X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regexec.c;h=d3f2065cbc09e53f09a53fc016a3d234336b84a5;hb=22d4bb9ccb8701e68f9243547d7e3a3c55f70908;hp=002b66ac85ae93dcd4f6bfffb835261e4dcfe546;hpb=4b19af017623bfa3bb72bb164598a517f586e0d3;p=p5sagit%2Fp5-mst-13.2.git

diff --git a/regexec.c b/regexec.c
index 002b66a..d3f2065 100644
--- a/regexec.c
+++ b/regexec.c
@@ -357,17 +357,18 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos,
 		     || ( (prog->reganch & ROPT_ANCH_BOL)
 			  && !PL_multiline ) );	/* Check after \n? */
 
-	if ((prog->check_offset_min == prog->check_offset_max) && !ml_anch) {
+	if (!ml_anch) {
+	  if ( !(prog->reganch & ROPT_ANCH_GPOS) /* Checked by the caller */
+	       /* SvCUR is not set on references: SvRV and SvPVX overlap */
+	       && sv && !SvROK(sv)
+	       && (strpos + SvCUR(sv) != strend)) {
+	      DEBUG_r(PerlIO_printf(Perl_debug_log, "Not at start...\n"));
+	      goto fail;
+	  }
+	  if (prog->check_offset_min == prog->check_offset_max) {
 	    /* Substring at constant offset from beg-of-str... */
 	    I32 slen;
 
-	    if ( !(prog->reganch & ROPT_ANCH_GPOS) /* Checked by the caller */
-		 /* SvCUR is not set on references: SvRV and SvPVX overlap */
-		 && sv && !SvROK(sv)
-		 && (strpos + SvCUR(sv) != strend)) {
-		DEBUG_r(PerlIO_printf(Perl_debug_log, "Not at start...\n"));
-		goto fail;
-	    }
 	    PL_regeol = strend;			/* Used in HOP() */
 	    s = HOPc(strpos, prog->check_offset_min);
 	    if (SvTAIL(check)) {
@@ -393,6 +394,7 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos,
 			 && memNE(SvPVX(check), s, slen)))
 		goto report_neq;
 	    goto success_at_start;
+	  }
 	}
 	/* Match is anchored, but substr is not anchored wrt beg-of-str. */
 	s = strpos;
@@ -690,6 +692,7 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos,
 	    SvREFCNT_dec(prog->check_substr);
 	    prog->check_substr = Nullsv;	/* disable */
 	    prog->float_substr = Nullsv;	/* clear */
+	    check = Nullsv;			/* abort */
 	    s = strpos;
 	    /* XXXX This is a remnant of the old implementation.  It
 	            looks wasteful, since now INTUIT can use many
@@ -752,6 +755,8 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos,
 					       "Could not match STCLASS...\n") );
 			goto fail;
 		    }
+		    if (!check)
+			goto giveup;
 		    DEBUG_r( PerlIO_printf(Perl_debug_log,
 				"Looking for %s substr starting at offset %ld...\n",
 				 what, (long)(s + start_shift - i_strpos)) );
@@ -762,6 +767,8 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos,
 		    goto retry_floating_check;
 		/* Recheck anchored substring, but not floating... */
 		s = check_at; 
+		if (!check)
+		    goto giveup;
 		DEBUG_r( PerlIO_printf(Perl_debug_log,
 			  "Looking for anchored substr starting at offset %ld...\n",
 			  (long)(other_last - i_strpos)) );
@@ -771,6 +778,8 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos,
                current position only: */
 	    if (ml_anch) {
 		s = t = t + 1;
+		if (!check)
+		    goto giveup;
 		DEBUG_r( PerlIO_printf(Perl_debug_log,
 			  "Looking for /%s^%s/m starting at offset %ld...\n",
 			  PL_colors[0],PL_colors[1], (long)(t - i_strpos)) );
@@ -792,8 +801,10 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos,
 		     PerlIO_printf(Perl_debug_log, 
 			"Does not contradict STCLASS...\n") );
     }
-    DEBUG_r(PerlIO_printf(Perl_debug_log, "%sGuessed:%s match at offset %ld\n",
-			  PL_colors[4], PL_colors[5], (long)(s - i_strpos)) );
+  giveup:
+    DEBUG_r(PerlIO_printf(Perl_debug_log, "%s%s:%s match at offset %ld\n",
+			  PL_colors[4], (check ? "Guessed" : "Giving up"),
+			  PL_colors[5], (long)(s - i_strpos)) );
     return s;
 
