integrate cfgperl contents into mainline

p4raw-id: //depot/perl@6400

9 files changed, 47 insertions, 50 deletions

diff --git a/Configure b/Configure
index 35b3878b99..bf3e857207 100755
--- a/Configure
+++ b/Configure
@@ -20,7 +20,7 @@
 
 # $Id: Head.U,v 3.0.1.9 1997/02/28 15:02:09 ram Exp $
 #
-# Generated on Wed Jul 12 07:59:35 EET DST 2000 [metaconfig 3.0 PL70]
+# Generated on Thu Jul 13 16:16:41 EET DST 2000 [metaconfig 3.0 PL70]
 # (with additional metaconfig patches by perlbug@perl.com)
 
 cat >/tmp/c1$$ <<EOF
diff --git a/embed.pl b/embed.pl
index 08e305ba33..0b4c5450ae 100755
--- a/embed.pl
+++ b/embed.pl
@@ -2474,7 +2474,7 @@ s	|char*	|scan_trans	|char *start
 s	|char*	|scan_word	|char *s|char *dest|STRLEN destlen \
 				|int allow_package|STRLEN *slp
 s	|char*	|skipspace	|char *s
-s	|char*	|swallow_bom	|char *s
+s	|char*	|swallow_bom	|U8 *s
 s	|void	|checkcomma	|char *s|char *name|char *what
 s	|void	|force_ident	|char *s|int kind
 s	|void	|incline	|char *s
diff --git a/global.sym b/global.sym
index 719e50a2f4..6ee8fc4879 100644
--- a/global.sym
+++ b/global.sym
@@ -21,6 +21,7 @@ Perl_get_context
 Perl_set_context
 Perl_amagic_call
 Perl_Gv_AMupdate
+Perl_apply_attrs_string
 Perl_avhv_delete_ent
 Perl_avhv_exists_ent
 Perl_avhv_fetch_ent
@@ -185,6 +186,7 @@ Perl_to_uni_upper_lc
 Perl_to_uni_title_lc
 Perl_to_uni_lower_lc
 Perl_is_utf8_char
+Perl_is_utf8_string
 Perl_is_utf8_alnum
 Perl_is_utf8_alnumc
 Perl_is_utf8_idfirst
@@ -460,6 +462,8 @@ Perl_utf16_to_utf8
 Perl_utf16_to_utf8_reversed
 Perl_utf8_distance
 Perl_utf8_hop
+Perl_utf8_to_bytes
+Perl_bytes_to_utf8
 Perl_utf8_to_uv
 Perl_uv_to_utf8
 Perl_warn
@@ -543,3 +547,4 @@ Perl_ptr_table_fetch
 Perl_ptr_table_store
 Perl_ptr_table_split
 Perl_sys_intern_clear
+Perl_sys_intern_init
diff --git a/proto.h b/proto.h
index bd222fe835..358f5300fa 100644
--- a/proto.h
+++ b/proto.h
@@ -1228,7 +1228,7 @@ STATIC char*	S_scan_subst(pTHX_ char *start);
 STATIC char*	S_scan_trans(pTHX_ char *start);
 STATIC char*	S_scan_word(pTHX_ char *s, char *dest, STRLEN destlen, int allow_package, STRLEN *slp);
 STATIC char*	S_skipspace(pTHX_ char *s);
-STATIC char*	S_swallow_bom(pTHX_ char *s);
+STATIC char*	S_swallow_bom(pTHX_ U8 *s);
 STATIC void	S_checkcomma(pTHX_ char *s, char *name, char *what);
 STATIC void	S_force_ident(pTHX_ char *s, int kind);
 STATIC void	S_incline(pTHX_ char *s);
diff --git a/regcomp.c b/regcomp.c
index 89b3e53bab..db099ad218 100644
--- a/regcomp.c
+++ b/regcomp.c
@@ -901,6 +901,9 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da
 				sv_catsv(data->last_found, last_str);
 				data->last_end += l * (mincount - 1);
 			    }
+			} else {
+			    /* start offset must point into the last copy */
+			    data->last_start_min += minnext * (mincount - 1);
 			}
 		    }
 		    /* It is counted once already... */
diff --git a/regexec.c b/regexec.c
index 6bddf2dc32..7472d09735 100644
--- a/regexec.c
+++ b/regexec.c
@@ -221,6 +221,22 @@ typedef struct re_cc_state
 
