X-Git-Url: https://perl5.git.perl.org/perl5.git/blobdiff_plain/28d03b21d43611f97565be8d077b7f8eb0bda26a..2f9365dc3b09bdf83c00a6d176d882057608308e:/regexp.h diff --git a/regexp.h b/regexp.h index a86963c..5dbab2e 100644 --- a/regexp.h +++ b/regexp.h @@ -36,13 +36,14 @@ struct regexp_engine; struct regexp; struct reg_substr_datum { - I32 min_offset; - I32 max_offset; + SSize_t min_offset; /* min pos (in chars) that substr must appear */ + SSize_t max_offset /* max pos (in chars) that substr must appear */; SV *substr; /* non-utf8 variant */ SV *utf8_substr; /* utf8 variant */ - I32 end_shift; + SSize_t end_shift; /* how many fixed chars must end the string */ }; struct reg_substr_data { + U8 check_ix; /* index into data[] of check substr */ struct reg_substr_datum data[3]; /* Actual array */ }; @@ -55,15 +56,15 @@ struct reg_substr_data { /* offsets within a string of a particular /(.)/ capture */ typedef struct regexp_paren_pair { - I32 start; - I32 end; + SSize_t start; + SSize_t end; /* 'start_tmp' records a new opening position before the matching end * has been found, so that the old start and end values are still * valid, e.g. * "abc" =~ /(.(?{print "[$1]"}))+/ *outputs [][a][b] * This field is not part of the API. */ - I32 start_tmp; + SSize_t start_tmp; } regexp_paren_pair; #if defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_UTF8_C) @@ -104,9 +105,9 @@ struct reg_code_block { /* Information about the match that the perl core uses to */ \ /* manage things */ \ U32 extflags; /* Flags used both externally and internally */ \ - I32 minlen; /* mininum possible number of chars in string to match */\ - I32 minlenret; /* mininum possible number of chars in $& */ \ - U32 gofs; /* chars left of pos that we search from */ \ + SSize_t minlen; /* mininum possible number of chars in string to match */\ + SSize_t minlenret; /* mininum possible number of chars in $& */ \ + STRLEN gofs; /* chars left of pos that we search from */ \ /* substring data about strings that must appear in the */ \ /* final match, used for optimisations */ \ struct reg_substr_data *substrs; \ @@ -124,10 +125,11 @@ struct reg_code_block { /* saved or original string so \digit works forever. */ \ char *subbeg; \ SV_SAVED_COPY /* If non-NULL, SV which is COW from original */\ - I32 sublen; /* Length of string pointed by subbeg */ \ - I32 suboffset; /* byte offset of subbeg from logical start of str */ \ - I32 subcoffset; /* suboffset equiv, but in chars (for @-/@+) */ \ + SSize_t sublen; /* Length of string pointed by subbeg */ \ + SSize_t suboffset; /* byte offset of subbeg from logical start of str */ \ + SSize_t subcoffset; /* suboffset equiv, but in chars (for @-/@+) */ \ /* Information about the match that isn't often used */ \ + SSize_t maxlen; /* mininum possible number of chars in string to match */\ /* offset from wrapped to the start of precomp */ \ PERL_BITFIELD32 pre_prefix:4; \ /* original flags used to compile the pattern, may differ */ \ @@ -146,7 +148,7 @@ typedef struct regexp { typedef struct re_scream_pos_data_s { char **scream_olds; /* match pos */ - I32 *scream_pos; /* Internal iterator of scream. */ + SSize_t *scream_pos; /* Internal iterator of scream. */ } re_scream_pos_data; /* regexp_engine structure. This is the dispatch table for regexes. @@ -155,7 +157,7 @@ typedef struct re_scream_pos_data_s typedef struct regexp_engine { REGEXP* (*comp) (pTHX_ SV * const pattern, U32 flags); I32 (*exec) (pTHX_ REGEXP * const rx, char* stringarg, char* strend, - char* strbeg, I32 minend, SV* screamer, + char* strbeg, SSize_t minend, SV* sv, void* data, U32 flags); char* (*intuit) (pTHX_ REGEXP * const rx, @@ -166,7 +168,7 @@ typedef struct regexp_engine { const U32 flags, re_scream_pos_data *data); SV* (*checkstr) (pTHX_ REGEXP * const rx); - void (*free) (pTHX_ REGEXP * const rx); + void (*rxfree) (pTHX_ REGEXP * const rx); void (*numbered_buff_FETCH) (pTHX_ REGEXP * const rx, const I32 paren, SV * const sv); void (*numbered_buff_STORE) (pTHX_ REGEXP * const rx, const I32 paren, @@ -183,7 +185,7 @@ typedef struct regexp_engine { #endif REGEXP* (*op_comp) (pTHX_ SV ** const patternp, int pat_count, OP *expr, const struct regexp_engine* eng, - REGEXP *VOL old_re, + REGEXP *old_re, bool *is_bare_re, U32 orig_rx_flags, U32 pm_flags); } regexp_engine; @@ -230,7 +232,7 @@ typedef struct regexp_engine { =for apidoc Am|REGEXP *|SvRX|SV *sv -Convenience macro to get the REGEXP from a SV. This is approximately +Convenience macro to get the REGEXP from a SV. This is approximately equivalent to the following snippet: if (SvMAGICAL(sv)) @@ -240,7 +242,7 @@ equivalent to the following snippet: if (SvTYPE(sv) == SVt_REGEXP) return (REGEXP*) sv; -NULL will be returned if a REGEXP* is not found. +C will be returned if a REGEXP* is not found. =for apidoc Am|bool|SvRXOK|SV* sv @@ -270,17 +272,25 @@ and check for NULL. #include "op_reg_common.h" -#define RXf_PMf_STD_PMMOD (RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_FOLD|RXf_PMf_EXTENDED) +#define RXf_PMf_STD_PMMOD (RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_FOLD|RXf_PMf_EXTENDED|RXf_PMf_NOCAPTURE) -#define CASE_STD_PMMOD_FLAGS_PARSE_SET(pmfl) \ - case IGNORE_PAT_MOD: *(pmfl) |= RXf_PMf_FOLD; break; \ - case MULTILINE_PAT_MOD: *(pmfl) |= RXf_PMf_MULTILINE; break; \ - case SINGLE_PAT_MOD: *(pmfl) |= RXf_PMf_SINGLELINE; break; \ - case XTENDED_PAT_MOD: *(pmfl) |= RXf_PMf_EXTENDED; break +#define CASE_STD_PMMOD_FLAGS_PARSE_SET(pmfl, x_count) \ + case IGNORE_PAT_MOD: *(pmfl) |= RXf_PMf_FOLD; break; \ + case MULTILINE_PAT_MOD: *(pmfl) |= RXf_PMf_MULTILINE; break; \ + case SINGLE_PAT_MOD: *(pmfl) |= RXf_PMf_SINGLELINE; break; \ + case XTENDED_PAT_MOD: *(pmfl) |= RXf_PMf_EXTENDED; (x_count)++; break;\ + case NOCAPTURE_PAT_MOD: *(pmfl) |= RXf_PMf_NOCAPTURE; break; + +#define STD_PMMOD_FLAGS_PARSE_X_WARN(x_count) \ + if (UNLIKELY((x_count) > 1)) { \ + Perl_ck_warner_d(aTHX_ packWARN2(WARN_DEPRECATED, WARN_REGEXP), \ + "Having more than one /%c regexp modifier is deprecated", \ + XTENDED_PAT_MOD); \ + } /* Note, includes charset ones, assumes 0 is the default for them */ #define STD_PMMOD_FLAGS_CLEAR(pmfl) \ - *(pmfl) &= ~(RXf_PMf_FOLD|RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_EXTENDED|RXf_PMf_CHARSET) + *(pmfl) &= ~(RXf_PMf_FOLD|RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_EXTENDED|RXf_PMf_CHARSET|RXf_PMf_NOCAPTURE) /* chars and strings used as regex pattern modifiers * Singular is a 'c'har, plural is a "string" @@ -292,6 +302,7 @@ and check for NULL. #define DEFAULT_PAT_MOD '^' /* Short for all the default modifiers */ #define EXEC_PAT_MOD 'e' #define KEEPCOPY_PAT_MOD 'p' +#define NOCAPTURE_PAT_MOD 'n' #define ONCE_PAT_MOD 'o' #define GLOBAL_PAT_MOD 'g' #define CONTINUE_PAT_MOD 'c' @@ -307,6 +318,7 @@ and check for NULL. #define ONCE_PAT_MODS "o" #define KEEPCOPY_PAT_MODS "p" +#define NOCAPTURE_PAT_MODS "n" #define EXEC_PAT_MODS "e" #define LOOP_PAT_MODS "gc" #define NONDESTRUCT_PAT_MODS "r" @@ -319,7 +331,7 @@ and check for NULL. /* This string is expected by regcomp.c to be ordered so that the first * character is the flag in bit RXf_PMf_STD_PMMOD_SHIFT of extflags; the next * character is bit +1, etc. */ -#define STD_PAT_MODS "msix" +#define STD_PAT_MODS "msixxn" #define CHARSET_PAT_MODS ASCII_RESTRICT_PAT_MODS DEPENDS_PAT_MODS LOCALE_PAT_MODS UNICODE_PAT_MODS @@ -328,7 +340,7 @@ and check for NULL. * extflags; the next character is in bit +1, etc. */ #define INT_PAT_MODS STD_PAT_MODS KEEPCOPY_PAT_MODS -#define EXT_PAT_MODS ONCE_PAT_MODS KEEPCOPY_PAT_MODS +#define EXT_PAT_MODS ONCE_PAT_MODS KEEPCOPY_PAT_MODS NOCAPTURE_PAT_MODS #define QR_PAT_MODS STD_PAT_MODS EXT_PAT_MODS CHARSET_PAT_MODS #define M_PAT_MODS QR_PAT_MODS LOOP_PAT_MODS #define S_PAT_MODS M_PAT_MODS EXEC_PAT_MODS NONDESTRUCT_PAT_MODS @@ -339,97 +351,89 @@ and check for NULL. * */ -/* Leave some space, so future bit allocations can go either in the shared or - * unshared area without affecting binary compatibility */ -#define RXf_BASE_SHIFT (_RXf_PMf_SHIFT_NEXT) - /* Set in Perl_pmruntime if op_flags & OPf_SPECIAL, i.e. split. Will be used by regex engines to check whether they should set RXf_SKIPWHITE */ -#define RXf_SPLIT (1<<(RXf_BASE_SHIFT-1)) -#if RXf_SPLIT != RXf_PMf_SPLIT -# error "RXf_SPLIT does not match RXf_PMf_SPLIT" -#endif - -/* Manually decorate this function with gcc-style attributes just to - * avoid having to restructure the header files and their called order, - * as proto.h would have to be included before this file, and isn't */ - -PERL_STATIC_INLINE const char * -get_regex_charset_name(const U32 flags, STRLEN* const lenp) - __attribute__warn_unused_result__; - -#define MAX_CHARSET_NAME_LENGTH 2 +#define RXf_SPLIT RXf_PMf_SPLIT + +/* Currently the regex flags occupy a single 32-bit word. Not all bits are + * currently used. The lower bits are shared with their corresponding PMf flag + * bits, up to but not