 12 Match any character except newline (\N) */ OP_ALLANY, /* 13 Match any character */ OP_ANYBYTE, /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */ OP_NOTPROP, /* 15 \P (not Unicode property) */ @@ -1309,8 +1925,8 @@ enum { OP_NOT_VSPACE, /* 20 \V (not vertical whitespace) */ OP_VSPACE, /* 21 \v (vertical whitespace) */ OP_EXTUNI, /* 22 \X (extended Unicode sequence */ - OP_EODN, /* 23 End of data or \n at end of data: \Z. */ - OP_EOD, /* 24 End of data: \z */ + OP_EODN, /* 23 End of data or \n at end of data (\Z) */ + OP_EOD, /* 24 End of data (\z) */ OP_CIRC, /* 25 Start of line - not multiline */ OP_CIRCM, /* 26 Start of line - multiline */ @@ -1433,8 +2049,8 @@ enum { OP_CLASS, /* 106 Match a character class, chars < 256 only */ OP_NCLASS, /* 107 Same, but the bitmap was created from a negative class - the difference is relevant only when a - UTF-8 character > 255 is encountered. */ - OP_XCLASS, /* 108 Extended class for handling UTF-8 chars within the + character > 255 is encountered. */ + OP_XCLASS, /* 108 Extended class for handling > 255 chars within the class. This does both positive and negative. */ OP_REF, /* 109 Match a back reference, casefully */ OP_REFI, /* 110 Match a back reference, caselessly */ @@ -1591,30 +2207,35 @@ in UTF-8 mode. The code that uses this table must know 2, /* noti */ \ /* Positive single-char repeats ** These are */ \ 2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \ - 4, 4, 4, /* upto, minupto, exact ** mode */ \ - 2, 2, 2, 4, /* *+, ++, ?+, upto+ */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* upto, minupto ** mode */ \ + 2+IMM2_SIZE, /* exact */ \ + 2, 2, 2, 2+IMM2_SIZE, /* *+, ++, ?+, upto+ */ \ 2, 2, 2, 2, 2, 2, /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8 */ \ - 4, 4, 4, /* upto I, minupto I, exact I */ \ - 2, 2, 2, 4, /* *+I, ++I, ?+I, upto+I */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* upto I, minupto I */ \ + 2+IMM2_SIZE, /* exact I */ \ + 2, 2, 2, 2+IMM2_SIZE, /* *+I, ++I, ?+I, upto+I */ \ /* Negative single-char repeats - only for chars < 256 */ \ 2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \ - 4, 4, 4, /* NOT upto, minupto, exact */ \ - 2, 2, 2, 4, /* Possessive NOT *, +, ?, upto */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* NOT upto, minupto */ \ + 2+IMM2_SIZE, /* NOT exact */ \ + 2, 2, 2, 2+IMM2_SIZE, /* Possessive NOT *, +, ?, upto */ \ 2, 2, 2, 2, 2, 2, /* NOT *I, *?I, +I, +?I, ?I, ??I */ \ - 4, 4, 4, /* NOT upto I, minupto I, exact I */ \ - 2, 2, 2, 4, /* Possessive NOT *I, +I, ?I, upto I */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* NOT upto I, minupto I */ \ + 2+IMM2_SIZE, /* NOT exact I */ \ + 2, 2, 2, 2+IMM2_SIZE, /* Possessive NOT *I, +I, ?I, upto I */ \ /* Positive type repeats */ \ 2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \ - 4, 4, 4, /* Type upto, minupto, exact */ \ - 2, 2, 2, 4, /* Possessive *+, ++, ?+, upto+ */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* Type upto, minupto */ \ + 2+IMM2_SIZE, /* Type exact */ \ + 2, 2, 2, 2+IMM2_SIZE, /* Possessive *+, ++, ?+, upto+ */ \ /* Character class & ref repeats */ \ 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \ - 5, 5, /* CRRANGE, CRMINRANGE */ \ - 33, /* CLASS */ \ - 33, /* NCLASS */ \ + 1+2*IMM2_SIZE, 1+2*IMM2_SIZE, /* CRRANGE, CRMINRANGE */ \ + 1+(32/sizeof(pcre_uchar)), /* CLASS */ \ + 1+(32/sizeof(pcre_uchar)), /* NCLASS */ \ 0, /* XCLASS - variable length */ \ - 3, /* REF */ \ - 3, /* REFI */ \ + 1+IMM2_SIZE, /* REF */ \ + 1+IMM2_SIZE, /* REFI */ \ 1+LINK_SIZE, /* RECURSE */ \ 2+2*LINK_SIZE, /* CALLOUT */ \ 1+LINK_SIZE, /* Alt */ \ @@ -1631,23 +2252,23 @@ in UTF-8 mode. The code that uses this table must know 1+LINK_SIZE, /* ONCE_NC */ \ 