diff -urN perl5.004_02/Changes.jp jperl5.004_02/Changes.jp
--- perl5.004_02/Changes.jp	Thu Jan  1 09:00:00 1970
+++ jperl5.004_02/Changes.jp	Fri Aug  8 10:10:23 1997
@@ -0,0 +1,159 @@
+Fri Aug  8 10:08:21 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.004_02.pat:
+
+	* perl5.004_02 対応.
+
+Thu Aug  7 13:00:23 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.004_01_03.pat:
+
+	* perl5.004_01_03 対応.
+
+	* $_ = 'abc0123"; tr/0-9/a/c; が "aaa0123" にならないバグを修正.
+	  bug report ありがとう新井康司さん.
+
+Tue Aug  5 18:28:32 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.004_01_02.pat:
+
+	* perl5.004_01_02 対応.
+
+Thu Jul 10 19:38:40 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.004_01-970710.pat:
+
+	* kanji.c で strcasecmp() を使うのをやめて jstricmp() を使うよう
+	にした.
+
+	* no I18N::Japanese; で $^H &= ~0xf000 になるようにした.
+	qw(string) も独立させた. これで jcode.pl の頭に
+	no I18N::Japanese; を入れれば jperl で使えるようになる.
+
+	* win32/Makefile に kanji.[co] を追加.
+
+Tue Jun 17 13:22:23 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.004_01-970617.pat:
+
+	* perl5.004_01 対応.
+
+Mon May 26 11:01:49 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.004-970526.pat:
+
+	* jcode.pl 対応. 文字クラス中の 16/8 進リテラルは 1 byte 単位で
+	アクセスすることした. いままでは 1 byte でも日本語 1 文字でもな
+	いので, 全く機能してなかった.
+
+	* tr も同じ.
+
+Sun May 18 17:49:11 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.004-970518.pat:
+
+	* perl5.004 Release 対応.
+
+Fri May 16 11:26:35 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.004RC2-970516.pat:
+
+	* perl5.004 Release Candidate 2 対応.
+
+Mon May 12 09:55:02 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.003_99a-970512.pat:
+
+	* perl5.003_99a 対応.
+
+Wed May  7 12:40:05 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.003_99-970507.pat:
+
+	* bit vector を 256+4096 に縮小.
+
+	* bit vector へのポインタの alignment を 4 バイト境界にした.
+
+Tue May  6 12:15:45 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* regclass(), reginclass() を bit vector(8192 bytes)にして高速化.
+
+Sun May  4 17:50:14 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.003_99-970504.pat:
+
+	* \Q, quotemeta() が日本語対応してなかった.
+
+Sat May  3 22:29:35 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* perl.5003_99 対応.
+
+Thu May  1 19:29:21 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.003_98-970501.pat:
+
+	* perl5.003_98 対応.
+
+Sun Apr 27 22:18:08 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.003_97i-970427.pat:
+
+	* regdump() の結果がめちゃくちゃになるので ANYOF_ENDMARK は 99 に
+	した. ついでに ANYOF で範囲を表示するようにした.
+
+	* KANJI MODE じゃないとき, ANY, SANY, ANYOF は SIMPLE になるように
+	した.
+
+Fri Apr 25 12:12:09 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.003_97h-970425.pat:
+
+	* patch 2.2 の -p の仕様が変更になったので, man
+	patch で推奨されてる形式にした.
+
+Wed Apr 23 12:25:16 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.003_97g-970423.pat:
+
+	* regexp.h: ROPT_KANJI と ROPT_IMPLICIT が重なってた.
+
+	* regcomp.h: ANY, SANY, ANYOF を simple[] へ移動. [] 
+        が速くなった(はず).
+
+Wed Apr 21 12:23:56 1997  WATANABE Hirofumi  <watanabe@ase.ptg.sony.co.jp>
+
+	* jperl5.003_97g-970421.pat:
+
+	* perl5.003_97g 対応.
+
+--------------
+
+jperl5.003_97e-970417.pat:
+
+  * perl5.003_97e 対応.
+
+jperl5.003_97d-970415.pat:
+
+  * perl5.003_97d 対応.
+
+jperl5.003_95-970331.pat:
+
+  * perl5.003_93, 94, 95 対応.
+
+  * /[\W]/, /[\S]/ がまったく逆の意味になっていたので修正.
+    誰も使ってないのかな? わたしも使ったことがない. :-)
+    make test で発覚.
+    regcomp.c は大幅に書き換えた.
+
+  * #! のラインに -Lsjis とかあると駄目だったので strncasecmp
+    を使おうと思ったけど util.c の ibcmp() を使った.
+
+  * -w option をつけると tr/// が "Use of uninitialized value"
+    という warning になっていたので修正. t/lib/db-*.t になぜか
+    -w がついてて鬱陶しくなった.
+
+  * toke.c の無意味なインデントの変更はやめた.
+
+jperl5.003_05.patch7_1:
+
+  * jperl4.036 とマージした.
diff -urN perl5.004_02/Configure jperl5.004_02/Configure
--- perl5.004_02/Configure	Thu Aug  7 23:08:44 1997
+++ jperl5.004_02/Configure	Fri Aug  8 10:07:23 1997
@@ -630,6 +630,7 @@
 usrinc=''
 defvoidused=''
 voidflags=''
+d_euc=''
 CONFIG=''
 
 define='define'
@@ -2490,6 +2491,19 @@
 test 0 -eq "$subversion" || $echo $n " subversion $subversion" $c
 echo ".)"
 
+if $test -r ../kanji.h; then
+    echo "You have Japanized version.";
+    dflt=y
+    rp="Do you want to use EUC as the kanji code(y=EUC, n=SJIS)?"
+    . ./myread
+    case "$ans" in
+    [yY]*)
+     d_euc='define';;
+    *)
+     d_euc='undef';;
+    esac
+fi
+    
 : set the prefixup variable, to restore leading tilda escape
 prefixup='case "$prefixexp" in
 "$prefix") ;;
@@ -10468,6 +10482,7 @@
 xlibpth='$xlibpth'
 zcat='$zcat'
 zip='$zip'
+d_euc='$d_euc'
 EOT
 
 : add special variables
diff -urN perl5.004_02/MANIFEST jperl5.004_02/MANIFEST
--- perl5.004_02/MANIFEST	Tue Jul 29 09:01:17 1997
+++ jperl5.004_02/MANIFEST	Fri Aug  8 10:07:23 1997
@@ -303,6 +303,8 @@
 installman		Perl script to install man pages for pods
 installperl		Perl script to do "make install" dirty work
 interp.sym		Interpreter specific symbols to hide in a struct
+kanji.c			Kanji handling routines.
+kanji.h			Header file for the above.
 keywords.h		The keyword numbers
 keywords.pl		Program to write keywords.h
 lib/AnyDBM_File.pm	Perl module to emulate dbmopen
@@ -356,6 +358,7 @@
 lib/Getopt/Long.pm	Fetch command options (GetOptions)
 lib/Getopt/Std.pm	Fetch command options (getopt, getopts)
 lib/I18N/Collate.pm	Routines to do strxfrm-based collation
+lib/I18N/Japanese.pm	Pragma to control Japanese.
 lib/IPC/Open2.pm	Open a two-ended pipe
 lib/IPC/Open3.pm	Open a three-ended pipe!
 lib/Math/BigFloat.pm	An arbitrary precision floating-point arithmetic package
@@ -535,6 +538,7 @@
 pod/Makefile		Make pods into something else
 pod/buildtoc		generate perltoc.pod
 pod/checkpods.PL	Tool to check for common errors in pods
+pod/jperl.pod		Jperl man page
 pod/perl.pod		Top level perl man page
 pod/perlapio.pod	IO API info
 pod/perlbook.pod	Book info
diff -urN perl5.004_02/Makefile.SH jperl5.004_02/Makefile.SH
--- perl5.004_02/Makefile.SH	Thu Aug  7 21:10:53 1997
+++ jperl5.004_02/Makefile.SH	Fri Aug  8 10:07:23 1997
@@ -194,7 +194,7 @@
 
 obj1 = $(mallocobj) gv$(OBJ_EXT) toke$(OBJ_EXT) perly$(OBJ_EXT) op$(OBJ_EXT) regcomp$(OBJ_EXT) dump$(OBJ_EXT) util$(OBJ_EXT) mg$(OBJ_EXT)
 obj2 = hv$(OBJ_EXT) av$(OBJ_EXT) run$(OBJ_EXT) pp_hot$(OBJ_EXT) sv$(OBJ_EXT) pp$(OBJ_EXT) scope$(OBJ_EXT) pp_ctl$(OBJ_EXT) pp_sys$(OBJ_EXT)
-obj3 = doop$(OBJ_EXT) doio$(OBJ_EXT) regexec$(OBJ_EXT) taint$(OBJ_EXT) deb$(OBJ_EXT) universal$(OBJ_EXT) globals$(OBJ_EXT) perlio$(OBJ_EXT)
+obj3 = doop$(OBJ_EXT) doio$(OBJ_EXT) regexec$(OBJ_EXT) taint$(OBJ_EXT) deb$(OBJ_EXT) universal$(OBJ_EXT) globals$(OBJ_EXT) perlio$(OBJ_EXT) kanji$(OBJ_EXT)
 
 obj = $(obj1) $(obj2) $(obj3) $(ARCHOBJS)
 
diff -urN perl5.004_02/README.j jperl5.004_02/README.j
--- perl5.004_02/README.j	Thu Jan  1 09:00:00 1970
+++ jperl5.004_02/README.j	Fri Aug  8 10:07:23 1997
@@ -0,0 +1,52 @@
+perl5.003に対する日本語パッチです。
+jperl4.019をperl5に移植した形になっています。
+
+* 正規表現
+* chop
+* split
+* tr
+* フォーマット
+* open ( 漢字2文字目の|を認識)
+
+が日本語化されています。逆に日本語化されていないのは，
+
+* reverse
+* index, rindex
+
+スイッチとして-bが追加されています。-bを付けると，日本語機能が
+offになります。
+
+新しいモジュールとして，
+I18N::Japaneseが追加されています。
+
+use I18N::Japanese qw(re tr format);
+
+  あるいは
+
+no I18N::Japanese qw(re tr);
+
+のように使います。
+
+起動ファイル名がjperlの場合には，
+自動的に use I18N::Japanese qw(re tr format string);
+が起動時に行われます。
+
+バグ : use I18N::Japanese qw(format)指定はブロックスコープを持ちません。
+要するに，プログラムテキスト中，最後に行ったformat指定がプログラム全体を
+通して有効になります。
+
+日本語コードは一応EUCとMS漢字をサポート。
+MS漢字はちょっとあやしい。
+
+Wed Jan 31 12:11:49 1996
+斎藤靖
+yasushi@cs.washington.edu
+http://cs.washington.edu/homes/yasushi
+
+
+Tue Aug 20 18:15:00 1996  Yasushi Saito  <yasushi@satin.cs.washington.edu>
+
+	* ^.が日本語にマッチしないバグ修正。渡辺博文＠ソニーさん。
+	  5.003対応。
+
+
diff -urN perl5.004_02/README.jp jperl5.004_02/README.jp
--- perl5.004_02/README.jp	Thu Jan  1 09:00:00 1970
+++ jperl5.004_02/README.jp	Fri Aug  8 10:08:06 1997
@@ -0,0 +1,61 @@
+1. はじめに
+
+  perl5.004_02 に対する日本語パッチです.
+  jperl5.003.patch7 と jperl 4.036 をミックスした形で
+  移植してます.
+
+  斎藤さんの変更は README.j をお読みください．
+
+  t/kanji/*.t は chmod +x し. sjis で make するときは
+  t/kanji/* は sjis に変換しといてください.
+
+2. warning message について
+
+  |perl: warning: Setting locale failed.
+  |perl: warning: Please check that your locale settings:
+  |	LC_ALL = (unset),
+  |	LANG = "ja_JP.ujis"
+  |    are supported and installed on your system.
+  |perl: warning: Falling back to the standard locale ("C").
+
+  という warning message は環境変数 PERL_BADLANG を設定する
+  ことで抑制することができます.
+  csh 系: setenv PERL_BADLANG 0
+  sh  系: PERL_BADLANG=0; export PERL_BADLANG
+  もしくは ./Configure -U d_setlocale してから make しましょう.
+
+3. jcode.pl を動かす
+
+  jcode.pl の先頭に no I18N::Japanese; を入れます.
+  後は script の中で適宜 use, no を使い分けてください.
+  基本的に jcode 内の subroutine を call する前に
+  no I18N::Japanese; して jperl 特有の機能を使う前に
+  use I18N::Japanese; します.
+
+
+4. jperl4.036 からの機能.
+
+  文字種のオプション
+
+    -Lsjis    日本語 Shift-JIS
+    -Leuc     日本語 EUC
+    -Ltca     中国語(台湾) TCA
+    -Lkseuc   韓国語EUC(KS C5601-1987)
+    -Llatin   1バイト文字のみ
+
+  の 5 通りに切り替わります．jperl という名前で起動された場合
+  はコンパイル時に指定した文字種になります(でも Configure では
+  euc か sjis しか選べない. もうしわけない).
+
+  jperl5.003.patch7 では full path で表わしたときにディレクト
+  リに jperl という文字が含まれていたら日本語機能が ON になっ
+  ていたので basename に jperl が含まれているときだけ ON にな
+  るようにしました(そうしないと jperl という名前のディレクトリ
+  で make するとデバッグできん :-) これは主に DOS での話). つ
+  いでに sjisperl, eucperl なんて名前で起動されればその文字種
+  になるようにしました.
+
+Fri Aug  8 10:08:04 JST 1997
+渡辺博文
+watanabe@ase.ptg.sony.so.jp
+$CPAN/authors/id/WATANABE
diff -urN perl5.004_02/config_h.SH jperl5.004_02/config_h.SH
--- perl5.004_02/config_h.SH	Fri May  9 01:57:47 1997
+++ jperl5.004_02/config_h.SH	Fri Aug  8 10:07:23 1997
@@ -123,6 +123,13 @@
 #define CPPSTDIN "$cppstdin"
 #define CPPMINUS "$cppminus"
 
