X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regcomp.h;h=72f415a0dab6719f80cc5c9c23f90f0e8e292f7c;hb=a2edbdc9c7ad63359455c7aff3db47e75ffe578b;hp=1075080d95385347ea04e001c0c1c07af5a7fde4;hpb=6d5c990f28c2a18264defef032a496e2e3dfb7cd;p=p5sagit%2Fp5-mst-13.2.git diff --git a/regcomp.h b/regcomp.h index 1075080..72f415a 100644 --- a/regcomp.h +++ b/regcomp.h @@ -1,7 +1,7 @@ /* regcomp.h * * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, - * 2000, 2001, 2002, 2003, 2005 by Larry Wall and others + * 2000, 2001, 2002, 2003, 2005, 2006 by Larry Wall and others * * You may distribute under the terms of either the GNU General Public * License or the Artistic License, as specified in the README file. @@ -18,11 +18,16 @@ typedef OP OP_4tree; /* Will be redefined later. */ #define PERL_ENABLE_EXTENDED_TRIE_OPTIMISATION 1 /* Should the optimiser take positive assertions into account? */ -#define PERL_ENABLE_POSITIVE_ASSERTION_STUDY 1 +#define PERL_ENABLE_POSITIVE_ASSERTION_STUDY 0 /* Not for production use: */ #define PERL_ENABLE_EXPERIMENTAL_REGEX_OPTIMISATIONS 0 +/* Activate offsets code - set to if 1 to enable */ +#ifdef DEBUGGING +#define RE_TRACK_PATTERN_OFFSETS +#endif + /* Unless the next line is uncommented it is illegal to combine lazy matching with possessive matching. Frankly it doesn't make much sense to allow it as X*?+ matches nothing, X+?+ matches a single char only, @@ -92,11 +97,28 @@ typedef OP OP_4tree; /* Will be redefined later. */ * special test to reverse the sign of BACK pointers since the offset is * stored negative.] */ -typedef struct regexp_internal { - regexp_paren_ofs *swap; /* Swap copy of *startp / *endp */ - U32 *offsets; /* offset annotations 20001228 MJD - data about mapping the program to the - string*/ + +/* This is the stuff that used to live in regexp.h that was truly + private to the engine itself. It now lives here. */ + +/* swap buffer for paren structs */ +typedef struct regexp_paren_ofs { + I32 *startp; + I32 *endp; +} regexp_paren_ofs; + + typedef struct regexp_internal { + int name_list_idx; /* Optional data index of an array of paren names */ + union { + U32 *offsets; /* offset annotations 20001228 MJD + data about mapping the program to the + string - + offsets[0] is proglen when this is used + */ + U32 proglen; + } u; + + regexp_paren_ofs *swap; /* Swap copy of *startp / *endp */ regnode *regstclass; /* Optional startclass as identified or constructed by the optimiser */ struct reg_data *data; /* Additional miscellaneous data used by the program. @@ -109,6 +131,20 @@ typedef struct regexp_internal { #define RXi_SET(x,y) (x)->pprivate = (void*)(y) #define RXi_GET(x) ((regexp_internal *)((x)->pprivate)) #define RXi_GET_DECL(r,ri) regexp_internal *ri = RXi_GET(r) +/* + * Flags stored in regexp->intflags + * These are used only internally to the regexp engine + * + * See regexp.h for flags used externally to the regexp engine + */ +#define PREGf_SKIP 0x00000001 +#define PREGf_IMPLICIT 0x00000002 /* Converted .* to ^.