X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regexec.c;h=6c0923fc8b89c4afcaff7ebc132ea21c21fbba4d;hb=aaf9c2b26697492a8234a7efe890beef8868ea9b;hp=e64846fdd047943416610af0d85678cb27c73e30;hpb=14a49a2428492a7a120f0254ff8085e99252f641;p=p5sagit%2Fp5-mst-13.2.git diff --git a/regexec.c b/regexec.c index e64846f..6c0923f 100644 --- a/regexec.c +++ b/regexec.c @@ -2,7 +2,11 @@ */ /* - * "One Ring to rule them all, One Ring to find them..." + * One Ring to rule them all, One Ring to find them + & + * [p.v of _The Lord of the Rings_, opening poem] + * [p.50 of _The Lord of the Rings_, I/iii: "The Shadow of the Past"] + * [p.254 of _The Lord of the Rings_, II/ii: "The Council of Elrond"] */ /* This file contains functions for executing a regular expression. See @@ -56,7 +60,8 @@ **** Alterations to Henry's code are... **** **** Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, - **** 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007 by Larry Wall and others + **** 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008 + **** 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. @@ -237,9 +242,10 @@ S_regcppop(pTHX_ const regexp *rex) dVAR; U32 i; char *input; - GET_RE_DEBUG_FLAGS_DECL; + PERL_ARGS_ASSERT_REGCPPOP; + /* Pop REGCP_OTHER_ELEMS before the parentheses loop starts. */ i = SSPOPINT; assert(i == SAVEt_REGCONTEXT); /* Check that the magic cookie is there. */ @@ -314,6 +320,8 @@ Perl_pregexec(pTHX_ REGEXP * const prog, char* stringarg, register char *strend, /* minend: end of match must be >=minend after stringarg. */ /* nosave: For optimizations. */ { + PERL_ARGS_ASSERT_PREGEXEC; + return regexec_flags(prog, stringarg, strend, strbeg, minend, screamer, NULL, nosave ? 0 : REXEC_COPY_STR); @@ -392,9 +400,10 @@ Perl_re_intuit_start(pTHX_ REGEXP * const rx, SV *sv, char *strpos, #ifdef DEBUGGING const char * const i_strpos = strpos; #endif - GET_RE_DEBUG_FLAGS_DECL; + PERL_ARGS_ASSERT_RE_INTUIT_START; + RX_MATCH_UTF8_set(rx,do_utf8); if (RX_UTF8(rx)) { @@ -1181,6 +1190,8 @@ S_find_byclass(pTHX_ regexp * prog, const regnode *c, char *s, register I32 tmp = 1; /* Scratch variable? */ register const bool do_utf8 = PL_reg_match_utf8; RXi_GET_DECL(prog,progi); + + PERL_ARGS_ASSERT_FIND_BYCLASS; /* We know what class it must start with. */ switch (OP(c)) { @@ -1509,7 +1520,7 @@ S_find_byclass(pTHX_ regexp * prog, const regnode *c, char *s, = (reg_ac_data*)progi->data->data[ ARG( c ) ]; reg_trie_data *trie = (reg_trie_data*)progi->data->data[ aho->trie ]; - HV *widecharmap = (HV*) progi->data->data[ aho->trie + 1 ]; + HV *widecharmap = MUTABLE_HV(progi->data->data[ aho->trie + 1 ]); const char *last_start = strend - trie->minlen; #ifdef DEBUGGING @@ -1720,9 +1731,12 @@ S_find_byclass(pTHX_ regexp * prog, const regnode *c, char *s, } static void -S_swap_match_buff (pTHX_ regexp *prog) { +S_swap_match_buff (pTHX_ regexp *prog) +{ regexp_paren_pair *t; + PERL_ARGS_ASSERT_SWAP_MATCH_BUFF; + if (!prog->swap) { /* We have to be careful. If the previous successful match was from this regex we don't want a