+/* EUC
+ *      It defined, the perl handles Japanese EUC characters.
+ *      Otherwise, it handles Shift JIS(aka MS Kanji) characters.
+ */
+#$d_euc EUC                     /**/
+
+
 /* HAS_ALARM:
  *	This symbol, if defined, indicates that the alarm routine is
  *	available.
diff -urN perl5.004_02/doio.c jperl5.004_02/doio.c
--- perl5.004_02/doio.c	Tue Jul 29 10:08:11 1997
+++ jperl5.004_02/doio.c	Fri Aug  8 10:07:23 1997
@@ -252,7 +252,8 @@
 	    else
 		fp = PerlIO_open(name,mode);
 	}
-	else if (name[len-1] == '|') {
+	else if (name[len-1] == '|'
+		 && kpart(name, name+len-1) == KPART_OTHER) {
 	    name[--len] = '\0';
 	    while (len && isSPACE(name[len-1]))
 		name[--len] = '\0';
diff -urN perl5.004_02/doop.c jperl5.004_02/doop.c
--- perl5.004_02/doop.c	Thu Apr 24 00:22:20 1997
+++ jperl5.004_02/doop.c	Fri Aug  8 10:07:23 1997
@@ -18,23 +18,55 @@
 #include <signal.h>
 #endif
 
+/* Cat 1 or 2-byte char into SV */
+static I32
+sv_catkanji(sv, tch)
+SV *sv;
+U32 tch;
+{
+    char xx;
+    /* horendously inefficent. I'll fix it later. hold on
+       -yasushi */
+    if (tch > 255) {
+	xx = tch >> 8;
+	sv_catpvn(sv, &xx, 1);
+	xx = tch;
+	sv_catpvn(sv, &xx, 1);
+	return 2;
+    } else {
+	xx = tch;
+	sv_catpvn(sv, &xx, 1);
+	return 1;
+    }
+}     
+
+
 I32
 do_trans(sv,arg)
 SV *sv;
 OP *arg;
 {
-    register short *tbl;
-    register U8 *s;
-    register U8 *send;
-    register U8 *d;
-    register I32 ch;
+    register U16 *tbl;
+    register U8 *s, *next_s;
     register I32 matches = 0;
+    register I32 ch;
+    register U8 *send;
+    STRLEN dlen = 0;
+    U32 last_rch;
+    SV *dest_sv = sv_newmortal();
+			    
     register I32 squash = op->op_private & OPpTRANS_SQUASH;
+    I32 delete = op->op_private & OPpTRANS_DELETE;
+    I32 complement = op->op_private & OPpTRANS_COMPLEMENT;
+    I32 kanji = op->op_private & OPpTRANS_KANJI;
+
     STRLEN len;
 
+    sv_setpvn(dest_sv, "", 0);
+
     if (SvREADONLY(sv))
 	croak(no_modify);
-    tbl = (short*)cPVOP->op_pv;
+    tbl = (U16 *)cPVOP->op_pv;
     s = (U8*)SvPV(sv, len);
     if (!len)
 	return 0;
@@ -45,37 +77,61 @@
     if (!tbl || !s)
 	croak("panic: do_trans");
     DEBUG_t( deb("2.TBL\n"));
-    if (!op->op_private) {
-	while (s < send) {
-	    if ((ch = tbl[*s]) >= 0) {
-		matches++;
-		*s = ch;
+
+    while (s < send) {
+	U32 tch, rch;
+	U8 *nest_s;
+	U16 *tp;
+	int matched;
+
+	if (kanji && iskanji(*s) && s < send-1) {
+	    tch = twochar_to_int(*s, *(s+1));
+	    next_s = s+2;
+	} else {
+	    tch = *(U8*)s;
+	    next_s = s+1;
+	}
+
+	/* look for ch in tbl */
+	if (!complement) {
+	    for (tp = tbl; *tp != (U16)(-1); tp += 2) {
+		if (*tp == tch) break;
+	    }
+	    matched = (*tp != (U16)(-1));
+	    rch = tp[1];
+	} else {
+	    for (tp = tbl; *tp != (U16)(-1); tp += 2) {
+		if (*tp == tch) break;
 	    }
-	    s++;
+	    matched = (*tp == (U16)(-1));
+	    if (matched && tbl != tp && tp[-1] != (U16)-1)
+		tch = tp[-1];
+	    rch = (U16)(delete ? -2 : -1);
 	}
-    }
-    else {
-	d = s;
-	while (s < send) {
-	    if ((ch = tbl[*s]) >= 0) {
-		*d = ch;
-		if (matches++ && squash) {
-		    if (d[-1] == *d)
-			matches--;
-		    else
-			d++;
+
+	
+	if (!matched) {
+	    dlen += sv_catkanji(dest_sv, tch);
+	} else {
+	    if (rch == (U16)(-2)) {
+		/* delete this character */;
+	    } else if (squash) {
+		if (last_rch == (rch==(U16)(-1)?tch:rch)) {
+		    ;
+		} else {
+		    dlen += sv_catkanji(dest_sv, rch == (U16)(-1) ? tch : rch);
+		    matches++;
 		}
-		else
-		    d++;
+	    } else {
+		dlen += sv_catkanji(dest_sv, rch==(U16)(-1) ? tch : rch);
+		matches++;
 	    }
-	    else if (ch == -1)		/* -1 is unmapped character */
-		*d++ = *s;		/* -2 is delete character */
-	    s++;
-	}
-	matches += send - d;	/* account for disappeared chars */
-	*d = '\0';
-	SvCUR_set(sv, d - (U8*)SvPVX(sv));
+	}
+	last_rch = ((rch==(U16)(-1)||rch==(U16)(-2)) ? tch : rch);
+	s = next_s;
     }
+    matches += (s-(U8*)SvPVX(sv)) - dlen;	/* account for disappeared chars */
+    sv_setpvn(sv, SvPVX(dest_sv), dlen);
     SvSETMAGIC(sv);
     return matches;
 }
@@ -241,8 +297,15 @@
     if (len && !SvPOK(sv))
 	s = SvPV_force(sv, len);
     if (s && len) {
-	s += --len;
-	sv_setpvn(astr, s, 1);
+	if (kpart(s, s+len-1) == KPART_KANJI_2) {
+	    len -= 2;
+	    s += len;
+	    sv_setpvn(astr, s, 2);
+	} else {
+	    len--;
+	    s += len;
+	    sv_setpvn(astr, s, 1);
+	}
 	*s = '\0';
 	SvCUR_set(sv, len);
 	SvNIOK_off(sv);
diff -urN perl5.004_02/embed.h jperl5.004_02/embed.h
--- perl5.004_02/embed.h	Tue Jul 29 09:33:25 1997
+++ jperl5.004_02/embed.h	Fri Aug  8 10:07:23 1997
@@ -1268,6 +1268,7 @@
 #define mess_sv			(curinterp->Imess_sv)
 #define minus_F			(curinterp->Iminus_F)
 #define minus_a			(curinterp->Iminus_a)
+#define minus_b			(curinterp->Iminus_b)
 #define minus_c			(curinterp->Iminus_c)
 #define minus_l			(curinterp->Iminus_l)
 #define minus_n			(curinterp->Iminus_n)
@@ -1422,6 +1423,7 @@
 #define Imess_sv		mess_sv
 #define Iminus_F		minus_F
 #define Iminus_a		minus_a
+#define Iminus_b		minus_b
 #define Iminus_c		minus_c
 #define Iminus_l		minus_l
 #define Iminus_n		minus_n
@@ -1585,6 +1587,7 @@
 #define mess_sv			Perl_mess_sv
 #define minus_F			Perl_minus_F
 #define minus_a			Perl_minus_a
+#define minus_b			Perl_minus_b
 #define minus_c			Perl_minus_c
 #define minus_l			Perl_minus_l
 #define minus_n			Perl_minus_n
diff -urN perl5.004_02/form.h jperl5.004_02/form.h
--- perl5.004_02/form.h	Fri Mar  7 00:46:34 1997
+++ jperl5.004_02/form.h	Fri Aug  8 10:07:23 1997
@@ -23,4 +23,5 @@
 #define FF_NEWLINE	13
 #define FF_BLANK	14
 #define FF_MORE		15
+#define FF_OPTION	16 /* c : hints value */
 