including _RXf_PMf_SHIFT_NEXT. The unused bits + * immediately follow; finally the used RXf-only (unshared) bits, so that the + * highest bit in the word is used. This gathers all the unused bits as a pool + * in the middle, like so: 11111111111111110000001111111111 + * where the '1's represent used bits, and the '0's unused. This design allows + * us to allocate off one end of the pool if we need to add a shared bit, and + * off the other end if we need a non-shared bit, without disturbing the other + * bits. This maximizes the likelihood of being able to change things without + * breaking binary compatibility. + * + * To add shared bits, do so in op_reg_common.h. This should change + * _RXf_PMf_SHIFT_NEXT so that things won't compile. Then come to regexp.h and + * op.h and adjust the constant adders in the definitions of RXf_BASE_SHIFT and + * Pmf_BASE_SHIFT down by the number of shared bits you added. That's it. + * Things should be binary compatible. But if either of these gets to having + * to subtract rather than add, leave at 0 and instead adjust all the entries + * that are in terms of it. But if the first one of those is already + * RXf_BASE_SHIFT+0, there are no bits left, and a redesign is in order. + * + * To remove unshared bits, just delete its entry. If you're where breaking + * binary compatibility is ok to do, you might want to adjust things to move + * the newly opened space so that it gets absorbed into the common pool. + * + * To add unshared bits, first use up any gaps in the middle. Otherwise, + * allocate off the low end until you get to RXf_BASE_SHIFT+0. If that isn't + * enough, move RXf_BASE_SHIFT down (if possible) and add the new bit at the + * other end instead; this preserves binary compatibility. + * + * For the regexp bits, PL_reg_extflags_name[] in regnodes.h has a comment + * giving which bits are used/unused */ -PERL_STATIC_INLINE const char * -get_regex_charset_name(const U32 flags, STRLEN* const lenp) -{ - /* Returns a string that corresponds to the name of the regex character set - * given by 'flags', and *lenp is set the length of that string, which - * cannot exceed MAX_CHARSET_NAME_LENGTH characters */ - - *lenp = 1; - switch (get_regex_charset(flags)) { - case REGEX_DEPENDS_CHARSET: return DEPENDS_PAT_MODS; - case REGEX_LOCALE_CHARSET: return LOCALE_PAT_MODS; - case REGEX_UNICODE_CHARSET: return UNICODE_PAT_MODS; - case REGEX_ASCII_RESTRICTED_CHARSET: return ASCII_RESTRICT_PAT_MODS; - case REGEX_ASCII_MORE_RESTRICTED_CHARSET: - *lenp = 2; - return ASCII_MORE_RESTRICT_PAT_MODS; - default: - return "?"; /* Unknown */ - } -} - -/* Anchor and GPOS related stuff */ -#define RXf_ANCH_BOL (1<<(RXf_BASE_SHIFT+0)) -#define RXf_ANCH_MBOL (1<<(RXf_BASE_SHIFT+1)) -#define RXf_ANCH_SBOL (1<<(RXf_BASE_SHIFT+2)) -#define RXf_ANCH_GPOS (1<<(RXf_BASE_SHIFT+3)) -#define RXf_GPOS_SEEN (1<<(RXf_BASE_SHIFT+4)) -#define RXf_GPOS_FLOAT (1<<(RXf_BASE_SHIFT+5)) -/* two bits here */ -#define RXf_ANCH (RXf_ANCH_BOL|RXf_ANCH_MBOL|RXf_ANCH_GPOS|RXf_ANCH_SBOL) -#define RXf_GPOS_CHECK (RXf_GPOS_SEEN|RXf_ANCH_GPOS) -#define RXf_ANCH_SINGLE (RXf_ANCH_SBOL|RXf_ANCH_GPOS) +#define RXf_BASE_SHIFT (_RXf_PMf_SHIFT_NEXT + 2) /* What we have seen */ -#define RXf_NO_INPLACE_SUBST (1<<(RXf_BASE_SHIFT+6)) -#define RXf_EVAL_SEEN (1<<(RXf_BASE_SHIFT+7)) -#define RXf_CANY_SEEN (1<<(RXf_BASE_SHIFT+8)) +#define RXf_NO_INPLACE_SUBST (1U<<(RXf_BASE_SHIFT+2)) +#define RXf_EVAL_SEEN (1U<<(RXf_BASE_SHIFT+3)) /* Special */ -#define RXf_NOSCAN (1<<(RXf_BASE_SHIFT+9)) -#define RXf_CHECK_ALL (1<<(RXf_BASE_SHIFT+10)) +#define RXf_UNBOUNDED_QUANTIFIER_SEEN (1U<<(RXf_BASE_SHIFT+4)) +#define RXf_CHECK_ALL (1U<<(RXf_BASE_SHIFT+5)) /* UTF8 related */ -#define RXf_MATCH_UTF8 (1<<(RXf_BASE_SHIFT+11)) /* $1 etc are utf8 */ +#define RXf_MATCH_UTF8 (1U<<(RXf_BASE_SHIFT+6)) /* $1 etc are utf8 */ /* Intuit related */ -#define RXf_USE_INTUIT_NOML (1<<(RXf_BASE_SHIFT+12)) -#define RXf_USE_INTUIT_ML (1<<(RXf_BASE_SHIFT+13)) -#define RXf_INTUIT_TAIL (1<<(RXf_BASE_SHIFT+14)) +#define RXf_USE_INTUIT_NOML (1U<<(RXf_BASE_SHIFT+7)) +#define RXf_USE_INTUIT_ML (1U<<(RXf_BASE_SHIFT+8)) +#define RXf_INTUIT_TAIL (1U<<(RXf_BASE_SHIFT+9)) #define RXf_USE_INTUIT (RXf_USE_INTUIT_NOML|RXf_USE_INTUIT_ML) +/* Do we have some sort of anchor? */ +#define RXf_IS_ANCHORED (1U<<(RXf_BASE_SHIFT+10)) + /* Copy and tainted info */ -#define RXf_COPY_DONE (1<<(RXf_BASE_SHIFT+16)) +#define RXf_COPY_DONE (1U<<(RXf_BASE_SHIFT+11)) -/* during execution: pattern temporarily tainted by executing locale ops; - * post-execution: $1 et al are tainted */ -#define RXf_TAINTED_SEEN (1<<(RXf_BASE_SHIFT+17)) +/* post-execution: $1 et al are tainted */ +#define RXf_TAINTED_SEEN (1U<<(RXf_BASE_SHIFT+12)) /* this pattern was tainted during compilation */ -#define RXf_TAINTED (1<<(RXf_BASE_SHIFT+18)) +#define RXf_TAINTED (1U<<(RXf_BASE_SHIFT+13)) /* Flags indicating special patterns */ -#define RXf_START_ONLY (1<<(RXf_BASE_SHIFT+19)) /* Pattern is /^/ */ -#define RXf_SKIPWHITE (1<<(RXf_BASE_SHIFT+20)) /* Pattern is for a split " " */ -#define RXf_WHITE (1<<(RXf_BASE_SHIFT+21)) /* Pattern is /\s+/ */ -#define RXf_NULL (1U<<(RXf_BASE_SHIFT+22)) /* Pattern is // */ -#if RXf_BASE_SHIFT+22 > 31 -# error Too many RXf_PMf bits used. See regnodes.h for any spare in middle +#define RXf_START_ONLY (1U<<(RXf_BASE_SHIFT+14)) /* Pattern is /^/ */ +#define RXf_SKIPWHITE (1U<<(RXf_BASE_SHIFT+15)) /* Pattern is for a */ + /* split " " */ +#define RXf_WHITE (1U<<(RXf_BASE_SHIFT+16)) /* Pattern is /\s+/ */ +#define RXf_NULL (1U<<(RXf_BASE_SHIFT+17)) /* Pattern is // */ + +/* See comments at the beginning of these defines about adding bits. The + * highest bit position should be used, so that if RXf_BASE_SHIFT gets + * increased, the #error below will be triggered so that you will be reminded + * to adjust things at the other end to keep the bit positions unchanged */ +#if RXf_BASE_SHIFT+17 > 31 +# error Too many RXf_PMf