* */ +#define PREGf_NAUGHTY 0x00000004 /* how exponential is this pattern? */ +#define PREGf_VERBARG_SEEN 0x00000008 +#define PREGf_CUTGROUP_SEEN 0x00000010 + + +/* this is where the old regcomp.h started */ struct regnode_string { U8 str_len; @@ -378,6 +414,7 @@ struct regnode_charclass_class { /* has [[:blah:]] classes */ #define REG_TOP_LEVEL_BRANCHES 0x00000040 #define REG_SEEN_VERBARG 0x00000080 #define REG_SEEN_CUTGROUP 0x00000100 +#define REG_SEEN_RUN_ON_COMMENT 0x00000200 START_EXTERN_C @@ -426,6 +463,8 @@ EXTCONST regexp_engine PL_core_reg_engine = { Perl_re_intuit_start, Perl_re_intuit_string, Perl_regfree_internal, + Perl_reg_numbered_buff_get, + Perl_reg_named_buff_get, #if defined(USE_ITHREADS) Perl_regdupe_internal #endif @@ -460,18 +499,8 @@ struct reg_data { void* data[1]; }; -struct reg_substr_datum { - I32 min_offset; - I32 max_offset; - SV *substr; /* non-utf8 variant */ - SV *utf8_substr; /* utf8 variant */ - I32 end_shift; -}; - -struct reg_substr_data { - struct reg_substr_datum data[3]; /* Actual array */ -}; - +/* Code in S_to_utf8_substr() and S_to_byte_substr() in regexec.c accesses + anchored* and float* via array indexes 0 and 1. */ #define anchored_substr substrs->data[0].substr #define anchored_utf8 substrs->data[0].utf8_substr #define anchored_offset substrs->data[0].min_offset @@ -666,6 +695,7 @@ re.pm, especially to the documentation. #define RE_DEBUG_EXTRA_OFFDEBUG 0x040000 #define RE_DEBUG_EXTRA_STATE 0x080000 #define RE_DEBUG_EXTRA_OPTIMISE 0x100000 +#define RE_DEBUG_EXTRA_BUFFERS 0x400000 /* combined */ #define RE_DEBUG_EXTRA_STACK 0x280000 @@ -703,6 +733,9 @@ re.pm, especially to the documentation. if (re_debug_flags & RE_DEBUG_EXTRA_STATE) x ) #define DEBUG_STACK_r(x) DEBUG_r( \ if (re_debug_flags & RE_DEBUG_EXTRA_STACK) x ) +#define DEBUG_BUFFERS_r(x) DEBUG_r( \ + if (re_debug_flags & RE_DEBUG_EXTRA_BUFFERS) x ) + #define DEBUG_OPTIMISE_MORE_r(x) DEBUG_r( \ if ((RE_DEBUG_EXTRA_OPTIMISE|RE_DEBUG_COMPILE_OPTIMISE) == \ (re_debug_flags & (RE_DEBUG_EXTRA_OPTIMISE|RE_DEBUG_COMPILE_OPTIMISE)) ) x ) @@ -738,22 +771,22 @@ re.pm, especially to the documentation. const char * const rpv = \ pv_pretty((dsv), (pv), (l), (m), \ PL_colors[(c1)],PL_colors[(c2)], \ - ((isuni) ? PERL_PV_ESCAPE_UNI : 0) ); \ + PERL_PV_ESCAPE_RE |((isuni) ? PERL_PV_ESCAPE_UNI : 0) ); \ const int rlen = SvCUR(dsv) #define RE_SV_ESCAPE(rpv,isuni,dsv,sv,m) \ const char * const rpv = \ pv_pretty((dsv), (SvPV_nolen_const(sv)), (SvCUR(sv)), (m), \ PL_colors[(c1)],PL_colors[(c2)], \ - ((isuni) ? PERL_PV_ESCAPE_UNI : 0) ) + PERL_PV_ESCAPE_RE |((isuni) ? PERL_PV_ESCAPE_UNI : 0) ) #define RE_PV_QUOTED_DECL(rpv,isuni,dsv,pv,l,m) \ const char * const rpv = \ pv_pretty((dsv), (pv), (l), (m), \ PL_colors[0], PL_colors[1], \ - ( PERL_PV_PRETTY_QUOTE | PERL_PV_PRETTY_ELIPSES | \ + ( PERL_PV_PRETTY_QUOTE | PERL_PV_ESCAPE_RE | PERL_PV_PRETTY_ELIPSES | \ ((isuni) ? PERL_PV_ESCAPE_UNI : 0)) \ - ) + ) #define RE_SV_DUMPLEN(ItEm) (SvCUR(ItEm) - (SvTAIL(ItEm)!=0)) #define RE_SV_TAIL(ItEm) (SvTAIL(ItEm) ? "$" : "")