diff -urN perl5.004_02/jcode.pl-2.3.pat jperl5.004_02/jcode.pl-2.3.pat
--- perl5.004_02/jcode.pl-2.3.pat	Thu Jan  1 09:00:00 1970
+++ jperl5.004_02/jcode.pl-2.3.pat	Fri Aug  8 10:07:23 1997
@@ -0,0 +1,5 @@
+--- jcode.pl-2.3	Tue Feb 25 14:09:15 1997
++++ jcode.pl	Thu Jul 10 19:56:31 1997
+@@ -1 +1,2 @@
++no I18N::Japanese;
+ package jcode;
diff -urN perl5.004_02/kanji.c jperl5.004_02/kanji.c
--- perl5.004_02/kanji.c	Thu Jan  1 09:00:00 1970
+++ jperl5.004_02/kanji.c	Fri Aug  8 10:07:23 1997
@@ -0,0 +1,401 @@
+/*kanji.c: kanji handling routines
+ *$Id$
+ */
+
+#include "EXTERN.h"
+#include "perl.h"
+
+unsigned char langtype_euc_tab[256] =
+{
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 0x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 1x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 2x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 3x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 4x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 5x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 6x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 7x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1,	/* 8x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 9x */
+    0, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Ax */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Bx */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Cx */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Dx */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Ex */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 0,	/* Fx */
+};
+unsigned char langtype_sjis_tab[256] =
+{
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 0x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 1x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 2x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 3x */
+    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,	/* 4x */
+    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,	/* 5x */
+    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,	/* 6x */
+    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0,	/* 7x */
+    2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* 8x */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* 9x */
+    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,	/* Ax */
+    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,	/* Bx */
+    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,	/* Cx */
+    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,	/* Dx */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Ex */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 0, 0, 0,	/* Fx */
+};
+unsigned char langtype_ks_euc_tab[256] =
+{
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 0x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 1x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 2x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 3x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 4x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 5x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 6x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 7x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 8x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 9x */
+    0, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Ax */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Bx */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Cx */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Dx */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Ex */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 0,	/* Fx */
+};
+unsigned char langtype_tca_tab[256] =
+{
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 0x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 1x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 2x */
+    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0, 0,	/* 3x */
+    0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,	/* 4x */
+    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0,	/* 5x */
+    0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,	/* 6x */
+    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0,	/* 7x */
+    1, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* 8x */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* 9x */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Ax */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Bx */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Cx */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Dx */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,	/* Ex */
+    3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 0, 0,	/* Fx */
+};
+unsigned char langtype_latin_tab[256] =
+{
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 0x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 1x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 2x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 3x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 4x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 5x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 6x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 7x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 8x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* 9x */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* Ax */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* Bx */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* Cx */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* Dx */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* Ex */
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,	/* Fx */
+};
+struct _langtypetable langtypetable[]=
+{
+    {langtype_sjis_tab, "SJIS"},
+    {langtype_euc_tab, "EUC"},
+    {langtype_ks_euc_tab, "KSEUC"},
+    {langtype_tca_tab, "TCA"},
+    {langtype_latin_tab, "LATIN"},
+    {0, 0},
+};
+#ifdef EUC
+char *langtypesymbol = "EUC";
+unsigned char *langtype = langtype_euc_tab;
+#else
+#ifdef KSEUC
+char *langtypesymbol = "KSEUC";
+unsigned char *langtype = langtype_kseuc_tab;
+#else
+#ifdef TCA
+char *langtypesymbol = "TCA";
+unsigned char *langtype = langtype_tca_tab;
+#else
+#ifdef LATIN
+char *langtypesymbol = "LATIN";
+unsigned char *langtype = langtype_latin_tab;
+#else
+char *langtypesymbol = "SJIS";
+unsigned char *langtype = langtype_sjis_tab;
+#endif
+#endif
+#endif
+#endif
+
+#define IsLower(c)	('a' <= (unsigned char)(c) && (unsigned char)(c) <= 'z')
+#define IsUpper(c)	('A' <= (unsigned char)(c) && (unsigned char)(c) <= 'Z')
+#define ToLower(c)	((c) - 'A' + 'a')
+#define ToUpper(c)	((c) - 'a' + 'A')
+
+int
+set_lang_type(symbol)
+char *symbol;
+{
+    int i;
+
+    if (!symbol || !*symbol)
+    	return 0;
+    for (i = 0; langtypetable[i].symbol; i++) {
+	if (jstricmp(langtypetable[i].symbol, symbol) == 0) {
+	    langtype = langtypetable[i].table;
+	    langtypesymbol = langtypetable[i].symbol;
+	    return 0;
+	}
+    }
+    return -1;
+}
+
+/* 
+   Combine two chars C1, C2 into single 16-bit integer
+   */
+I32 twochar_to_int(c1, c2)
+int c1, c2;
+{
+    return (((unsigned char)c1)<<8 | ((unsigned char)c2));
+}
+I32 int_to_1stbyte(ch)
+I32 ch;
+{
+    return (ch >> 8);
+}
+I32 int_to_2ndbyte(ch)
+I32 ch;
+{
+    return ch & 0xFF;
+}
+
+/* kpart(pLim, pChr);
+ * char * pLim;  Buffer Top or Limit for scanning
+ * char * pChr;  Pointer to the char
+ *		return whith 1 : *pChr is First byte of MS-Kanji
+ *		return whith 2 : *pChr is Second byte of MS-Kanji
+ *		return whith 0 : otherwize
+ */
+int
+kpart(pLim, pChr)
+char *pLim;
+char *pChr;
+{
+    register char *p = pChr - 1;
+    register int ct = 0;
+
+    while (p >= pLim && iskanji(*p)) {
+	p--;
+	ct++;
+    }
+    return (ct & 1) ? 2 : iskanji(*pChr);
+}
+
+/* jstrlen: return the number of charctors in string.
+ */
+int 
+jstrlen(s)
+char *s;
+{
+    int len;
+
+    for (len = 0; *s; s++, len++) {
+	if (iskanji(*s) && s[1])
+	    s++;
+    }
+    return len;
+}
+/* jstlen: return the number of charctors in string.
+ */
+int 
+jstlen(s, len)
+char *s;
+int len;
+{
+    int l;
+    char *send = s + len;
+
+    for (l = 0; s < send; s++, l++) {
+	if (iskanji(*s) && s[1])
+	    s++;
+    }
+    return l;
+}
+
+/* jnthchar: return with the pointer to n'th charactor in string.
+ *           return (char *)0, when jstrlen(s) < n
+ */
+char *
+jnthchar(s, n)
+char *s;
+int n;
+{
+    if (n) {
+	while (--n) {
+	    if (iskanji(*s))
+		s++;
+	    s++;
+	}
+    }
+    return s;
+}
+
+/* jindex: return with the pointer to 'ch' in "string"
+ *         return with (char *)0, when not found
+ */
+char *
+jindex(s, c)
+char *s;
+int c;
+{
+    while (*s) {
+	if (UCH(*s) == UCH(c))
+	    return s;
+
+	if (iskanji(*s) && s[1]) {
+	    s++;
+	}
+	s++;
+    }
+    return (char *) 0;
+}
+
+/* jrindex: return with the pointer to Right end 'ch' in "string"
+ *          return with (char *)0, when not found
+ */
+char *
+jrindex(s, c)
+char *s;
+int c;
+{
+    char *olds = (char *) 0;
+
+    while (*s) {
+	if (UCH(*s) == UCH(c))
+	    olds = s;
+
+	if (iskanji(*s) && s[1]) {
+	    s++;
+	}
+	s++;
+    }
+    return olds;
+}
+
+char *
+jstrlower(s)
+char *s;
+{
+    char *ws = s;
+
+    while (*ws) {
+	if (iskanji(*ws) && ws[1])
+	    ws++;
+	else
+	    *ws = IsUpper(*ws) ? ToLower(*ws) : *ws;
+	ws++;
+    }
+    return s;
+}
+
+int
+jstricmp(s1, s2)
+char *s1;
+char *s2;
+{
+    int c1, c2;
+
+    for (;;) {
+	c1 = UCH(*s1++);
+	c2 = UCH(*s2++);
+	if (!iskanji(c1)) {
+	    c1 = IsUpper(c1) ? ToLower(c1) : c1;
+	    c2 = IsUpper(c2) ? ToLower(c2) : c2;
+	}
+	if (c1 != c2)
+	    return c1 - c2;
+	if (c1 == 0)
+	    return 0;
+	if (iskanji(c1)) {
+	    c1 = UCH(*s1++);
+	    c2 = UCH(*s2++);
+	    if (c1 != c2)
+		return c1 - c2;
+	    if (c1 == 0)
+		return 0;
+	}
+    }
+    return 0;
+}
+
+int
+jstrnicmp(s1, s2, n)
+char *s1;
+char *s2;
+int n;
+{
+    int c1, c2;
+
+    for (;;) {
+	if (n--)
+	    return 0;
+	c1 = UCH(*s1++);
+	c2 = UCH(*s2++);
+	if (!iskanji(c1)) {
+	    c1 = IsUpper(c1) ? ToLower(c1) : c1;
+	    c2 = IsUpper(c2) ? ToLower(c2) : c2;
+	}
+	if (c1 != c2)
+	    break;
+	if (c1 == 0)
+	    return 0;
+	if (iskanji(c1)) {
+	    if (n--)
+		return 0;
+	    c1 = UCH(*s1++);
+	    c2 = UCH(*s2++);
+	    if (c1 != c2)
+		break;
+	    if (c1 == 0)
+		return 0;
+	}
+    }
+    return c1 - c2;
+}
+
+unsigned int
+jnextcode(cc)
+unsigned int cc;
+{
+    unsigned int c1, c2;
+
+    if (langtype == langtype_latin_tab) {
+	return cc + 1; /* ?:-P */
+    }
+    c1 = ((cc >> 8) & 255);
+    c2 = (cc & 255);
+    c2++;
+    if (iskanji2(c2)) {
+	return (c1 << 8)|c2;
+    }
+    while (!iskanji2(c2)) {
+	c2++;
+    }
+    if (c2 <= 0x100) {
+	return (c1 << 8)|c2;
+    }
+    c2 &= 0xFF;
+    c1++;
+    if (iskanji(c1)) {
+	return (c1 << 8)|c2;
+    }
+    while (!iskanji(c1)) {
+	c1++;
+    }
+    return ((c1 & 255) << 8)|c2;    
+}
diff -urN perl5.004_02/kanji.h jperl5.004_02/kanji.h
--- perl5.004_02/kanji.h	Thu Jan  1 09:00:00 1970
+++ jperl5.004_02/kanji.h	Fri Aug  8 10:07:23 1997
@@ -0,0 +1,65 @@
+#ifndef _KANJI_H_INCLUDE
+#ifdef iskanji
+#undef iskanji
+#endif
+
+#define UCH(c)        ((unsigned char)(c))
+#define BND(l, h, c)  (((l) <= (c))&&((c) <= (h)))
+
+#define iskanji(c)  (!minus_b && (langtype[(int)((unsigned char)(c))] & 1))
+#define iskanji2(c) (!minus_b && (langtype[(int)((unsigned char)(c))] & 2))
+
+struct _langtypetable {
+	unsigned char *table;
+	char          *symbol;
+};
+extern char *langtypesymbol;
+extern unsigned char *langtype;
+extern struct _langtypetable langtypetable[];
+
+/* return values of kpart */
+#define KPART_KANJI_1 1 /* kanji 1st byte */
+#define KPART_KANJI_2 2 /* kanji 2nd byte */
+#define KPART_OTHER   0 /* other (ASCII) */
+
+I32 twochar_to_int _((int c1, int c2));
+I32 int_to_1stbyte _((I32 c1));
+I32 int_to_2ndbyte _((I32 c1));
+
+#if defined(__STDC__) || defined(MSDOS)
+int   set_lang_type(char *symbol);
+int   kpart(char *pLim,char *pChr);
+int   jstrlen(char *s);
+int   jstlen(char *s, int len);
+char *jnthchar(char *s,int n);
+char *jindex(char *s,int c);
+char *jrindex(char *s,int c);
+char *jstrlower(char *s);
+int   jstricmp(char *s1, char *s2);
+int   jstrnicmp(char *s1, char *s2, int n);
+#else
+int   set_lang_type();
+int   kpart();
+int   jstrlen();
+int   jstlen();
+char *jnthchar();
+char *jindex();
+char *jrindex();
+char *jstrlower();
+int   jstricmp();
+int   jstrnicmp();
+#endif
+#ifdef strlwr
+#undef strlwr
+#endif
+#define strlwr  jstrlower
+#ifdef index
+#undef index
+#endif
+#define index   jindex
+#ifdef rindex
+#undef rindex
+#endif
+#define rindex  jrindex
+#define _KANJI_H_INCLUDE 1
+#endif /* _KANJI_H_INCLUDE */
diff -urN perl5.004_02/lib/I18N/Japanese.pm jperl5.004_02/lib/I18N/Japanese.pm
--- perl5.004_02/lib/I18N/Japanese.pm	Thu Jan  1 09:00:00 1970
+++ jperl5.004_02/lib/I18N/Japanese.pm	Fri Aug  8 10:07:23 1997
@@ -0,0 +1,71 @@
+package I18N::Japanese;
+
+=head1 NAME
+
+Japanese - Perl pragma to control whether some built-in operations understand
+    composite characters or not.
+
+=head1 SYNOPSIS
+
+use Japanese qw(re tr format);
+nouse Japanese qw(re tr format);
+
+=head1 DESCRIPTION
+
+This module is actually a pragma module, though it begins with capital J.
+Import/unimport takes list of strings.
+
+=over 5
+    
+=re
+
+    Enables Japanized regexp, including split.
+
+=tr 
+
+    Enables Japanized tr//.
+
+=format
+
+    Enables Japanized format.
+
+=back
+
+    As usual, omitting the arguments means specifying all the bits.
+    The setting is valid only within the current block.
+    
+=cut
+
+# ウラワザ : 0x8000 は，文字列定数で2バイト文字を認識するかどうか
+sub import {
+    shift;
+    my $bits = 0;
+    @_ = qw(re tr format string) if (@_ == 0);
+    for (@_) {
+	$bits |= 0x1000 if $_ eq 're';
+	$bits |= 0x2000 if $_ eq 'tr';
+	$bits |= 0x4000 if $_ eq 'format';
+	$bits |= 0x8000 if $_ eq 'string';
+    }
+#    $bits |= 0x8000 if ($bits != 0);
+    
+    $^H |= $bits;
+    #print "imp : hints = $bits\n";
+}
+
+sub unimport {
+    shift;
+    my $bits = 0;
+    @_ = qw(re tr format string) if (@_ == 0);
+    for (@_) {
+	$bits |= 0x1000 if $_ eq 're';
+	$bits |= 0x2000 if $_ eq 'tr';
+	$bits |= 0x4000 if $_ eq 'format';
+	$bits |= 0x8000 if $_ eq 'string';
+    }
+#    $bits = 0x8000 if ($^H != 0x8000 && $bits == 0);
+    $^H &= ~$bits;
+    #print "unimp : hints = $bits\n";
+}
+
+1;
diff -urN perl5.004_02/op.c jperl5.004_02/op.c
--- perl5.004_02/op.c	Tue Aug  5 22:26:00 1997
+++ jperl5.004_02/op.c	Fri Aug  8 10:07:23 1997
@@ -1927,57 +1927,74 @@
     SV *rstr = ((SVOP*)repl)->op_sv;
     STRLEN tlen;
     STRLEN rlen;
-    register U8 *t = (U8*)SvPV(tstr, tlen);
-    register U8 *r = (U8*)SvPV(rstr, rlen);
-    register I32 i;
-    register I32 j;
+    register char *t = SvPV(tstr, tlen);
+    register char *r = SvPV(rstr, rlen);
+    register I32 i; /* indexes t */
+    register I32 j; /* indexes j */
+    register I32 k; /* indexes tbl */
+    I32 lasttch = -3;
+    I32 lastrch = -1;
+    I32 tbl_size = 256;
+      
     I32 delete;
     I32 complement;
-    register short *tbl;
+    I32 kanji;
+    /* the even index holds the t-char(in 2byte), and the odd index
+       holds the r-char(in 2 byte) if t-char is to be removed, then
+       r-char is -2. */
+    register U16 *tbl;
 