subsequent paritally @@ -1768,9 +1782,9 @@ Perl_regexec_flags(pTHX_ REGEXP * const rx, char *stringarg, register char *stre RXi_GET_DECL(prog,progi); regmatch_info reginfo; /* create some info to pass to regtry etc */ bool swap_on_fail = 0; - GET_RE_DEBUG_FLAGS_DECL; + PERL_ARGS_ASSERT_REGEXEC_FLAGS; PERL_UNUSED_ARG(data); /* Be paranoid... */ @@ -2211,6 +2225,9 @@ S_regtry(pTHX_ regmatch_info *reginfo, char **startpos) regexp *const prog = (struct regexp *)SvANY(rx); RXi_GET_DECL(prog,progi); GET_RE_DEBUG_FLAGS_DECL; + + PERL_ARGS_ASSERT_REGTRY; + reginfo->cutpoint=NULL; if ((prog->extflags & RXf_EVAL_SEEN) && !PL_reg_eval_set) { @@ -2233,7 +2250,7 @@ S_regtry(pTHX_ regmatch_info *reginfo, char **startpos) /* Make $_ available to executed code. */ if (reginfo->sv != DEFSV) { SAVE_DEFSV; - DEFSV = reginfo->sv; + DEFSV_set(reginfo->sv); } if (!(SvTYPE(reginfo->sv) >= SVt_PVMG && SvMAGIC(reginfo->sv) @@ -2566,11 +2583,14 @@ S_debug_start_match(pTHX_ const REGEXP *prog, const bool do_utf8, const char *start, const char *end, const char *blurb) { const bool utf8_pat = RX_UTF8(prog) ? 1 : 0; + + PERL_ARGS_ASSERT_DEBUG_START_MATCH; + if (!PL_colorset) reginitcolors(); { RE_PV_QUOTED_DECL(s0, utf8_pat, PERL_DEBUG_PAD_ZERO(0), - RX_PRECOMP(prog), RX_PRELEN(prog), 60); + RX_PRECOMP_const(prog), RX_PRELEN(prog), 60); RE_PV_QUOTED_DECL(s1, do_utf8, PERL_DEBUG_PAD_ZERO(1), start, end - start, 60); @@ -2609,6 +2629,8 @@ S_dump_exec_pos(pTHX_ const char *locinput, ? (5 + taill) - l : locinput - loc_bostr; int pref0_len; + PERL_ARGS_ASSERT_DUMP_EXEC_POS; + while (do_utf8 && UTF8_IS_CONTINUATION(*(U8*)(locinput - pref_len))) pref_len++; pref0_len = pref_len - (locinput - loc_reg_starttry); @@ -2659,11 +2681,15 @@ S_dump_exec_pos(pTHX_ const char *locinput, * or 0 if non of the buffers matched. */ STATIC I32 -S_reg_check_named_buff_matched(pTHX_ const regexp *rex, const regnode *scan) { +S_reg_check_named_buff_matched(pTHX_ const regexp *rex, const regnode *scan) +{ I32 n; RXi_GET_DECL(rex,rexi); - SV *sv_dat=(SV*)rexi->data->data[ ARG( scan ) ]; + SV *sv_dat= MUTABLE_SV(rexi->data->data[ ARG( scan ) ]); I32 *nums=(I32*)SvPVX(sv_dat); + + PERL_ARGS_ASSERT_REG_CHECK_NAMED_BUFF_MATCHED; + for ( n=0; n= nums[n] && PL_regoffs[nums[n]].end != -1) @@ -2707,16 +2733,12 @@ S_regmatch(pTHX_ regmatch_info *reginfo, regnode *prog) dVAR; register const bool do_utf8 = PL_reg_match_utf8; const U32 uniflags = UTF8_ALLOW_DEFAULT; - REGEXP *rex_sv = reginfo->prog; regexp *rex = (struct regexp *)SvANY(rex_sv); RXi_GET_DECL(rex,rexi); - I32 oldsave; - /* the current state. This is a cached copy of PL_regmatch_state */ register regmatch_state *st; - /* cache heavy used fields of st in registers */ register regnode *scan; register regnode *next; @@ -2731,13 +2753,11 @@ S_regmatch(pTHX_ regmatch_info *reginfo, regnode *prog) const U32 max_nochange_depth = (3 * rex->nparens > MAX_RECURSE_EVAL_NOCHANGE_DEPTH) ? 