bits used. See comments at beginning of these for what to do #endif /* @@ -438,7 +442,7 @@ get_regex_charset_name(const U32 flags, STRLEN* const lenp) * */ -#if NO_TAINT_SUPPORT +#ifdef NO_TAINT_SUPPORT # define RX_ISTAINTED(prog) 0 # define RX_TAINT_on(prog) NOOP # define RXp_MATCH_TAINTED(prog) 0 @@ -501,6 +505,10 @@ get_regex_charset_name(const U32 flags, STRLEN* const lenp) #define RX_LASTPAREN(prog) (ReANY(prog)->lastparen) #define RX_LASTCLOSEPAREN(prog) (ReANY(prog)->lastcloseparen) #define RX_SAVED_COPY(prog) (ReANY(prog)->saved_copy) +/* last match was zero-length */ +#define RX_ZERO_LEN(prog) \ + (RX_OFFS(prog)[0].start + (SSize_t)RX_GOFS(prog) \ + == RX_OFFS(prog)[0].end) #endif /* PLUGGABLE_RE_EXTENSION */ @@ -534,16 +542,24 @@ get_regex_charset_name(const U32 flags, STRLEN* const lenp) /* Whether the pattern stored at RX_WRAPPED is in UTF-8 */ #define RX_UTF8(prog) SvUTF8(prog) -#define REXEC_COPY_STR 0x01 /* Need to copy the string. */ -#define REXEC_CHECKED 0x02 /* check_substr already checked. */ -#define REXEC_SCREAM 0x04 /* use scream table. */ -#define REXEC_IGNOREPOS 0x08 /* \G matches at start. */ -#define REXEC_NOT_FIRST 0x10 /* This is another iteration of //g. */ - /* under REXEC_COPY_STR, it's ok for the - * engine (modulo PL_sawamperand etc) - * to skip copying ... */ -#define REXEC_COPY_SKIP_PRE 0x20 /* ...the $` part of the string, or */ -#define REXEC_COPY_SKIP_POST 0x40 /* ...the $' part of the string */ + +/* bits in flags arg of Perl_regexec_flags() */ + +#define REXEC_COPY_STR 0x01 /* Need to copy the string for captures. */ +#define REXEC_CHECKED 0x02 /* re_intuit_start() already called. */ +#define REXEC_SCREAM 0x04 /* currently unused. */ +#define REXEC_IGNOREPOS 0x08 /* use stringarg, not pos(), for \G match */ +#define REXEC_NOT_FIRST 0x10 /* This is another iteration of //g: + no need to copy string again */ + + /* under REXEC_COPY_STR, it's ok for the + engine (modulo PL_sawamperand etc) + to skip copying: ... */ +#define REXEC_COPY_SKIP_PRE 0x20 /* ...the $` part of the string, or */ +#define REXEC_COPY_SKIP_POST 0x40 /* ...the $' part of the string */ +#define REXEC_FAIL_ON_UNDERFLOW 0x80 /* fail the match if $& would start before + the start pos (so s/.\G// would fail + on second iteration */ #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN) # define ReREFCNT_inc(re) \ @@ -594,7 +610,8 @@ typedef struct { STRLEN suboffset; /* saved suboffset field from rex */ STRLEN subcoffset; /* saved subcoffset field from rex */ MAGIC *pos_magic; /* pos() magic attached to $_ */ - I32 pos; /* the original value of pos() in pos_magic */ + SSize_t pos; /* the original value of pos() in pos_magic */ + U8 pos_flags; /* flags to be restored; currently only MGf_BYTES*/ } regmatch_info_aux_eval; @@ -831,11 +848,5 @@ typedef struct regmatch_slab { /* - * Local variables: - * c-indentation-style: bsd - * c-basic-offset: 4 - * indent-tabs-mode: nil - * End: - * * ex: set ts=8 sts=4 sw=4 et: */