-    tbl = (short*)cPVOP->op_pv;
+    New(803, tbl, tbl_size, U16);
+    
+    
     complement	= op->op_private & OPpTRANS_COMPLEMENT;
     delete	= op->op_private & OPpTRANS_DELETE;
+    kanji	= op->op_private & OPpTRANS_KANJI;
     /* squash	= op->op_private & OPpTRANS_SQUASH; */
 
-    if (complement) {
-	Zero(tbl, 256, short);
-	for (i = 0; i < tlen; i++)
-	    tbl[t[i]] = -1;
-	for (i = 0, j = 0; i < 256; i++) {
-	    if (!tbl[i]) {
-		if (j >= rlen) {
-		    if (delete)
-			tbl[i] = -2;
-		    else if (rlen)
-			tbl[i] = r[j-1];
-		    else
-			tbl[i] = i;
-		}
-		else
-		    tbl[i] = r[j++];
+    for (i = 0, j = 0, k = 0; i < tlen || j < rlen; ) {
+	U32 tch, rch;
+	if (i >= tlen) {
+	    tch = lasttch;
+	} else {
+	    if (kanji && iskanji(t[i]) && i < tlen-1) {
+		tch = twochar_to_int(t[i], t[i+1]);
+		i+=2;
+	    } else {
+		tch = (unsigned char)t[i]; 
+		i++;
 	    }
+	    lasttch = tch;
 	}
-    }
-    else {
-	if (!rlen && !delete) {
-	    r = t; rlen = tlen;
-	}
-	for (i = 0; i < 256; i++)
-	    tbl[i] = -1;
-	for (i = 0, j = 0; i < tlen; i++,j++) {
-	    if (j >= rlen) {
-		if (delete) {
-		    if (tbl[t[i]] == -1)
-			tbl[t[i]] = -2;
-		    continue;
-		}
-		--j;
+	if (j >= rlen) {
+	    if (delete) rch = -2;
+	    else rch = lastrch;
+	} else {
+	    if (kanji && iskanji(r[j]) && j < rlen-1) {
+		rch = twochar_to_int(r[j], r[j+1]);
+		j += 2;
+	    } else {
+		rch = (unsigned char)r[j];
+		j++;
 	    }
-	    if (tbl[t[i]] == -1)
-		tbl[t[i]] = r[j];
+	    lastrch = rch;
 	}
+	if (k >= tbl_size) {
+	    tbl_size += 256;
+	    Renew(tbl, tbl_size, U16);
+	}
+	tbl[k++] = tch;
+	tbl[k++] = rch;
+    }
+    if (k >= tbl_size) {
+	tbl_size += 4;
+	Renew(tbl, tbl_size, U16);
     }
+    /* mark the end */
+    tbl[k++] = -1;
+    tbl[k++] = -1;
+    cPVOP->op_pv = (char*)tbl;
+    
     op_free(expr);
     op_free(repl);
 
diff -urN perl5.004_02/op.h jperl5.004_02/op.h
--- perl5.004_02/op.h	Wed May 14 03:59:40 1997
+++ jperl5.004_02/op.h	Fri Aug  8 10:07:23 1997
@@ -96,6 +96,7 @@
 #define OPpTRANS_SQUASH		16
 #define OPpTRANS_DELETE		32
 #define OPpTRANS_COMPLEMENT	64
+#define OPpTRANS_KANJI		128
 
 /* Private for OP_REPEAT */
 #define OPpREPEAT_DOLIST	64	/* List replication. */
diff -urN perl5.004_02/perl.c jperl5.004_02/perl.c
--- perl5.004_02/perl.c	Tue Aug  5 22:38:36 1997
+++ jperl5.004_02/perl.c	Fri Aug  8 10:07:42 1997
@@ -453,6 +453,8 @@
     Safefree(sv_interp);
 }
 
+int default_hints;
+
 int
 perl_parse(sv_interp, xsinit, argc, argv, env)
 PerlInterpreter *sv_interp;
@@ -506,6 +508,29 @@
 	return 0;
     }
 
+    s = rindex(origargv[0], '/');
+    if (!s)
+	s = origargv[0];
+    if (instr(s, "jperl"))
+	s = "";
+    else if (instr(s, "kseucperl"))
+	s = "KSEUC";
+    else if (instr(s, "tcaperl"))
+	s = "TCA";
+    else if (instr(s, "eucperl"))
+	s = "EUC";
+    else if (instr(s, "sjisperl"))
+	s = "SJIS";
+    else if (instr(s, "latinperl"))
+	s = "LATIN";
+    else
+    	s = 0;
+    if (s) {
+	default_hints = hints = HINT_KANJI_REGEXP | HINT_KANJI_TR
+	  | HINT_KANJI_FORMAT | HINT_KANJI_STRING;
+	set_lang_type(s);
+    }
+    
     if (main_root) {
 	curpad = AvARRAY(comppad);
 	op_free(main_root);
@@ -559,6 +584,7 @@
 	case '0':
 	case 'F':
 	case 'a':
+	case 'b':
 	case 'c':
 	case 'd':
 	case 'D':
@@ -574,6 +600,7 @@
 	case 'U':
 	case 'v':
 	case 'w':
+	case 'L':
 	    if (s = moreswitches(s))
 		goto reswitch;
 	    break;
@@ -1345,6 +1372,14 @@
 	minus_a = TRUE;
 	s++;
 	return s;
+    case 'b':
+	/* No kanji support */
+	minus_b = TRUE;
+	default_hints = hints &= ~(HINT_KANJI_REGEXP | HINT_KANJI_TR
+	  | HINT_KANJI_FORMAT | HINT_KANJI_STRING);
+	set_lang_type("LATIN");
+	s++;
+	return s;
     case 'c':
 	minus_c = TRUE;
 	s++;
@@ -1497,6 +1532,11 @@
 #endif
 
 	printf("\n\nCopyright 1987-1997, Larry Wall\n");
+ 	printf("\nJapanization patch 4 by Yasushi Saito, 1996\n");
+	printf("\nModified by Hirofumi Watanabe, 1996, 1997\n");
+	printf("jperl5.004_02-970808\n");
+    	printf("%s version\n", langtypesymbol);
+	
 #ifdef MSDOS
 	printf("\n\nMS-DOS port Copyright (c) 1989, 1990, Diomidis Spinellis\n");
 #endif
@@ -1518,6 +1558,32 @@
 	dowarn = TRUE;
 	s++;
 	return s;
+    case 'L':
+	{
+	    int f, i;
+	    int len;
+    	    char *lang;
+	    s++;
+	    for (f = i = 0; langtypetable[i].symbol; i++) {
+		len = strlen(langtypetable[i].symbol);
+		if (ibcmp(s, langtypetable[i].symbol, len) == 0) {
+		    set_lang_type(s);
+		    f = 1;
+		    default_hints = hints =
+		    	HINT_KANJI_REGEXP |
+		    	HINT_KANJI_TR |
+			HINT_KANJI_FORMAT |
+			HINT_KANJI_STRING;
+		    break;
+		}
+	    }
+	    if (!f)
+	        croak("Unknown language type '%s'", s);
+	    minus_b = FALSE;
+	    while (*s)
+		s++;
+	    return s;
+	}
     case '*':
     case ' ':
 	if (s[1] == '-')	/* Additional switches on #! line. */
diff -urN perl5.004_02/perl.h jperl5.004_02/perl.h
--- perl5.004_02/perl.h	Tue Aug  5 23:51:17 1997
+++ jperl5.004_02/perl.h	Fri Aug  8 10:07:23 1997
@@ -1661,6 +1661,11 @@
 #define HINT_STRICT_VARS	0x00000400
 #define HINT_LOCALE		0x00000800
 
+#define HINT_KANJI_REGEXP	0x00001000
+#define HINT_KANJI_TR		0x00002000
+#define HINT_KANJI_FORMAT	0x00004000
+#define HINT_KANJI_STRING	0x00008000
+
 /**************************************************************************/
 /* This regexp stuff is global since it always happens within 1 expr eval */
 /**************************************************************************/
@@ -1728,6 +1733,7 @@
 IEXT bool	Iminus_p;
 IEXT bool	Iminus_l;
 IEXT bool	Iminus_a;
+IEXT bool	Iminus_b;
 IEXT bool	Iminus_F;
 IEXT bool	Idoswitches;
 IEXT bool	Idowarn;
@@ -2027,6 +2033,21 @@
 EXT MGVTBL vtbl_amagic;
 EXT MGVTBL vtbl_amagicelem;
 #endif /* OVERLOAD */
+
+#include "kanji.h"
+#undef isalpha
+#undef isascii
+#undef isdigit
+#undef islower
+#undef isspace
+#undef isupper
+#define isalpha(c) (BND('A', 'Z', UCH(c))||BND('a', 'z', UCH(c)))
+#define isascii(c) (UCH(c) <= 0x7F)
+#define isdigit(c) BND('0', '9', UCH(c))
+#define islower(c) BND('a', 'z', UCH(c))
+#define isspace(c) (BND(9, 13, UCH(c)) || (c) == ' ')
+#define isupper(c) BND('A', 'Z', UCH(c))
+
 