1+LINK_SIZE, /* BRA */ \ 1+LINK_SIZE, /* BRAPOS */ \ - 3+LINK_SIZE, /* CBRA */ \ - 3+LINK_SIZE, /* CBRAPOS */ \ + 1+LINK_SIZE+IMM2_SIZE, /* CBRA */ \ + 1+LINK_SIZE+IMM2_SIZE, /* CBRAPOS */ \ 1+LINK_SIZE, /* COND */ \ 1+LINK_SIZE, /* SBRA */ \ 1+LINK_SIZE, /* SBRAPOS */ \ - 3+LINK_SIZE, /* SCBRA */ \ - 3+LINK_SIZE, /* SCBRAPOS */ \ + 1+LINK_SIZE+IMM2_SIZE, /* SCBRA */ \ + 1+LINK_SIZE+IMM2_SIZE, /* SCBRAPOS */ \ 1+LINK_SIZE, /* SCOND */ \ - 3, 3, /* CREF, NCREF */ \ - 3, 3, /* RREF, NRREF */ \ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* CREF, NCREF */ \ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* RREF, NRREF */ \ 1, /* DEF */ \ 1, 1, 1, /* BRAZERO, BRAMINZERO, BRAPOSZERO */ \ 3, 1, 3, /* MARK, PRUNE, PRUNE_ARG */ \ 1, 3, /* SKIP, SKIP_ARG */ \ 1, 3, /* THEN, THEN_ARG */ \ 1, 1, 1, 1, /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT */ \ - 3, 1 /* CLOSE, SKIPZERO */ + 1+IMM2_SIZE, 1 /* CLOSE, SKIPZERO */ /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion" condition. */ @@ -1665,51 +2286,112 @@ enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49, ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59, ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69, - ERR70, ERR71, ERR72, ERRCOUNT }; + ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERRCOUNT }; +/* JIT compiling modes. The function list is indexed by them. */ +enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE, + JIT_NUMBER_OF_COMPILE_MODES }; + /* The real format of the start of the pcre block; the index of names and the code vector run on as long as necessary after the end. We store an explicit offset to the name table so that if a regex is compiled on one host, saved, and then run on another where the size of pointers is different, all might still -be well. For the case of compiled-on-4 and run-on-8, we include an extra -pointer that is always NULL. For future-proofing, a few dummy fields were -originally included - even though you can never get this planning right - but -there is only one left now. +be well. -NOTE NOTE NOTE: -Because people can now save and re-use compiled patterns, any additions to this -structure should be made at the end, and something earlier (e.g. a new -flag in the options or one of the dummy fields) should indicate that the new -fields are present. Currently PCRE always sets the dummy fields to zero. -NOTE NOTE NOTE +The size of the structure must be a multiple of 8 bytes. For the case of +compiled-on-4 and run-on-8, we include an extra pointer that is always NULL so +that there are an even number of pointers which therefore are a multiple of 8 +bytes. + +It is necessary to fork the struct for the 32 bit library, since it needs to +use pcre_uint32 for first_char and req_char. We can't put an ifdef inside the +typedef because pcretest needs access to the struct of the 8-, 16- and 32-bit +variants. + +*** WARNING *** +When new fields are added to these structures, remember to adjust the code in +pcre_byte_order.c that is concerned with swapping the byte order of the fields +when a compiled regex is reloaded on a host with different endianness. +*** WARNING *** +There is also similar byte-flipping code in pcretest.c, which is used for +testing the byte-flipping features. It must also be kept in step. +*** WARNING *** */ -typedef struct real_pcre { +typedef struct real_pcre8_or_16 { pcre_uint32 magic_number; pcre_uint32 size; /* Total that was malloced */ pcre_uint32 options; /* Public options */ - pcre_uint16 flags; /* Private flags */ - pcre_uint16 dummy1; /* For future use */ - pcre_uint16 top_bracket; - pcre_uint16 top_backref; - pcre_uint16 first_byte; - pcre_uint16 req_byte; + pcre_uint32 flags; /* Private flags */ + pcre_uint32 