   fail_finish:				/* Substring not found */
@@ -905,7 +916,7 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, char *strend, char *sta
 	    PL_reg_flags |= RF_tainted;
 	    /* FALL THROUGH */
 	case BOUNDUTF8:
-	    tmp = (I32)(s != startpos) ? utf8_to_uv(reghop((U8*)s, -1), 0) : '\n';
+	    tmp = (I32)(s != startpos) ? utf8_to_uv_chk(reghop((U8*)s, -1), 0, 0) : '\n';
 	    tmp = ((OP(c) == BOUNDUTF8 ? isALNUM_uni(tmp) : isALNUM_LC_uni(tmp)) != 0);
 	    while (s < strend) {
 		if (tmp == !(OP(c) == BOUNDUTF8 ?
@@ -941,7 +952,7 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, char *strend, char *sta
 	    PL_reg_flags |= RF_tainted;
 	    /* FALL THROUGH */
 	case NBOUNDUTF8:
-	    tmp = (I32)(s != startpos) ? utf8_to_uv(reghop((U8*)s, -1), 0) : '\n';
+	    tmp = (I32)(s != startpos) ? utf8_to_uv_chk(reghop((U8*)s, -1), 0, 0) : '\n';
 	    tmp = ((OP(c) == NBOUNDUTF8 ? isALNUM_uni(tmp) : isALNUM_LC_uni(tmp)) != 0);
 	    while (s < strend) {
 		if (tmp == !(OP(c) == NBOUNDUTF8 ?
@@ -1986,7 +1997,7 @@ S_regmatch(pTHX_ regnode *prog)
 		while (s < e) {
 		    if (l >= PL_regeol)
 			sayNO;
-		    if (utf8_to_uv((U8*)s, 0) != (c1 ?
+		    if (utf8_to_uv_chk((U8*)s, 0, 0) != (c1 ?
 						  toLOWER_utf8((U8*)l) :
 						  toLOWER_LC_utf8((U8*)l)))
 		    {
@@ -2124,7 +2135,7 @@ S_regmatch(pTHX_ regnode *prog)
 	case NBOUNDUTF8:
 	    /* was last char in word? */
 	    ln = (locinput != PL_regbol)
-		? utf8_to_uv(reghop((U8*)locinput, -1), 0) : PL_regprev;
+		? utf8_to_uv_chk(reghop((U8*)locinput, -1), 0, 0) : PL_regprev;
 	    if (OP(scan) == BOUNDUTF8 || OP(scan) == NBOUNDUTF8) {
 		ln = isALNUM_uni(ln);
 		n = swash_fetch(PL_utf8_alnum, (U8*)locinput);
@@ -3616,7 +3627,11 @@ S_reginclass(pTHX_ register regnode *p, register I32 c)
 	    (ANYOF_CLASS_TEST(p, ANYOF_UPPER)   &&  isUPPER_LC(c))  ||
 	    (ANYOF_CLASS_TEST(p, ANYOF_NUPPER)  && !isUPPER_LC(c))  ||
 	    (ANYOF_CLASS_TEST(p, ANYOF_XDIGIT)  &&  isXDIGIT(c))    ||
-	    (ANYOF_CLASS_TEST(p, ANYOF_NXDIGIT) && !isXDIGIT(c))
+	    (ANYOF_CLASS_TEST(p, ANYOF_NXDIGIT) && !isXDIGIT(c))    ||
+	    (ANYOF_CLASS_TEST(p, ANYOF_PSXSPC)  &&  isPSXSPC(c))    ||
+	    (ANYOF_CLASS_TEST(p, ANYOF_NPSXSPC) && !isPSXSPC(c))    ||
+	    (ANYOF_CLASS_TEST(p, ANYOF_BLANK)   &&  isBLANK(c))    ||
+	    (ANYOF_CLASS_TEST(p, ANYOF_NBLANK)  && !isBLANK(c))
 	    ) /* How's that for a conditional? */
 	{
 	    match = TRUE;
@@ -3722,4 +3737,3 @@ restore_pos(pTHXo_ void *arg)
 	PL_curpm = PL_reg_oldcurpm;
     }	
 }
-