 #endif /* !DOINIT */
 
diff -urN perl5.004_02/pod/jperl.pod jperl5.004_02/pod/jperl.pod
--- perl5.004_02/pod/jperl.pod	Thu Jan  1 09:00:00 1970
+++ jperl5.004_02/pod/jperl.pod	Fri Aug  8 10:07:23 1997
@@ -0,0 +1,114 @@
+=head1 NAME
+
+jperl - Japanized Perl.
+
+=head1 DESCRIPTION
+
+Jperl is a variation of L<perl> that recognizes 16-bit Japanese characters.
+It supports Japanese EUC and Shift JIS(aka MS Kanji code).
+
+Currently, it understands Japanese in the following circumstances.
+
+=over 5
+
+=item * Regular Expression
+
+ '.' matches either 1-byte ASCII character or 2-byte Japanese character.
+Character class notation('[X-Y]', '[^X-Y]') recognizes a
+Japanese letter as a character.
+
+=item * tr
+
+Similarly, tr recognizes a Japanese letter as a letter.
+
+=item * split
+
+Similar.
+
+=item * formats
+
+It doesn't cut 2-byte letter at the middle when linebreaking.
+
+=item * chop
+
+It understands 2-byte letters.
+
+=back
+
+=head1 ADDITIONAL COMMAND LINE ARGUMENT
+
+=over 5
+
+=item B<-b>
+
+turns off Japanese mode. This is supported only in Japanized perl.
+Without this option, most 8-bit characters are interpreted to be
+composite character, and some programs that
+use 8th bit for some control information(L<pod2man> is an example) won't 
+work. With B<-b>, jperl will be compatible with the original perl.
+
+=back
+
+=head1 AUTHOR
+
+Yasushi Saito
+
+yasushi@cs.washington.edu
+
+=head1 NAME
+
+jperl - Japanized Perl.
+
+=head1 DESCRIPTION
+
+Jperl is a variation of L<perl> that recognizes 16-bit Japanese characters.
+It supports Japanese EUC and Shift JIS(aka MS Kanji code).
+
+Currently, it understands Japanese in the following circumstances.
+
+=over 5
+
+=item * Regular Expression
+
+ '.' matches either 1-byte ASCII character or 2-byte Japanese character.
+Character class notation('[X-Y]', '[^X-Y]') recognizes a
+Japanese letter as a character.
+
+=item * tr
+
+Similarly, tr recognizes a Japanese letter as a letter.
+
+=item * split
+
+Similar.
+
+=item * formats
+
+It doesn't cut 2-byte letter at the middle when linebreaking.
+
+=item * chop
+
+It understands 2-byte letters.
+
+=back
+
+=head1 ADDITIONAL COMMAND LINE ARGUMENT
+
+=over 5
+
+=item B<-b>
+
+turns off Japanese mode. This is supported only in Japanized perl.
+Without this option, most 8-bit characters are interpreted to be
+composite character, and some programs that
+use 8th bit for some control information(L<pod2man> is an example) won't 
+work. With B<-b>, jperl will be compatible with the original perl.
+
+=back
+
+=head1 AUTHOR
+
+Yasushi Saito
+
+yasushi@cs.washington.edu
+
diff -urN perl5.004_02/pod/perlrun.pod jperl5.004_02/pod/perlrun.pod
--- perl5.004_02/pod/perlrun.pod	Fri Aug  1 04:52:32 1997
+++ jperl5.004_02/pod/perlrun.pod	Fri Aug  8 10:07:23 1997
@@ -207,6 +207,14 @@
 
 An alternate delimiter may be specified using B<-F>.
 
+=item B<-b>
+
+turns off Japanese mode. This is supported only in Japanized perl.
+Without this option, most 8-bit characters are interpreted to be
+composite character(i.e., Japanese Kanji), and some programs that
+use 8th bit for some control information(L<pod2man> is an example) won't 
+work. With -b, jperl will be compatible with the original perl.
+
 =item B<-c>
 
 causes Perl to check the syntax of the script and then exit without
diff -urN perl5.004_02/pod/pod2man.PL jperl5.004_02/pod/pod2man.PL
--- perl5.004_02/pod/pod2man.PL	Fri Aug  1 06:03:43 1997
+++ jperl5.004_02/pod/pod2man.PL	Fri Aug  8 10:07:23 1997
@@ -25,7 +25,7 @@
 # You can use $Config{...} to use Configure variables.
 
 print OUT <<"!GROK!THIS!";
-$Config{startperl}
+$Config{startperl} -b
     eval 'exec $Config{perlpath} -S \$0 \${1+"\$@"}'
 	if \$running_under_some_shell;
 
diff -urN perl5.004_02/pp.c jperl5.004_02/pp.c
--- perl5.004_02/pp.c	Tue Aug  5 22:07:43 1997
+++ jperl5.004_02/pp.c	Fri Aug  8 10:07:23 1997
@@ -1996,11 +1996,17 @@
 	    TAINT;
 	    SvTAINTED_on(sv);
 	    for (; s < send; s++)
-		*s = toUPPER_LC(*s);
+		if (hints & HINT_KANJI_STRING && iskanji(*s) && s+1 < send)
+		    s++;
+		else
+		    *s = toUPPER_LC(*s);
 	}
 	else {
 	    for (; s < send; s++)
-		*s = toUPPER(*s);
+		if (hints & HINT_KANJI_STRING && iskanji(*s) && s+1 < send)
+		    s++;
+		else
+		    *s = toUPPER(*s);
 	}
     }
     RETURN;
@@ -2028,11 +2034,17 @@
 	    TAINT;
 	    SvTAINTED_on(sv);
 	    for (; s < send; s++)
-		*s = toLOWER_LC(*s);
+		if (hints & HINT_KANJI_STRING && iskanji(*s) && s+1 < send)
+		    s++;
+		else
+		    *s = toLOWER_LC(*s);
 	}
 	else {
 	    for (; s < send; s++)
-		*s = toLOWER(*s);
+		if (hints & HINT_KANJI_STRING && iskanji(*s) && s+1 < send)
+		    s++;
+		else
+		    *s = toLOWER(*s);
 	}
     }
     RETURN;
@@ -2051,7 +2063,10 @@
 	SvGROW(TARG, (len * 2) + 1);
 	d = SvPVX(TARG);
 	while (len--) {
-	    if (!isALNUM(*s))
+	    if (hints & HINT_KANJI_STRING && iskanji(*s) && len) {
+		*d++ = *s++;
+		len--;
+	    } else if (!isALNUM(*s))
 		*d++ = '\\';
 	    *d++ = *s++;
 	}
diff -urN perl5.004_02/pp_ctl.c jperl5.004_02/pp_ctl.c
--- perl5.004_02/pp_ctl.c	Tue Aug  5 22:26:00 1997
+++ jperl5.004_02/pp_ctl.c	Fri Aug  8 10:07:23 1997
@@ -243,6 +243,7 @@
     I32 itemsize;
     I32 fieldsize;
     I32 lines = 0;
+    I32 kanji = (hints & HINT_KANJI_FORMAT);
     bool chopspace = (strchr(chopset, ' ') != Nullch);
     char *chophere;
     char *linemark;
@@ -328,6 +329,11 @@
 	    itemsize = len;
 	    if (itemsize > fieldsize)
 		itemsize = fieldsize;
+	    if (kanji
+		&& kpart(item, item+itemsize) == KPART_KANJI_2) {
+		/* don't split kanji at the middle. */
+		itemsize--;
+	    }
 	    send = chophere = s + itemsize;
 	    while (s < send) {
 		if (*s & ~31)
@@ -355,6 +361,11 @@
 	    }
 	    else {
 		itemsize = fieldsize;
+		if (kanji
+		    && kpart(item, item+itemsize) == KPART_KANJI_2) {
+		    /* don't split kanji at the middle. */
+		    itemsize--;
+		}
 		send = chophere = s + itemsize;
 		while (s < send || (s == send && isSPACE(*s))) {
 		    if (isSPACE(*s)) {
@@ -2231,6 +2242,7 @@
 
 PP(pp_require)
 {
+    extern int default_hints;
     dSP;
     register CONTEXT *cx;
     SV *sv;
@@ -2352,7 +2364,7 @@
     name = savepv(name);
     SAVEFREEPV(name);
     SAVEI32(hints);
-    hints = 0;
+    hints = default_hints;
  
     /* switch to eval mode */
 
@@ -2603,6 +2615,7 @@
     U16 *linepc;
     register I32 arg;
     bool ischop;
+    I32 kanji = (hints & HINT_KANJI_FORMAT);
 
     if (len == 0)
 	croak("Null picture in formline");
@@ -2618,6 +2631,12 @@
     }
 
     while (s <= send) {
+	if (kanji && iskanji(*s)) {
+	    s += 2;
+	    skipspaces = 0;
+	    continue;
+	}
+
 	switch (*s++) {
 	default:
 	    skipspaces = 0;
diff -urN perl5.004_02/pp_sys.c jperl5.004_02/pp_sys.c
--- perl5.004_02/pp_sys.c	Thu Aug  7 23:42:07 1997
+++ jperl5.004_02/pp_sys.c	Fri Aug  8 10:07:23 1997
@@ -2533,8 +2533,9 @@
 	if (!*s) {			/* null never allowed in text */
 	    odd += len;
 	    break;
-	}
-	else if (*s & 128)
+	} else if (iskanji(*s) && s[1]) {
+	    ;
+	} else if (*s & 128)
 	    odd++;
 	else if (*s < 32 &&
 	  *s != '\n' && *s != '\r' && *s != '\b' &&
diff -urN perl5.004_02/regcomp.c jperl5.004_02/regcomp.c
--- perl5.004_02/regcomp.c	Fri Aug  1 03:18:44 1997
+++ jperl5.004_02/regcomp.c	Fri Aug  8 10:07:23 1997
@@ -71,6 +71,8 @@
 #define	STATIC	static
 #endif
 
+#define ISKANJI(c) ((hints & HINT_KANJI_REGEXP) && iskanji(c))
+
 #define	ISMULT1(c)	((c) == '*' || (c) == '+' || (c) == '?')
 #define	ISMULT2(s)	((*s) == '*' || (*s) == '+' || (*s) == '?' || \
 	((*s) == '{' && regcurly(s)))
@@ -111,6 +113,7 @@
 static void regtail _((char *, char *));
 static char* regwhite _((char *, char *));
 static char* nextchar _((void));
+static void regclassfree _((struct regexp *));
 