limit_match; /* Limit set from regex */ + pcre_uint32 limit_recursion; /* Limit set from regex */ + pcre_uint16 first_char; /* Starting character */ + pcre_uint16 req_char; /* This character must be seen */ + pcre_uint16 max_lookbehind; /* Longest lookbehind (characters) */ + pcre_uint16 top_bracket; /* Highest numbered group */ + pcre_uint16 top_backref; /* Highest numbered back reference */ pcre_uint16 name_table_offset; /* Offset to name table that follows */ pcre_uint16 name_entry_size; /* Size of any name items */ pcre_uint16 name_count; /* Number of name items */ pcre_uint16 ref_count; /* Reference count */ + pcre_uint16 dummy1; /* To ensure size is a multiple of 8 */ + pcre_uint16 dummy2; /* To ensure size is a multiple of 8 */ + pcre_uint16 dummy3; /* To ensure size is a multiple of 8 */ + const pcre_uint8 *tables; /* Pointer to tables or NULL for std */ + void *nullpad; /* NULL padding */ +} real_pcre8_or_16; - const unsigned char *tables; /* Pointer to tables or NULL for std */ - const unsigned char *nullpad; /* NULL padding */ -} real_pcre; +typedef struct real_pcre8_or_16 real_pcre; +typedef struct real_pcre8_or_16 real_pcre16; +typedef struct real_pcre32 { + pcre_uint32 magic_number; + pcre_uint32 size; /* Total that was malloced */ + pcre_uint32 options; /* Public options */ + pcre_uint32 flags; /* Private flags */ + pcre_uint32 limit_match; /* Limit set from regex */ + pcre_uint32 limit_recursion; /* Limit set from regex */ + pcre_uint32 first_char; /* Starting character */ + pcre_uint32 req_char; /* This character must be seen */ + pcre_uint16 max_lookbehind; /* Longest lookbehind (characters) */ + pcre_uint16 top_bracket; /* Highest numbered group */ + pcre_uint16 top_backref; /* Highest numbered back reference */ + pcre_uint16 name_table_offset; /* Offset to name table that follows */ + pcre_uint16 name_entry_size; /* Size of any name items */ + pcre_uint16 name_count; /* Number of name items */ + pcre_uint16 ref_count; /* Reference count */ + pcre_uint16 dummy; /* To ensure size is a multiple of 8 */ + const pcre_uint8 *tables; /* Pointer to tables or NULL for std */ + void *nullpad; /* NULL padding */ +} real_pcre32; + +#if defined COMPILE_PCRE8 +#define REAL_PCRE real_pcre +#elif defined COMPILE_PCRE16 +#define REAL_PCRE real_pcre16 +#elif defined COMPILE_PCRE32 +#define REAL_PCRE real_pcre32 +#endif + +/* Assert that the size of REAL_PCRE is divisible by 8 */ +typedef int __assert_real_pcre_size_divisible_8[(sizeof(REAL_PCRE) % 8) == 0 ? 1 : -1]; + +/* Needed in pcretest to access some fields in the real_pcre* structures + * directly. They're unified for 8/16/32 bits since the structs only differ + * after these fields; if that ever changes, need to fork those defines into + * 8/16 and 32 bit versions. */ +#define REAL_PCRE_MAGIC(re) (((REAL_PCRE*)re)->magic_number) +#define REAL_PCRE_SIZE(re) (((REAL_PCRE*)re)->size) +#define REAL_PCRE_OPTIONS(re) (((REAL_PCRE*)re)->options) +#define REAL_PCRE_FLAGS(re) (((REAL_PCRE*)re)->flags) + /* The format of the block used to store data from pcre_study(). The same remark (see NOTE above) about extending this structure applies. */ typedef struct pcre_study_data { pcre_uint32 size; /* Total that was malloced */ pcre_uint32 flags; /* Private flags */ - uschar start_bits[32]; /* Starting char bits */ + pcre_uint8 start_bits[32]; /* Starting char bits */ pcre_uint32 minlength; /* Minimum subject length */ } pcre_study_data; @@ -1728,33 +2410,35 @@ typedef struct open_capitem { doing the compiling, so that they are thread-safe. */ typedef struct compile_data { - const uschar *lcc; /* Points to lower casing table */ - const uschar *fcc; /* Points to case-flipping table */ - const uschar *cbits; /* Points to character