 #define regcpblow(cp) LEAVE_SCOPE(cp)
 
+#define TRYPAREN(paren, n, input) {				\
+    if (paren) {						\
+	if (n) {						\
+	    PL_regstartp[paren] = HOPc(input, -1) - PL_bostr;	\
+	    PL_regendp[paren] = input - PL_bostr;		\
+	}							\
+	else							\
+	    PL_regendp[paren] = -1;				\
+    }								\
+    if (regmatch(next))						\
+	sayYES;							\
+    if (paren && n)						\
+	PL_regendp[paren] = -1;					\
+}
+
+
 /*
  * pregexec and friends
  */
@@ -3001,8 +3017,6 @@ S_regmatch(pTHX_ regnode *prog)
 	    else
 		c1 = c2 = -1000;
 	    PL_reginput = locinput;
-	    if (paren)
-		PL_regendp[paren] = -1;
 	    if (minmod) {
 		CHECKPOINT lastcp;
 		minmod = 0;
@@ -3037,16 +3051,7 @@ S_regmatch(pTHX_ regnode *prog)
 				sayNO;
 			}
 			/* PL_reginput == locinput now */
-			if (paren) {
-			    if (ln) {
-				PL_regstartp[paren] = HOPc(locinput, -1) - PL_bostr;
-				PL_regendp[paren] = locinput - PL_bostr;
-			    }
-			    else
-				PL_regendp[paren] = -1;
-			}
-			if (regmatch(next))
-			    sayYES;
+			TRYPAREN(paren, ln, locinput);
 			PL_reginput = locinput;	/* Could be reset... */
 			REGCP_UNWIND;
 			/* Couldn't or didn't -- move forward. */
@@ -3060,16 +3065,7 @@ S_regmatch(pTHX_ regnode *prog)
 			UCHARAT(PL_reginput) == c1 ||
 			UCHARAT(PL_reginput) == c2)
 		    {
-			if (paren) {
-			    if (n) {
-				PL_regstartp[paren] = HOPc(PL_reginput, -1) - PL_bostr;
-				PL_regendp[paren] = PL_reginput - PL_bostr;
-			    }
-			    else
-				PL_regendp[paren] = -1;
-			}
-			if (regmatch(next))
-			    sayYES;
+			TRYPAREN(paren, n, PL_reginput);
 			REGCP_UNWIND;
 		    }
 		    /* Couldn't or didn't -- move forward. */
@@ -3103,16 +3099,7 @@ S_regmatch(pTHX_ regnode *prog)
 			    UCHARAT(PL_reginput) == c1 ||
 			    UCHARAT(PL_reginput) == c2)
 			    {
-				if (paren && n) {
-				    if (n) {
-					PL_regstartp[paren] = HOPc(PL_reginput, -1) - PL_bostr;
-					PL_regendp[paren] = PL_reginput - PL_bostr;
-				    }
-				    else
-					PL_regendp[paren] = -1;
-				}
-				if (regmatch(next))
-				    sayYES;
+				TRYPAREN(paren, n, PL_reginput);
 				REGCP_UNWIND;
 			    }
 			/* Couldn't or didn't -- back up. */
@@ -3127,8 +3114,7 @@ S_regmatch(pTHX_ regnode *prog)
 			    UCHARAT(PL_reginput) == c1 ||
 			    UCHARAT(PL_reginput) == c2)
 			    {
-				if (regmatch(next))
-				    sayYES;
+				TRYPAREN(paren, n, PL_reginput);
 				REGCP_UNWIND;
 			    }
 			/* Couldn't or didn't -- back up. */
diff --git a/t/op/re_tests b/t/op/re_tests
index 263bded7c9..b3137c737d 100644
--- a/t/op/re_tests
+++ b/t/op/re_tests
@@ -751,5 +751,10 @@ tt+$	xxxtt	y	-	-
 '^\S\s+aa$'m	\nx aa	y	-	-
 (^|a)b	ab	y	-	-
 ^([ab]*?)(b)?(c)$	abac	y	-$2-	--
+(\w)?(abc)\1b	abcab	n	-	-
+^(?:.,){2}c	a,b,c	y	-	-
+^(.,){2}c	a,b,c	y	$1	b,
+^(?:[^,]*,){2}c	a,b,c	y	-	-
+^([^,]*,){2}c	a,b,c	y	$1	b,
 (?i)		y	-	-
 '(?!\A)x'm	a\nxb\n	y	-	-
diff --git a/toke.c b/toke.c
index d9b42a8905..672acb71bf 100644
--- a/toke.c
+++ b/toke.c
@@ -7407,21 +7407,21 @@ Perl_yyerror(pTHX_ char *s)
 }
 