 /*
  - pregcomp - compile a regular expression into internal code
@@ -200,7 +203,7 @@
     /* Dig out information for optimizations. */
     pm->op_pmflags = regflags;
     r->regstart = Nullsv;	/* Worst-case defaults. */
-    r->reganch = 0;
+    r->reganch = ((hints & HINT_KANJI_REGEXP) ? ROPT_KANJI : 0);
     r->regmust = Nullsv;
     r->regback = -1;
     r->regstclass = Nullch;
@@ -793,12 +796,12 @@
 	else
 	    ret = regnode(ANY);
 	regnaughty++;
-	*flagp |= HASWIDTH|SIMPLE;
+	*flagp |= HASWIDTH;
 	break;
     case '[':
 	regparse++;
 	ret = regclass();
-	*flagp |= HASWIDTH|SIMPLE;
+	*flagp |= HASWIDTH;
 	break;
     case '(':
 	nextchar();
@@ -949,7 +952,7 @@
 			  : EXACT);
 	    regc(0);		/* save spot for len */
 	    for (len = 0, p = regparse - 1;
-	      len < 127 && p < regxend;
+	      len < 126 && p < regxend;
 	      len++)
 	    {
 		oldp = p;
@@ -1040,7 +1043,22 @@
 		}
 		if (regflags & PMf_EXTENDED)
 		    p = regwhite(p, regxend);
-		if (ISMULT2(p)) { /* Back off on ?+*. */
+		if (ISKANJI(ender) && *p) {
+		    if (ISMULT2((p+1))) {
+			if (len) {
+			    p = oldp;
+			} else {
+			    len += 2;
+			    regc(ender);
+			    ender = *p++;
+			    regc(ender);
+			}
+			break;
+		    }
+		    len++;
+		    regc(ender);
+		    ender = *p++;
+		} else if (ISMULT2(p)) { /* Back off on ?+*. */
 		    if (len)
 			p = oldp;
 		    else {
@@ -1090,14 +1108,47 @@
 }
 
 static void
+block_on(b, x, y)
+char *b;
+int x;
+int y;
+{
+    int lo;
+    int hi;
+    int r_lo = x & 7;
+    int r_hi = y & 7;
+
+    if (!b)
+	return;
+
+    if (0x8000 <= x)
+	x -= 0x8000 - 0x100;
+    if (0x8000 <= y)
+	y -= 0x8000 - 0x100;
+
+    lo = x >> 3;
+    hi = y >> 3;
+
+    if (lo == hi) {
+	b[lo] |= (1 << (r_hi+1)) - (1 << r_lo);
+    } else {
+	int i;
+	for (i = lo + 1; i <  hi; i++)
+	    b[i] = 0xff;
+	b[lo] |= (0xff << r_lo);
+	b[hi] |= ~(0xff << (r_hi+1));
+    }
+}
+
+static void
 regset(opnd, c)
 char *opnd;
 register I32 c;
 {
-    if (opnd == &regdummy)
+    if (!opnd)
 	return;
     c &= 0xFF;
-    opnd[1 + (c >> 3)] |= (1 << (c & 7));
+    opnd[c >> 3] |= (1 << (c & 7));
 }
 
 static char *
@@ -1110,11 +1161,27 @@
     register char *ret;
     register I32 def;
     I32 numlen;
+    int ismb = 0;
+    char **bitvec = 0;
 
     ret = regnode(ANYOF);
     opnd = regcode;
-    for (class = 0; class < 33; class++)
+    regc(0);
+
+    if (opnd != &regdummy) {
+	if ((long)regcode & 1) /* fix alignment */
+	    regc(127);
+	if ((long)regcode & 2) { /* and fix alignment */
+	    regc(127);
+	    regc(127);
+	}
+	bitvec = (char **)regcode;
+    } else
+	    regsize += 3;
+
+    for (class = 0; class < sizeof (char *); class++)
 	regc(0);
+
     if (*regparse == '^') {	/* Complement of range. */
 	regnaughty++;
 	regparse++;
@@ -1122,6 +1189,7 @@
 	    *opnd |= ANYOF_INVERT;
     }
     if (opnd != &regdummy) {
+	Newz(1003, *bitvec, (0x8000 + 0x100) / 8, char);
 	if (regflags & PMf_FOLD)
 	    *opnd |= ANYOF_FOLD;
 	if (regflags & PMf_LOCALE)
@@ -1132,67 +1200,81 @@
     while (regparse < regxend && *regparse != ']') {
        skipcond:
 	class = UCHARAT(regparse++);
+	if (ISKANJI(class) && regparse < regxend) {
+	    if (*opnd & ANYOF_LITERAL)
+		FAIL("kanji and hex/octal literal");
+	    class = (class << 8) + UCHARAT(regparse++);
+	    ismb = 1;
+	}
 	if (class == '\\') {
 	    class = UCHARAT(regparse++);
 	    switch (class) {
 	    case 'w':
-		if (regflags & PMf_LOCALE) {
-		    if (opnd != &regdummy)
+		if (bitvec) {
+		    if (regflags & PMf_LOCALE) {
 			*opnd |= ANYOF_ALNUML;
-		}
-		else {
-		    for (class = 0; class < 256; class++)
-			if (isALNUM(class))
-			    regset(opnd, class);
+		    }
+		    else {
+			for (class = 0; class < 256; class++)
+			    if (isALNUM(class))
+				regset(*bitvec, class);
+		    }
 		}
 		lastclass = 1234;
 		continue;
 	    case 'W':
-		if (regflags & PMf_LOCALE) {
-		    if (opnd != &regdummy)
+		if (bitvec) {
+		    if (regflags & PMf_LOCALE) {
 			*opnd |= ANYOF_NALNUML;
-		}
-		else {
-		    for (class = 0; class < 256; class++)
-			if (!isALNUM(class))
-			    regset(opnd, class);
+		    }
+		    else {
+			for (class = 0; class < 256; class++)
+			    if (!isALNUM(class))
+				regset(*bitvec, class);
+			block_on(*bitvec, 0x8000, 0xffff);
+		    }
 		}
 		lastclass = 1234;
 		continue;
 	    case 's':
-		if (regflags & PMf_LOCALE) {
-		    if (opnd != &regdummy)
+		if (bitvec) {
+		    if (regflags & PMf_LOCALE) {
 			*opnd |= ANYOF_SPACEL;
-		}
-		else {
-		    for (class = 0; class < 256; class++)
-			if (isSPACE(class))
-			    regset(opnd, class);
+		    }
+		    else {
+			for (class = 0; class < 256; class++)
+			    if (isSPACE(class))
+				regset(*bitvec, class);
+		    }
 		}
 		lastclass = 1234;
 		continue;
 	    case 'S':
-		if (regflags & PMf_LOCALE) {
-		    if (opnd != &regdummy)
+		if (bitvec) {
+		    if (regflags & PMf_LOCALE) {
 			*opnd |= ANYOF_NSPACEL;
-		}
-		else {
-		    for (class = 0; class < 256; class++)
-			if (!isSPACE(class))
-			    regset(opnd, class);
+		    }
+		    else {
+			for (class = 0; class < 256; class++)
+			    if (!isSPACE(class))
+				regset(*bitvec, class);
+			block_on(*bitvec, 0x8000, 0xffff);
+		    }
 		}
 		lastclass = 1234;
 		continue;
 	    case 'd':
-		for (class = '0'; class <= '9'; class++)
-		    regset(opnd, class);
+		if (bitvec) {
+		    block_on(*bitvec, '0', '9');
+		}
 		lastclass = 1234;
 		continue;
 	    case 'D':
-		for (class = 0; class < '0'; class++)
-		    regset(opnd, class);
-		for (class = '9' + 1; class < 256; class++)
-		    regset(opnd, class);
+		if (bitvec) {
+		    block_on(*bitvec, 0, '0' - 1);
+		    block_on(*bitvec, '9' + 1, 255);
+		    block_on(*bitvec, 0x8000, 0xffff);
+		}
 		lastclass = 1234;
 		continue;
 	    case 'n':
@@ -1219,6 +1301,11 @@
 	    case 'x':
 		class = scan_hex(regparse, 2, &numlen);
 		regparse += numlen;
+		if (bitvec) {
+		    if (ismb)
+			FAIL("kanji and hex/octal literal");
+		    *opnd |= ANYOF_LITERAL;
+		}
 		break;
 	    case 'c':
 		class = UCHARAT(regparse++);
@@ -1228,6 +1315,11 @@
 	    case '5': case '6': case '7': case '8': case '9':
 		class = scan_oct(--regparse, 3, &numlen);
 		regparse += numlen;
+		if (bitvec) {
+		    if (ismb)
+			FAIL("kanji and hex/octal literal");
+		    *opnd |= ANYOF_LITERAL;
+		}
 		break;
 	    }
 	}
@@ -1245,8 +1337,8 @@
 		continue;	/* do it next time */
 	    }
 	}
-	for ( ; lastclass <= class; lastclass++)
-	    regset(opnd, lastclass);
+	if (bitvec)
+	    block_on(*bitvec, lastclass, class);
 	lastclass = class;
     }
     if (*regparse != ']')
@@ -1312,7 +1404,7 @@
 #ifdef REGALIGN
 #ifndef lint
     if (!((long)ret & 1))
-      *ret++ = 127;
+      *ret++ = 120;
 #endif
 #endif
     ptr = ret;
@@ -1517,7 +1609,6 @@
 }
 
 #ifdef DEBUGGING
-
 /*
  - regdump - dump a regexp onto Perl_debug_log in vaguely comprehensible form
  */
@@ -1548,7 +1639,16 @@
 	    PerlIO_printf(Perl_debug_log, "(%d)", (s-r->program)+(next-s));
 	s += 3;
 	if (op == ANYOF) {
-	    s += 33;
+	    char **bitvec;
+	    s++;
+	    if ((long)s & 1) /* fix alignment */
+		s++;
+	    if ((long)s & 2) /* fix alignment */
+		s += 2;
+	    bitvec = (char **)s;
+	    PerlIO_printf(Perl_debug_log, "bitvec: %08x, ", bitvec);
+	    PerlIO_printf(Perl_debug_log, " %08x", *bitvec);
+	    s += sizeof (char *);
 	}
 	if (regkind[(U8)op] == EXACT) {
 	    /* Literal string, where present. */
@@ -1771,5 +1871,42 @@
     }
     Safefree(r->startp);
     Safefree(r->endp);
+    regclassfree(r);
     Safefree(r);
+}
+
+static void
+regclassfree(r)
+struct regexp *r;
+{
+    register char *s;
+    register char op = EXACT;	/* Arbitrary non-END op. */
+
+
+    s = r->program + 1;
+    while (op != END) {	/* While that wasn't END last time... */
+#ifdef REGALIGN
+	if (!((long)s & 1))
+	    s++;
+#endif
+	op = OP(s);
+	s += regarglen[(U8)op] + 3;
+	if (op == ANYOF) {
+	    char **bitvec;
+	    s++;
+	    if ((long)s & 1) /* fix alignment */
+		s++;
+	    if ((long)s & 2) /* fix alignment */
+		s += 2;
+	    bitvec = (char **)s;
+	    Safefree(*bitvec);
+	    s += sizeof (char *);
+	}
+	if (regkind[(U8)op] == EXACT) {
+	    /* Literal string, where present. */
+	    s++;
+	    while (*s++)
+		;
+	}
+    }
 }
diff -urN perl5.004_02/regcomp.h jperl5.004_02/regcomp.h
--- perl5.004_02/regcomp.h	Sat Jun  7 09:42:21 1997
+++ jperl5.004_02/regcomp.h	Fri Aug  8 10:07:23 1997
@@ -256,6 +256,7 @@
 #define MAGIC 0234
 
 /* Flags for first parameter byte of ANYOF */
+#define ANYOF_LITERAL	0x80
 #define ANYOF_INVERT	0x40
 #define ANYOF_FOLD	0x20
 #define ANYOF_LOCALE	0x10
diff -urN perl5.004_02/regexec.c jperl5.004_02/regexec.c
--- perl5.004_02/regexec.c	Tue Jul 29 06:16:23 1997
+++ jperl5.004_02/regexec.c	Fri Aug  8 10:07:23 1997
@@ -58,6 +58,12 @@
 #ifndef STATIC
 #define	STATIC	static
 #endif
+#undef isascii
+#define isascii(c) (0 < (c) && (c) < 0x7F)
+
+#define KANJI_MODE (_prog->reganch & ROPT_KANJI)
+static regexp *_prog; /* save the latest regexp* here.
+		       XXX this won't work in multi-threaded environment */
 
 #ifdef DEBUGGING
 static I32 regnarrate = 0;
@@ -214,6 +220,8 @@
 	croak("NULL regexp parameter");
 	return 0;
     }
+    
+    _prog = prog; /* save the thing */
 
     if (startpos == strbeg)	/* is ^ valid at stringarg? */
 	regprev = '\n';
@@ -267,6 +275,9 @@
 	    s = startpos;
 	    minlen = SvCUR(prog->regmust);
 	}
+	if (KANJI_MODE && s && kpart(startpos, s) == KPART_KANJI_2) {
+	    s = startpos;
+	}
     }
 