type table */ - const uschar *ctypes; /* Points to table of type maps */ - const uschar *start_workspace;/* The start of working space */ - const uschar *start_code; /* The start of the compiled code */ - const uschar *start_pattern; /* The start of the pattern */ - const uschar *end_pattern; /* The end of the pattern */ - open_capitem *open_caps; /* Chain of open capture items */ - uschar *hwm; /* High watermark of workspace */ - uschar *name_table; /* The name/number table */ - int names_found; /* Number of entries so far */ - int name_entry_size; /* Size of each entry */ - int workspace_size; /* Size of workspace */ - int bracount; /* Count of capturing parens as we compile */ - int final_bracount; /* Saved value after first pass */ - int top_backref; /* Maximum back reference */ - unsigned int backref_map; /* Bitmap of low back refs */ - int assert_depth; /* Depth of nested assertions */ - int external_options; /* External (initial) options */ - int external_flags; /* External flag bits to be set */ - int req_varyopt; /* "After variable item" flag for reqbyte */ - BOOL had_accept; /* (*ACCEPT) encountered */ - BOOL check_lookbehind; /* Lookbehinds need later checking */ - int nltype; /* Newline type */ - int nllen; /* Newline string length */ - uschar nl[4]; /* Newline string when fixed length */ + const pcre_uint8 *lcc; /* Points to lower casing table */ + const pcre_uint8 *fcc; /* Points to case-flipping table */ + const pcre_uint8 *cbits; /* Points to character type table */ + const pcre_uint8 *ctypes; /* Points to table of type maps */ + const pcre_uchar *start_workspace;/* The start of working space */ + const pcre_uchar *start_code; /* The start of the compiled code */ + const pcre_uchar *start_pattern; /* The start of the pattern */ + const pcre_uchar *end_pattern; /* The end of the pattern */ + open_capitem *open_caps; /* Chain of open capture items */ + pcre_uchar *hwm; /* High watermark of workspace */ + pcre_uchar *name_table; /* The name/number table */ + int names_found; /* Number of entries so far */ + int name_entry_size; /* Size of each entry */ + int workspace_size; /* Size of workspace */ + unsigned int bracount; /* Count of capturing parens as we compile */ + int final_bracount; /* Saved value after first pass */ + int max_lookbehind; /* Maximum lookbehind (characters) */ + int top_backref; /* Maximum back reference */ + unsigned int backref_map; /* Bitmap of low back refs */ + int assert_depth; /* Depth of nested assertions */ + pcre_uint32 external_options; /* External (initial) options */ + pcre_uint32 external_flags; /* External flag bits to be set */ + int req_varyopt; /* "After variable item" flag for reqbyte */ + BOOL had_accept; /* (*ACCEPT) encountered */ + BOOL had_pruneorskip; /* (*PRUNE) or (*SKIP) encountered */ + BOOL check_lookbehind; /* Lookbehinds need later checking */ + int nltype; /* Newline type */ + int nllen; /* Newline string length */ + pcre_uchar nl[4]; /* Newline string when fixed length */ } compile_data; /* Structure for maintaining a chain of pointers to the currently incomplete @@ -1762,7 +2446,7 @@ branches, for testing for left recursion while compili typedef struct branch_chain { struct branch_chain *outer; - uschar *current_branch; + pcre_uchar *current_branch; } branch_chain; /* Structure for items in a linked list that represents an explicit recursive @@ -1770,10 +2454,11 @@ call within the pattern; used by pcre_exec(). */ typedef struct recursion_info { struct recursion_info *prevrec; /* Previous recursion record (or NULL) */ - int group_num; /* Number of group that was called */ + unsigned int group_num; /* Number of group that was called */ int *offset_save; /* Pointer to start of saved offsets */ int