 STATIC char*
-S_swallow_bom(pTHX_ char *s)
+S_swallow_bom(pTHX_ U8 *s)
 {
     STRLEN slen;
     slen = SvCUR(PL_linestr);
     switch (*s) {
-    case -1:       
-	if ((s[1] & 255) == 254) { 
+    case 0xFF:       
+	if (s[1] == 0xFE) { 
 	    /* UTF-16 little-endian */
 #ifndef PERL_NO_UTF16_FILTER
 	    U8 *news;
 #endif
-	    s += 2;
-	    if (*s == 0 && s[1] == 0)  /* UTF-32 little-endian */
+	    if (s[2] == 0 && s[3] == 0)  /* UTF-32 little-endian */
 		Perl_croak(aTHX_ "Unsupported script encoding");
 #ifndef PERL_NO_UTF16_FILTER
+	    s += 2;
 	    filter_add(S_utf16rev_textfilter, NULL);
 	    New(898, news, (PL_bufend - s) * 3 / 2 + 1, U8);
 	    PL_bufend = utf16_to_utf8((U16*)s, news, PL_bufend - s);
@@ -7431,9 +7431,8 @@ S_swallow_bom(pTHX_ char *s)
 #endif
 	}
 	break;
-
-    case -2:
-	if ((s[1] & 255) == 255) {   /* UTF-16 big-endian */
+    case 0xFE:
+	if (s[1] == 0xFF) {   /* UTF-16 big-endian */
 #ifndef PERL_NO_UTF16_FILTER
 	    U8 *news;
 	    filter_add(S_utf16_textfilter, NULL);
@@ -7445,15 +7444,14 @@ S_swallow_bom(pTHX_ char *s)
 #endif
 	}
 	break;
-
-    case -17:
-	if (slen > 2 && (s[1] & 255) == 187 && (s[2] & 255) == 191) {
+    case 0xEF:
+	if (slen > 2 && s[1] == 0xBB && s[2] == 0xBF) {
 	    s += 3;                      /* UTF-8 */
 	}
 	break;
     case 0:
 	if (slen > 3 && s[1] == 0 &&  /* UTF-32 big-endian */
-	    s[2] & 255 == 254 && s[3] & 255 == 255)
+	    s[2] == 0xFE && s[3] == 0xFF)
 	{
 	    Perl_croak(aTHX_ "Unsupported script encoding");
 	}
diff --git a/vms/subconfigure.com b/vms/subconfigure.com
index ce291f8cba..b986820038 100644
--- a/vms/subconfigure.com
+++ b/vms/subconfigure.com
@@ -72,7 +72,7 @@ $! ##ADD NEW CONSTANTS HERE##
 $ perl_i_prot="undef"
 $ perl_d_getespwnam="undef"
 $ perl_d_getprpwnam="undef"
-$ perl_d_isnan= = "define"
+$ perl_d_isnan= "define"
 $ perl_sizesize = "4"
 $ perl_shmattype = ""
 $ perl_mmaptype = ""