     /* Mark beginning of line for ^ . */
@@ -311,9 +322,12 @@
 		if (*s == ch) {
 		    if (regtry(prog, s))
 			goto got_it;
+		    if (KANJI_MODE && iskanji(*s) && s+1<strend) s++; 
 		    s++;
-		    while (s < strend && *s == ch)
+		    while (s < strend && *s == ch) {
+			if (KANJI_MODE && iskanji(*s) && s+1<strend) s++; 
 			s++;
+		    }
 		}
 		s++;
 	    }
@@ -325,6 +339,7 @@
 	    {
 		if (regtry(prog, s))
 		    goto got_it;
+		if (KANJI_MODE && iskanji(*s) && s+1<strend) s++; 
 		s++;
 	    }
 	}
@@ -334,6 +349,7 @@
 	    {
 		if (regtry(prog, s))
 		    goto got_it;
+		if (KANJI_MODE && iskanji(*s) && s+1<strend) s++; 
 		s++;
 	    }
 	}
@@ -352,7 +368,13 @@
 	case ANYOF:
 	    c = OPERAND(c);
 	    while (s < strend) {
-		if (reginclass(c, *s)) {
+  		register I32 i = UCHARAT(s);
+		if (KANJI_MODE && !(*c & ANYOF_LITERAL) && iskanji(i) && s+1 < strend) {
+		    s++;
+		    i = (i << 8) + UCHARAT(s);
+		    s--;
+		}
+		if (reginclass(c, i)) {
 		    if (tmp && regtry(prog, s))
 			goto got_it;
 		    else
@@ -360,6 +382,7 @@
 		}
 		else
 		    tmp = 1;
+		if (KANJI_MODE && !(*c & ANYOF_LITERAL) && iskanji(*s)) s++;
 		s++;
 	    }
 	    break;
@@ -377,6 +400,7 @@
 		    if (regtry(prog, s))
 			goto got_it;
 		}
+		if (KANJI_MODE && iskanji(*s)) s++;
 		s++;
 	    }
 	    if ((minlen || tmp) && regtry(prog,s))
@@ -395,6 +419,7 @@
 		    tmp = !tmp;
 		else if (regtry(prog, s))
 		    goto got_it;
+		if (KANJI_MODE && iskanji(*s)) s++;
 		s++;
 	    }
 	    if ((minlen || !tmp) && regtry(prog,s))
@@ -410,6 +435,7 @@
 		}
 		else
 		    tmp = 1;
+		if (KANJI_MODE && iskanji(*s)) s++;
 		s++;
 	    }
 	    break;
@@ -424,6 +450,7 @@
 		}
 		else
 		    tmp = 1;
+		if (KANJI_MODE && iskanji(*s)) s++;
 		s++;
 	    }
 	    break;
@@ -437,6 +464,7 @@
 		}
 		else
 		    tmp = 1;
+		if (KANJI_MODE && iskanji(*s)) s++;
 		s++;
 	    }
 	    break;
@@ -451,6 +479,7 @@
 		}
 		else
 		    tmp = 1;
+		if (KANJI_MODE && iskanji(*s)) s++;
 		s++;
 	    }
 	    break;
@@ -464,6 +493,7 @@
 		}
 		else
 		    tmp = 1;
+		if (KANJI_MODE && iskanji(*s)) s++;
 		s++;
 	    }
 	    break;
@@ -478,6 +508,7 @@
 		}
 		else
 		    tmp = 1;
+		if (KANJI_MODE && iskanji(*s)) s++;
 		s++;
 	    }
 	    break;
@@ -491,6 +522,7 @@
 		}
 		else
 		    tmp = 1;
+		if (KANJI_MODE && iskanji(*s)) s++;
 		s++;
 	    }
 	    break;
@@ -505,6 +537,7 @@
 		}
 		else
 		    tmp = 1;
+		if (KANJI_MODE && iskanji(*s)) s++;
 		s++;
 	    }
 	    break;
@@ -544,6 +577,7 @@
 	do {
 	    if (regtry(prog, s))
 		goto got_it;
+	    if (KANJI_MODE && iskanji(*s)) s++;
 	} while (s++ < strend);
     }
 
@@ -728,11 +762,13 @@
 	case SANY:
 	    if (!nextchar && locinput >= regeol)
 		sayNO;
+	    if (KANJI_MODE && iskanji(nextchar) && locinput[1]) ++locinput;
 	    nextchar = UCHARAT(++locinput);
 	    break;
 	case ANY:
 	    if (!nextchar && locinput >= regeol || nextchar == '\n')
 		sayNO;
+	    if (KANJI_MODE && iskanji(nextchar) && locinput[1]) ++locinput;
 	    nextchar = UCHARAT(++locinput);
 	    break;
 	case EXACT:
@@ -772,6 +808,11 @@
 	    s = OPERAND(scan);
 	    if (nextchar < 0)
 		nextchar = UCHARAT(locinput);
+	    if (KANJI_MODE && !(*s & ANYOF_LITERAL)
+		&& iskanji(nextchar) && locinput[1]) {
+		locinput++;
+		nextchar = (nextchar << 8) + UCHARAT(locinput);
+	    }
 	    if (!reginclass(s, nextchar))
 		sayNO;
 	    if (!nextchar && locinput >= regeol)
@@ -787,6 +828,8 @@
 	    if (!(OP(scan) == ALNUM
 		  ? isALNUM(nextchar) : isALNUM_LC(nextchar)))
 		sayNO;
+	    if (KANJI_MODE && iskanji(nextchar) && locinput[1])
+	      ++locinput;
 	    nextchar = UCHARAT(++locinput);
 	    break;
 	case NALNUML:
@@ -798,6 +841,8 @@
 	    if (OP(scan) == NALNUM
 		? isALNUM(nextchar) : isALNUM_LC(nextchar))
 		sayNO;
+	    if (KANJI_MODE && iskanji(nextchar) && locinput[1])
+	      ++locinput;
 	    nextchar = UCHARAT(++locinput);
 	    break;
 	case BOUNDL:
@@ -839,6 +884,8 @@
 	    if (OP(scan) == SPACE
 		? isSPACE(nextchar) : isSPACE_LC(nextchar))
 		sayNO;
+	    if (KANJI_MODE && iskanji(nextchar) && locinput[1])
+	      ++locinput;
 	    nextchar = UCHARAT(++locinput);
 	    break;
 	case DIGIT:
@@ -851,6 +898,8 @@
 		sayNO;
 	    if (isDIGIT(nextchar))
 		sayNO;
+	    if (KANJI_MODE && iskanji(nextchar) && locinput[1])
+	      ++locinput;
 	    nextchar = UCHARAT(++locinput);
 	    break;
 	case REFFL:
@@ -1088,13 +1137,14 @@
 		if (ln && regrepeat(scan, ln) < ln)
 		    sayNO;
 		while (n >= ln || (n == 32767 && ln > 0)) { /* ln overflow ? */
-		    /* If it could work, try it. */
-		    if (c1 == -1000 ||
-			UCHARAT(reginput) == c1 ||
-			UCHARAT(reginput) == c2)
-		    {
-			if (regmatch(next))
-			    sayYES;
+		    if (!KANJI_MODE
+			|| kpart(locinput, reginput) != KPART_KANJI_2) {
+			/* If it could work, try it. */
+			if (c1 == -1000 ||
+			    UCHARAT(reginput) == c1 ||
+			    UCHARAT(reginput) == c2)
+			    if (regmatch(next))
+				sayYES;
 		    }
 		    /* Couldn't or didn't -- back up. */
 		    reginput = locinput + ln;
@@ -1112,13 +1162,14 @@
 		    (!multiline || OP(next) == SEOL))
 		    ln = n;			/* why back off? */
 		while (n >= ln) {
-		    /* If it could work, try it. */
-		    if (c1 == -1000 ||
-			UCHARAT(reginput) == c1 ||
-			UCHARAT(reginput) == c2)
-		    {
-			if (regmatch(next))
-			    sayYES;
+		    if (!KANJI_MODE
+			|| kpart(locinput, reginput) != KPART_KANJI_2) {
+			/* If it could work, try it. */
+			if (c1 == -1000 ||
+			    UCHARAT(reginput) == c1 ||
+			    UCHARAT(reginput) == c2)
+			    if (regmatch(next))
+				sayYES;
 		    }
 		    /* Couldn't or didn't -- back up. */
 		    n--;
@@ -1188,6 +1239,7 @@
     register char *opnd;
     register I32 c;
     register char *loceol = regeol;
+    char *oldscan;
 
     scan = reginput;
     if (max != 32767 && max < loceol - scan)
@@ -1195,8 +1247,10 @@
     opnd = OPERAND(p);
     switch (OP(p)) {
     case ANY:
-	while (scan < loceol && *scan != '\n')
+	while (scan < loceol && *scan != '\n') {
+	    if (KANJI_MODE && iskanji(*scan)) scan++;
 	    scan++;
+	}
 	break;
     case SANY:
 	scan = loceol;
@@ -1220,8 +1274,23 @@
 	    scan++;
 	break;
     case ANYOF:
-	while (scan < loceol && reginclass(opnd, *scan))
+  	c = UCHARAT(scan);
+	oldscan = scan;
+	if (KANJI_MODE && !(*opnd & ANYOF_LITERAL)
+	    && iskanji(c) && scan < loceol) {
 	    scan++;
+	    c = (c << 8) + UCHARAT(scan);
+	}
+	while (scan < loceol && reginclass(opnd, c)) {
+	    scan++;
+	    c = UCHARAT(scan);
+	    oldscan = scan;
+	    if (KANJI_MODE && iskanji(c) && scan < loceol) {
+		scan++;
+		c = (c << 8) + UCHARAT(scan);
+	    }
+	}
+	scan = oldscan;
 	break;
     case ALNUM:
 	while (scan < loceol && isALNUM(*scan))
@@ -1233,8 +1302,10 @@
 	    scan++;
 	break;
     case NALNUM:
-	while (scan < loceol && !isALNUM(*scan))
+	while (scan < loceol && !isALNUM(*scan)) {
+	    if (KANJI_MODE && iskanji(*scan)) scan++;
 	    scan++;
+	}
 	break;
     case NALNUML:
 	regtainted = TRUE;
@@ -1251,6 +1322,10 @@
 	    scan++;
 	break;
     case NSPACE:
+	while (scan < loceol && !isSPACE(*scan)) {
+	    if (KANJI_MODE && iskanji(*scan)) scan++;
+	    scan++;
+	}
 	while (scan < loceol && !isSPACE(*scan))
 	    scan++;
 	break;
@@ -1264,8 +1339,10 @@
 	    scan++;
 	break;
     case NDIGIT:
-	while (scan < loceol && !isDIGIT(*scan))
+	while (scan < loceol && !isDIGIT(*scan)) {
+	    if (KANJI_MODE && iskanji(*scan)) scan++;
 	    scan++;
+	}
 	break;
     default:		/* Called on something of 0 width. */
 	break;		/* So match right here or not at all. */
@@ -1288,10 +1365,22 @@
 {
     char flags = *p;
     bool match = FALSE;
+    char *bitvec;
 
-    c &= 0xFF;
-    if (p[1 + (c >> 3)] & (1 << (c & 7)))
+    p++;
+    if ((long)p & 1) /* fix alignment */
+	p++;
+    if ((long)p & 2) /* fix alignment */
+	p += 2;
+    bitvec = *(char **)p;
+
+    c &= 0xFFFF;
+    if (0x8000 <= c)
+	c -= 0x8000 - 0x100;
+    if (bitvec[c >> 3] & (1 << (c & 7)))
 	match = TRUE;
+    else if (0x100 <= c)
+	goto end;
     else if (flags & ANYOF_FOLD) {
 	I32 cf;
 	if (flags & ANYOF_LOCALE) {
@@ -1300,7 +1389,7 @@
 	}
 	else
 	    cf = fold[c];
-	if (p[1 + (cf >> 3)] & (1 << (cf & 7)))
+	if (bitvec[cf >> 3] & (1 << (cf & 7)))
 	    match = TRUE;
     }
 