saved_max; /* Number of saved offsets */ - USPTR subject_position; /* Position at start of recursion */ + int saved_capture_last; /* Last capture number */ + PCRE_PUCHAR subject_position; /* Position at start of recursion */ } recursion_info; /* A similar structure for pcre_dfa_exec(). */ @@ -1781,7 +2466,7 @@ typedef struct recursion_info { typedef struct dfa_recursion_info { struct dfa_recursion_info *prevrec; int group_num; - USPTR subject_position; + PCRE_PUCHAR subject_position; } dfa_recursion_info; /* Structure for building a chain of data for holding the values of the subject @@ -1791,7 +2476,7 @@ pcre_exec(). */ typedef struct eptrblock { struct eptrblock *epb_prev; - USPTR epb_saved_eptr; + PCRE_PUCHAR epb_saved_eptr; } eptrblock; @@ -1802,67 +2487,71 @@ typedef struct match_data { unsigned long int match_call_count; /* As it says */ unsigned long int match_limit; /* As it says */ unsigned long int match_limit_recursion; /* As it says */ - int *offset_vector; /* Offset vector */ - int offset_end; /* One past the end */ - int offset_max; /* The maximum usable for return data */ - int nltype; /* Newline type */ - int nllen; /* Newline string length */ - int name_count; /* Number of names in name table */ - int name_entry_size; /* Size of entry in names table */ - uschar *name_table; /* Table of names */ - uschar nl[4]; /* Newline string when fixed */ - const uschar *lcc; /* Points to lower casing table */ - const uschar *ctypes; /* Points to table of type maps */ - BOOL offset_overflow; /* Set if too many extractions */ - BOOL notbol; /* NOTBOL flag */ - BOOL noteol; /* NOTEOL flag */ - BOOL utf8; /* UTF8 flag */ - BOOL jscript_compat; /* JAVASCRIPT_COMPAT flag */ - BOOL use_ucp; /* PCRE_UCP flag */ - BOOL endonly; /* Dollar not before final \n */ - BOOL notempty; /* Empty string match not wanted */ - BOOL notempty_atstart; /* Empty string match at start not wanted */ - BOOL hitend; /* Hit the end of the subject at some point */ - BOOL bsr_anycrlf; /* \R is just any CRLF, not full Unicode */ - BOOL hasthen; /* Pattern contains (*THEN) */ - BOOL ignore_skip_arg; /* For re-run when SKIP name not found */ - const uschar *start_code; /* For use when recursing */ - USPTR start_subject; /* Start of the subject string */ - USPTR end_subject; /* End of the subject string */ - USPTR start_match_ptr; /* Start of matched string */ - USPTR end_match_ptr; /* Subject position at end match */ - USPTR start_used_ptr; /* Earliest consulted character */ - int partial; /* PARTIAL options */ - int end_offset_top; /* Highwater mark at end of match */ - int capture_last; /* Most recent capture number */ - int start_offset; /* The start offset value */ - int match_function_type; /* Set for certain special calls of MATCH() */ - eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */ - int eptrn; /* Next free eptrblock */ - recursion_info *recursive; /* Linked list of recursion data */ - void *callout_data; /* To pass back to callouts */ - const uschar *mark; /* Mark pointer to pass back on success */ - const uschar *nomatch_mark; /* Mark pointer to pass back on failure */ - const uschar *once_target; /* Where to back up to for atomic groups */ + int *offset_vector; /* Offset vector */ + int offset_end; /* One past the end */ + int offset_max; /* The maximum usable for return data */ + int nltype; /* Newline type */ + int nllen; /* Newline string length */ + int name_count; /* Number of names in name table */ + int name_entry_size; /* Size of entry in names table */ + unsigned int skip_arg_count; /* For counting SKIP_ARGs */ + unsigned int ignore_skip_arg; /* For re-run when SKIP arg name not found */ + pcre_uchar *name_table; /* Table of names */ + pcre_uchar nl[4]; /* Newline string