3 * rex->nparens : MAX_RECURSE_EVAL_NOCHANGE_DEPTH; - regmatch_state *yes_state = NULL; /* state to pop to on success of subpattern */ /* mark_state piggy backs on the yes_state logic so that when we unwind the stack on success we can update the mark_state as we go */ regmatch_state *mark_state = NULL; /* last mark state we have seen */ - regmatch_state *cur_eval = NULL; /* most recent EVAL_AB state */ struct regmatch_state *cur_curlyx = NULL; /* most recent curlyx */ U32 state_num; @@ -2750,10 +2770,7 @@ S_regmatch(pTHX_ regmatch_info *reginfo, regnode *prog) during a successfull match */ U32 lastopen = 0; /* last open we saw */ bool has_cutgroup = RX_HAS_CUTGROUP(rex) ? 1 : 0; - SV* const oreplsv = GvSV(PL_replgv); - - /* these three flags are set by various ops to signal information to * the very next op. They have a useful lifetime of exactly one loop * iteration, and are not preserved or restored by state pushes/pops @@ -2768,11 +2785,12 @@ S_regmatch(pTHX_ regmatch_info *reginfo, regnode *prog) false: plain (?=foo) true: used as a condition: (?(?=foo)) */ - #ifdef DEBUGGING GET_RE_DEBUG_FLAGS_DECL; #endif + PERL_ARGS_ASSERT_REGMATCH; + DEBUG_OPTIMISE_r( DEBUG_EXECUTE_r({ PerlIO_printf(Perl_debug_log,"regmatch start\n"); })); @@ -2928,7 +2946,7 @@ S_regmatch(pTHX_ regmatch_info *reginfo, regnode *prog) /* what trie are we using right now */ reg_trie_data * const trie = (reg_trie_data*)rexi->data->data[ ARG( scan ) ]; - HV * widecharmap = (HV *)rexi->data->data[ ARG( scan ) + 1 ]; + HV * widecharmap = MUTABLE_HV(rexi->data->data[ ARG( scan ) + 1 ]); U32 state = trie->startstate; if (trie->bitmap && trie_type != trie_utf8_fold && @@ -3085,7 +3103,7 @@ S_regmatch(pTHX_ regmatch_info *reginfo, regnode *prog) /* only one choice left - just continue */ DEBUG_EXECUTE_r({ AV *const trie_words - = (AV *) rexi->data->data[ARG(ST.me)+TRIE_WORDS_OFFSET]; + = MUTABLE_AV(rexi->data->data[ARG(ST.me)+TRIE_WORDS_OFFSET]); SV ** const tmp = av_fetch( trie_words, ST.accept_buff[ 0 ].wordnum-1, 0 ); SV *sv= tmp ? sv_newmortal() : NULL; @@ -3167,7 +3185,7 @@ S_regmatch(pTHX_ regmatch_info *reginfo, regnode *prog) DEBUG_EXECUTE_r({ AV *const trie_words - = (AV *) rexi->data->data[ARG(ST.me)+TRIE_WORDS_OFFSET]; + = MUTABLE_AV(rexi->data->data[ARG(ST.me)+TRIE_WORDS_OFFSET]); SV ** const tmp = av_fetch( trie_words, ST.accept_buff[ best ].wordnum - 1, 0 ); regnode *nextop=(!ST.jump || !ST.jump[ST.accept_buff[best].wordnum]) ? @@ -3783,7 +3801,7 @@ S_regmatch(pTHX_ regmatch_info *reginfo, regnode *prog) /* This isn't a first class regexp. Instead, it's caching a regexp onto an existing, Perl visible scalar. */ - sv_magic(ret, (SV*) rx, PERL_MAGIC_qr, 0, 0); + sv_magic(ret, MUTABLE_SV(rx), PERL_MAGIC_qr, 0, 0); } PL_regsize = osize; } @@ -4329,7 +4347,7 @@ NULL case CUTGROUP: PL_reginput = locinput; sv_yes_mark = st->u.mark.mark_name = scan->flags ? NULL : - (SV*)rexi->data->data[ ARG( scan ) ]; + MUTABLE_SV(rexi->data->data[ ARG( scan ) ]); PUSH_STATE_GOTO(CUTGROUP_next,next); /* NOTREACHED */ case CUTGROUP_next_fail: @@ -4965,6 +4983,8 @@ NULL do_ifmatch: ST.me = scan; ST.logical = logical; + logical = 0; /* XXX: reset state of logical once it has been saved into ST */ + /* execute body of (?...A) */ PUSH_YES_STATE_GOTO(IFMATCH_A, NEXTOPER(NEXTOPER(scan))); /* NOTREACHED */ @@ -5004,7 +5024,7 @@ NULL case PRUNE: PL_reginput = locinput; if (!scan->flags) - sv_yes_mark = sv_commit = (SV*)rexi->data->data[ ARG( scan ) ]; + sv_yes_mark = sv_commit = MUTABLE_SV(rexi->data->data[ ARG( scan ) ]); PUSH_STATE_GOTO(COMMIT_next,next); /* NOTREACHED */ case COMMIT_next_fail: @@ -5018,7 +5038,7 @@ NULL case MARKPOINT: ST.prev_mark = mark_state; ST.mark_name = sv_commit = sv_yes_mark - = (SV*)rexi->data->data[ ARG( scan ) ]; + = MUTABLE_SV(rexi->data->data[ ARG( scan ) ]); mark_state = st; ST.mark_loc = PL_reginput = locinput; PUSH_YES_STATE_GOTO(MARKPOINT_next,next); @@ -5059,7 +5079,7 @@ NULL otherwise do nothing. Meaning we need to scan */ regmatch_state *cur = mark_state; - SV *find = (SV*)rexi->data->data[ ARG( scan ) ]; + SV *find = MUTABLE_SV(rexi->data->data[ ARG( scan ) ]); while (cur) { if ( sv_eq( cur->u.mark.mark_name, @@ -5348,6 +5368,8 @@ S_regrepeat(pTHX_ const regexp *prog, const regnode *p, I32 max, int depth) PERL_UNUSED_ARG(depth); #endif + PERL_ARGS_ASSERT_REGREPEAT; + scan = PL_reginput; if (max == REG_INFTY) max = I32_MAX; @@ -5663,12 +5685,14 @@ Perl_regclass_swash(pTHX_ const regexp *prog, register const regnode* node, bool RXi_GET_DECL(prog,progi); const struct reg_data * const data = prog ? progi->data : NULL; + PERL_ARGS_ASSERT_REGCLASS_SWASH; + if (data && data->count) { const U32 n = ARG(node); if (data->what[n] == 's') { - SV * const rv = (SV*)data->data[n]; - AV * const av = (AV*)SvRV((SV*)rv); + SV * const rv = MUTABLE_SV(data->data[n]); + AV * const av = MUTABLE_AV(SvRV(rv)); SV **const ary = AvARRAY(av); SV **a, **b; @@ -5719,6 +5743,8 @@ S_reginclass(pTHX_ const regexp *prog, register const regnode *n, register const STRLEN len = 0; STRLEN plen; + PERL_ARGS_ASSERT_REGINCLASS; + if (do_utf8 && !UTF8_IS_INVARIANT(c)) { c = utf8n_to_uvchr(p, UTF8_MAXBYTES, &len, (UTF8_ALLOW_DEFAULT & UTF8_ALLOW_ANYUV) | UTF8_CHECK_ONLY); @@ -5836,6 +5862,9 @@ STATIC U8 * S_reghop3(U8 *s, I32 off, const U8* lim) { dVAR; + + PERL_ARGS_ASSERT_REGHOP3; + if (off >= 0) { while (off-- && s < lim) { /* XXX could check well-formedness here */ @@ -5864,6 +5893,9 @@ STATIC U8 * S_reghop4(U8 *s, I32 off, const U8* llim, const U8* rlim) { dVAR; + + PERL_ARGS_ASSERT_REGHOP4; + if (off >= 0) { while (off-- && s < rlim) { /* XXX could check well-formedness here */ @@ -5888,6 +5920,9 @@ STATIC U8 * S_reghopmaybe3(U8* s, I32 off, const U8* lim) { dVAR; + + PERL_ARGS_ASSERT_REGHOPMAYBE3; + if (off >= 0) { while (off-- && s < lim) { /* XXX could check well-formedness here */ @@ -5935,6 +5970,9 @@ STATIC void S_to_utf8_substr(pTHX_ register regexp *prog) { int i = 1; + + PERL_ARGS_ASSERT_TO_UTF8_SUBSTR; + do { if (prog->substrs->data[i].substr && !prog->substrs->data[i].utf8_substr) { @@ -5964,6 +6002,9 @@ S_to_byte_substr(pTHX_ register regexp *prog) { dVAR; int i = 1; + + PERL_ARGS_ASSERT_TO_BYTE_SUBSTR; + do { if (prog->substrs->data[i].utf8_substr && !prog->substrs->data[i].substr) {