when fixed */ + const pcre_uint8 *lcc; /* Points to lower casing table */ + const pcre_uint8 *fcc; /* Points to case-flipping table */ + const pcre_uint8 *ctypes; /* Points to table of type maps */ + BOOL notbol; /* NOTBOL flag */ + BOOL noteol; /* NOTEOL flag */ + BOOL utf; /* UTF-8 / UTF-16 flag */ + BOOL jscript_compat; /* JAVASCRIPT_COMPAT flag */ + BOOL use_ucp; /* PCRE_UCP flag */ + BOOL endonly; /* Dollar not before final \n */ + BOOL notempty; /* Empty string match not wanted */ + BOOL notempty_atstart; /* Empty string match at start not wanted */ + BOOL hitend; /* Hit the end of the subject at some point */ + BOOL bsr_anycrlf; /* \R is just any CRLF, not full Unicode */ + BOOL hasthen; /* Pattern contains (*THEN) */ + const pcre_uchar *start_code; /* For use when recursing */ + PCRE_PUCHAR start_subject; /* Start of the subject string */ + PCRE_PUCHAR end_subject; /* End of the subject string */ + PCRE_PUCHAR start_match_ptr; /* Start of matched string */ + PCRE_PUCHAR end_match_ptr; /* Subject position at end match */ + PCRE_PUCHAR start_used_ptr; /* Earliest consulted character */ + int partial; /* PARTIAL options */ + int end_offset_top; /* Highwater mark at end of match */ + pcre_int32 capture_last; /* Most recent capture number + overflow flag */ + int start_offset; /* The start offset value */ + int match_function_type; /* Set for certain special calls of MATCH() */ + eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */ + int eptrn; /* Next free eptrblock */ + recursion_info *recursive; /* Linked list of recursion data */ + void *callout_data; /* To pass back to callouts */ + const pcre_uchar *mark; /* Mark pointer to pass back on success */ + const pcre_uchar *nomatch_mark;/* Mark pointer to pass back on failure */ + const pcre_uchar *once_target; /* Where to back up to for atomic groups */ +#ifdef NO_RECURSE + void *match_frames_base; /* For remembering malloc'd frames */ +#endif } match_data; /* A similar structure is used for the same purpose by the DFA matching functions. */ typedef struct dfa_match_data { - const uschar *start_code; /* Start of the compiled pattern */ - const uschar *start_subject; /* Start of the subject string */ - const uschar *end_subject; /* End of subject string */ - const uschar *start_used_ptr; /* Earliest consulted character */ - const uschar *tables; /* Character tables */ - int start_offset; /* The start offset value */ - int moptions; /* Match options */ - int poptions; /* Pattern options */ - int nltype; /* Newline type */ - int nllen; /* Newline string length */ - uschar nl[4]; /* Newline string when fixed */ - void *callout_data; /* To pass back to callouts */ - dfa_recursion_info *recursive; /* Linked list of recursion data */ + const pcre_uchar *start_code; /* Start of the compiled pattern */ + const pcre_uchar *start_subject ; /* Start of the subject string */ + const pcre_uchar *end_subject; /* End of subject string */ + const pcre_uchar *start_used_ptr; /* Earliest consulted character */ + const pcre_uint8 *tables; /* Character tables */ + int start_offset; /* The start offset value */ + int moptions; /* Match options */ + int poptions; /* Pattern options */ + int nltype; /* Newline type */ + int nllen; /* Newline string length */ + pcre_uchar nl[4]; /* Newline string when fixed */ + void *callout_data; /* To pass back to callouts */ + dfa_recursion_info *recursive; /* Linked list of recursion data */ } dfa_match_data; /* Bit definitions for entries in the pcre_ctypes table. */ @@ -1898,6 +2587,33 @@ total length. */ #define ctypes_offset (cbits_offset + cbit_length) #define tables_length (ctypes_offset + 256) +/* Internal function and data prefixes. */ + +#if defined COMPILE_PCRE8 +#ifndef PUBL +#define PUBL(name) pcre_##name +#endif +#ifndef PRIV +#define PRIV(name) _pcre_##name +#endif +#elif defined COMPILE_PCRE16 +#ifndef PUBL +#define PUBL(name) pcre16_##name +#endif +#ifndef PRIV +#define PRIV(name) _pcre16_##name +#endif +#elif defined COMPILE_PCRE32 +#ifndef PUBL +#define PUBL(name) pcre32_##name +#endif +#ifndef PRIV +#define PRIV(name) _pcre32_##name +#endif +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE[8|16|32] */ + /* Layout of the UCP type table that translates property names into types and codes. Each entry used to point directly to a name, but to reduce the number of relocations in shared libraries, it now has an offset into a single string @@ -1915,74 +2631,140 @@ of the exported public functions. They have to be "ext but are not part of the PCRE public API. The data for these tables is in the pcre_tables.c module. */ -extern const int _pcre_utf8_table1[]; -extern const int _pcre_utf8_table2[]; -extern const int _pcre_utf8_table3[]; -extern const uschar _pcre_utf8_table4[]; +#ifdef COMPILE_PCRE8 +extern const int PRIV(utf8_table1)[]; +extern const int PRIV(utf8_table1_size); +extern const int PRIV(utf8_table2)[]; +extern const int PRIV(utf8_table3)[]; +extern const pcre_uint8 PRIV(utf8_table4)[]; +#endif /* COMPILE_PCRE8 */ -#ifdef SUPPORT_JIT -extern const uschar _pcre_utf8_char_sizes[]; -#endif +extern const char PRIV(utt_names)[]; +extern const ucp_type_table PRIV(utt)[]; +extern const int PRIV(utt_size); -extern const int _pcre_utf8_table1_size; +extern const pcre_uint8 PRIV(OP_lengths)[]; +extern const pcre_uint8 PRIV(default_tables)[]; -extern const char _pcre_utt_names[]; -extern const ucp_type_table _pcre_utt[]; -extern const int _pcre_utt_size; +extern const pcre_uint32 PRIV(hspace_list)[]; +extern const pcre_uint32 PRIV(vspace_list)[]; -extern const uschar _pcre_default_tables[]; -extern const uschar _pcre_OP_lengths[]; - - /* Internal shared functions. These are functions that are used by more than one of the exported public functions. They have to be "external" in the C sense, but are not part of the PCRE public API. */ -extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int); -extern BOOL _pcre_is_newline(USPTR, int, USPTR, int *, BOOL); -extern int _pcre_ord2utf8(int, uschar *); -extern real_pcre *_pcre_try_flipped(const real_pcre *, real_pcre *, - const pcre_study_data *, pcre_study_data *); -extern int _pcre_valid_utf8(USPTR, int, int *); -extern BOOL _pcre_was_newline(USPTR, int, USPTR, int *, BOOL); -extern BOOL _pcre_xclass(int, const uschar *); +/* String comparison functions. */ +#if defined COMPILE_PCRE8 +#define STRCMP_UC_UC(str1, str2) \ + strcmp((char *)(str1), (char *)(str2)) +#define STRCMP_UC_C8(str1, str2) \ + strcmp((char *)(str1), (str2)) +#define STRNCMP_UC_UC(str1, str2, num) \ + strncmp((char *)(str1), (char *)(str2), (num)) +#define STRNCMP_UC_C8(str1, str2, num) \ + strncmp((char *)(str1), (str2), (num)) +#define STRLEN_UC(str) strlen((const char *)str) + +#elif defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + +extern int PRIV(strcmp_uc_uc)(const pcre_uchar *, + const pcre_uchar *); +extern int PRIV(strcmp_uc_c8)(const pcre_uchar *, + const char *); +extern int PRIV(strncmp_uc_uc)(const pcre_uchar *, + const pcre_uchar *, unsigned int num); +extern int PRIV(strncmp_uc_c8)(const pcre_uchar *, + const char *, unsigned int num); +extern unsigned int PRIV(strlen_uc)(const pcre_uchar *str); + +#define STRCMP_UC_UC(str1, str2) \ + PRIV(strcmp_uc_uc)((str1), (str2)) +#define STRCMP_UC_C8(str1, str2) \ + PRIV(strcmp_uc_c8)((str1), (str2)) +#define STRNCMP_UC_UC(str1, str2, num) \ + PRIV(strncmp_uc_uc)((str1), (str2), (num)) +#define STRNCMP_UC_C8(str1, str2, num) \ + PRIV(strncmp_uc_c8)((str1), (str2), (num)) +#define STRLEN_UC(str) PRIV(strlen_uc)(str) + +#endif /* COMPILE_PCRE[8|16|32] */ + +#if defined COMPILE_PCRE8 || defined COMPILE_PCRE16 + +#define STRCMP_UC_UC_TEST(str1, str2) STRCMP_UC_UC(str1, str2) +#define STRCMP_UC_C8_TEST(str1, str2) STRCMP_UC_C8(str1, str2) + +#elif defined COMPILE_PCRE32 + +extern int PRIV(strcmp_uc_uc_utf)(const pcre_uchar *, + const pcre_uchar *); +extern int PRIV(strcmp_uc_c8_utf)(const pcre_uchar *, + const char *); + +#define STRCMP_UC_UC_TEST(str1, str2) \ + (utf ? PRIV(strcmp_uc_uc_utf)((str1), (str2)) : PRIV(strcmp_uc_uc)((str1), (str2))) +#define STRCMP_UC_C8_TEST(str1, str2) \ + (utf ? PRIV(strcmp_uc_c8_utf)((str1), (str2)) : PRIV(strcmp_uc_c8)((str1), (str2))) + +#endif /* COMPILE_PCRE[8|16|32] */ + +extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int); +extern BOOL PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR, + int *, BOOL); +extern unsigned int PRIV(ord2utf)(pcre_uint32, pcre_uchar *); +extern int PRIV(valid_utf)(PCRE_PUCHAR, int, int *); +extern BOOL PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR, + int *, BOOL); +extern BOOL PRIV(xclass)(pcre_uint32, const pcre_uchar *, BOOL); + #ifdef SUPPORT_JIT -extern void _pcre_jit_compile(const real_pcre *, pcre_extra *); -extern int _pcre_jit_exec(const real_pcre *, void *, PCRE_SPTR, - int, int, int, int, int *, int); -extern void _pcre_jit_free(void *); -extern int _pcre_jit_get_size(void *); +extern void PRIV(jit_compile)(const REAL_PCRE *, + PUBL(extra) *, int); +extern int PRIV(jit_exec)(const PUBL(extra) *, + const pcre_uchar *, int, int, int, int *, int); +extern void PRIV(jit_free)(void *); +extern int PRIV(jit_get_size)(void *); +extern const char* PRIV(jit_get_target)(void); #endif /* Unicode character database (UCD) */ typedef struct { - uschar script; - uschar chartype; - pcre_int32 other_case; + pcre_uint8 script; /* ucp_Arabic, etc. */ + pcre_uint8 chartype; /* ucp_Cc, etc. (general categories) */ + pcre_uint8 gbprop; /* ucp_gbControl, etc. (grapheme break property) */ + pcre_uint8 caseset; /* offset to multichar other cases or zero */ + pcre_int32 other_case; /* offset to other case, or zero if none */ } ucd_record; -extern const ucd_record _pcre_ucd_records[]; -extern const uschar _pcre_ucd_stage1[]; -extern const pcre_uint16 _pcre_ucd_stage2[]; -extern const int _pcre_ucp_gentype[]; +extern const pcre_uint32 PRIV(ucd_caseless_sets)[]; +extern const ucd_record PRIV(ucd_records)[]; +extern const pcre_uint8 PRIV(ucd_stage1)[]; +extern const pcre_uint16 PRIV(ucd_stage2)[]; +extern const pcre_uint32 PRIV(ucp_gentype)[]; +extern const pcre_uint32 PRIV(ucp_gbtable)[]; #ifdef SUPPORT_JIT -extern const int _pcre_ucp_typerange[]; +extern const int PRIV(ucp_typerange)[]; #endif +#ifdef SUPPORT_UCP /* UCD access macros */ #define UCD_BLOCK_SIZE 128 -#define GET_UCD(ch) (_pcre_ucd_records + \ - _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \ - UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE]) +#define GET_UCD(ch) (PRIV(ucd_records) + \ + PRIV(ucd_stage2)[PRIV(ucd_stage1)[(int)(ch) / UCD_BLOCK_SIZE] * \ + UCD_BLOCK_SIZE + (int)(ch) % UCD_BLOCK_SIZE]) -#define UCD_CHARTYPE(ch) GET_UCD(ch)->chartype -#define UCD_SCRIPT(ch) GET_UCD(ch)->script -#define UCD_CATEGORY(ch) _pcre_ucp_gentype[UCD_CHARTYPE(ch)] -#define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case) +#define UCD_CHARTYPE(ch) GET_UCD(ch)->chartype +#define UCD_SCRIPT(ch) GET_UCD(ch)->script +#define UCD_CATEGORY(ch) PRIV(ucp_gentype)[UCD_CHARTYPE(ch)] +#define UCD_GRAPHBREAK(ch) GET_UCD(ch)->gbprop +#define UCD_CASESET(ch) GET_UCD(ch)->caseset +#define UCD_OTHERCASE(ch) ((pcre_uint32)((int)ch + (int)(GET_UCD(ch)->other_case))) + +#endif /* SUPPORT_UCP */ #endif