@@ -1315,7 +1404,8 @@
 	    match = TRUE;
 	}
     }
-
+end:
+    ;
     return match ^ ((flags & ANYOF_INVERT) != 0);
 }
 
diff -urN perl5.004_02/regexp.h jperl5.004_02/regexp.h
--- perl5.004_02/regexp.h	Sat Jan 18 03:05:21 1997
+++ jperl5.004_02/regexp.h	Fri Aug  8 10:07:23 1997
@@ -35,3 +35,4 @@
 #define  ROPT_ANCH_GPOS	 2
 #define ROPT_SKIP	4
 #define ROPT_IMPLICIT	8
+#define ROPT_KANJI	16
diff -urN perl5.004_02/t/TEST jperl5.004_02/t/TEST
--- perl5.004_02/t/TEST	Sat Aug  2 00:25:01 1997
+++ jperl5.004_02/t/TEST	Fri Aug  8 10:07:23 1997
@@ -21,7 +21,7 @@
 
 if ($#ARGV == -1) {
     @ARGV = split(/[ \n]/,
-      `echo base/*.t comp/*.t cmd/*.t io/*.t; echo op/*.t pragma/*.t lib/*.t`);
+      `echo base/*.t comp/*.t cmd/*.t io/*.t; echo op/*.t pragma/*.t lib/*.t kanji/*.t`);
 }
 
 if ($^O eq 'os2' || $^O eq 'qnx') {
diff -urN perl5.004_02/t/harness jperl5.004_02/t/harness
--- perl5.004_02/t/harness	Fri Mar 21 11:50:32 1997
+++ jperl5.004_02/t/harness	Fri Aug  8 10:07:23 1997
@@ -15,5 +15,5 @@
 $Test::Harness::verbose = shift if @ARGV && $ARGV[0] eq '-v';
 
 @tests = @ARGV;
-@tests = <base/*.t comp/*.t cmd/*.t io/*.t op/*.t pragma/*.t lib/*.t> unless @tests;
+@tests = <base/*.t comp/*.t cmd/*.t io/*.t op/*.t pragma/*.t lib/*.t kanji/*.t> unless @tests;
 Test::Harness::runtests @tests;
diff -urN perl5.004_02/t/kanji/format.t jperl5.004_02/t/kanji/format.t
--- perl5.004_02/t/kanji/format.t	Thu Jan  1 09:00:00 1970
+++ jperl5.004_02/t/kanji/format.t	Fri Aug  8 10:07:23 1997
@@ -0,0 +1,23 @@
+#!./perl
+BEGIN {
+    $^H |= 0xf000;   # Turn on all japanese features.
+}
+
+
+format XXX =
+^>>>>
+$var
+^>>>>
+$var
+.
+
+open XXX, ">xxx";
+
+$var = "あいうえお";
+write XXX;
+close XXX;
+
+print "1..1\n";
+
+if (`cat xxx` eq " あい\n うえ\n") {print "ok 1\n";} else {print "not ok 1\n";}
+
diff -urN perl5.004_02/t/kanji/op.t jperl5.004_02/t/kanji/op.t
--- perl5.004_02/t/kanji/op.t	Thu Jan  1 09:00:00 1970
+++ jperl5.004_02/t/kanji/op.t	Fri Aug  8 10:07:23 1997
@@ -0,0 +1,49 @@
+#!./perl
+#
+# chop, chomp
+#
+BEGIN {
+    $^H |= 0xf000;   # Turn on all japanese features.
+}
+
+print "1..11\n";
+$y = "あいう";
+$x = chomp($y);
+if ($x ne 0 || $y ne "あいう") {print "not ok 1\n";} else {print "ok 1\n";};
+
+$y = "あいう\n";
+$x = chomp($y);
+if ($x ne 1 || $y ne "あいう") {print "not ok 2\n";} else {print "ok 2\n";};
+
+$y = "あいう";
+$x = chop($y);
+if ($x ne "う" || $y ne "あい") {print "not ok 3\n";} else {print "ok 3\n";};
+
+$y = "あいうt";
+$x = chop($y);
+if ($x ne "t" || $y ne "あいう") {print "not ok 4\n";} else {print "ok 4\n";};
+
+#
+# tr
+#
+$y = "あいうえお"; $y =~ tr/あ-う//cd;
+if ($y ne "あいう") {print "not ok 5\n";} else {print "ok 5\n";};
+$y = "あいうえお"; $y =~ tr/あ-う/か-く/;
+if ($y ne "かきくえお") {print "not ok 6\n";} else {print "ok 6\n";};
+
+$y = 'abcabcabc'; $y =~ tr/a-c/d-f/;
+if ($y ne 'defdefdef') {print "not ok 7\n";} else {print "ok 7\n";};
+$y = 'abc'; $y =~ tr/abc/def/;
+if ($y ne 'def') {print "not ok 8\n";} else {print "ok 8\n";};
+$y = 'abcabcabc'; $y =~ tr/abc/def/;
+if ($y ne 'defdefdef') {print "not ok 9\n";} else {print "ok 9\n";};
+
+#
+# split
+#
+@X = split(/(.)/, "abcde");
+if ($#X != 9) {print "not ok 10\n";} else {print "ok 10\n";};
+
+@X = split(/(.)/, "あいうtえお");
+$x = $#X;
+if ($x != 11) {print "not ok 11\n";} else {print "ok 11\n";};
diff -urN perl5.004_02/t/kanji/re_tests jperl5.004_02/t/kanji/re_tests
--- perl5.004_02/t/kanji/re_tests	Thu Jan  1 09:00:00 1970
+++ jperl5.004_02/t/kanji/re_tests	Fri Aug  8 10:07:23 1997
@@ -0,0 +1,255 @@
+あいう	あいう	y	$&	あいう
+あいう	xいう	n	-	-
+あいう	あxう	n	-	-
+あいう	あいx	n	-	-
+あいう	xあいうy	y	$&	あいう
+あいう	あいあいう	y	$&	あいう
+あい*う	あいう	y	$&	あいう
+あい*いう	あいう	y	$&	あいう
+あい*いう	あいいう	y	$&	あいいう
+あい*いう	あいいいいう	y	$&	あいいいいう
+あい{0,}いう	あいいいいう	y	$&	あいいいいう
+あい+いう	あいいう	y	$&	あいいう
+あい+いう	あいう	n	-	-
+あい+いう	あいq	n	-	-
+あい{1,}いう	あいq	n	-	-
+あい+いう	あいいいいう	y	$&	あいいいいう
+あい{1,}いう	あいいいいう	y	$&	あいいいいう
+あい{1,3}いう	あいいいいう	y	$&	あいいいいう
+あい{3,4}いう	あいいいいう	y	$&	あいいいいう
+あい{4,5}いう	あいいいいう	n	-	-
+あい?いう	あいいう	y	$&	あいいう
+あい?いう	あいう	y	$&	あいう
+あい{0,1}いう	あいう	y	$&	あいう
+あい?いう	あいいいいう	n	-	-
+あい?う	あいう	y	$&	あいう
+あい{0,1}う	あいう	y	$&	あいう
+^あいう$	あいう	y	$&	あいう
+^あいう$	あいうう	n	-	-
+^あいう	あいうう	y	$&	あいう
+^あいう$	ああいう	n	-	-
+あいう$	ああいう	y	$&	あいう
+^	あいう	y	$&	
+$	あいう	y	$&	
+あ.う	あいう	y	$&	あいう
+あ.う	あxう	y	$&	あxう
+あ.*う	あxyzう	y	$&	あxyzう
+あ.*う	あxyzえ	n	-	-
+あ[いう]え	あいう	n	-	-
+あ[いう]え	あいえ	y	$&	あいえ
+あ[い-え]e	あいえ	n	-	-
+あ[い-え]e	あうe	y	$&	あうe
+あ[い-え]	ああう	y	$&	あう
+あ[-い]	あ-	y	$&	あ-
+あ[い-]	あ-	y	$&	あ-
+あ[い-あ]	-	c	-	-
+あ[]い	-	c	-	-
+あ[	-	c	-	-
+あ]	あ]	y	$&	あ]
+あ[]]い	あ]い	y	$&	あ]い
+あ[^いう]え	あeえ	y	$&	あeえ
+あ[^いう]え	あいえ	n	-	-
+あ[^-い]う	あえう	y	$&	あえう
+あ[^-い]う	あ-う	n	-	-
+あ[^]い]う	あ]う	n	-	-
+あ[^]い]う	あえう	y	$&	あえう
+あい|うえ	あいう	y	$&	あい
+あい|うえ	あいうえ	y	$&	あい
+()ef	えef	y	$&-$1	ef-
+*あ	-	c	-	-
+(*)い	-	c	-	-
+$い	い	n	-	-
+あ\	-	c	-	-
+あ\(い	あ(い	y	$&-$1	あ(い-
+あ\(*い	あい	y	$&	あい
+あ\(*い	あ((い	y	$&	あ((い
+あ\\い	あ\い	y	$&	あ\い
+あいう)	-	c	-	-
+(あいう	-	c	-	-
+((あ))	あいう	y	$&-$1-$2	あ-あ-あ
+(あ)い(う)	あいう	y	$&-$1-$2	あいう-あ-う
+あ+い+う	ああいいあいう	y	$&	あいう
+あ{1,}い{1,}う	ああいいあいう	y	$&	あいう
+あ**	-	c	-	-
+あ.+?う	あいうあいう	y	$&	あいう
+(あ+|い)*	あい	y	$&-$1	あい-い
+(あ+|い){0,}	あい	y	$&-$1	あい-い
+(あ+|い)+	あい	y	$&-$1	あい-い
+(あ+|い){1,}	あい	y	$&-$1	あい-い
+(あ+|い)?	あい	y	$&-$1	あ-あ
+(あ+|い){0,1}	あい	y	$&-$1	あ-あ
+)(	-	c	-	-
+[^あい]*	うえe	y	$&	うえe
+あいう		n	-	-
+あ*		y	$&	
+([あいう])*え	あいいいうえ	y	$&-$1	あいいいうえ-う
+([あいう])*いうえ	あいうえ	y	$&-$1	あいうえ-あ
+あ|い|う|え|e	e	y	$&	e
+(あ|い|う|え|e)f	ef	y	$&-$1	ef-e
+あいうえ*efg	あいうえefg	y	$&	あいうえefg
+あい*	xあいyあいいいz	y	$&	あい
+あい*	xあyあいいいz	y	$&	あ
+(あい|うえ)e	あいうえe	y	$&-$1	うえe-うえ
+[あいhgefえう]ij	hij	y	$&	hij
+^(あい|うえ)e	あいうえe	n	x$1y	xy
+(あいう|)ef	あいうえef	y	$&-$1	ef-
+(あ|い)う*え	あいうえ	y	$&-$1	いうえ-い
+(あい|あい*)いう	あいう	y	$&-$1	あいう-あ
+あ([いう]*)う*	あいう	y	$&-$1	あいう-いう
+あ([いう]*)(う*え)	あいうえ	y	$&-$1-$2	あいうえ-いう-え
+あ([いう]+)(う*え)	あいうえ	y	$&-$1-$2	あいうえ-いう-え
+あ([いう]*)(う+え)	あいうえ	y	$&-$1-$2	あいうえ-い-うえ
+あ[いうえ]*えうえうえe	あえうえうえe	y	$&	あえうえうえe
+あ[いうえ]+えうえうえe	あえうえうえe	n	-	-
+(あい|あ)い*う	あいう	y	$&-$1	あいう-あい
+((あ)(い)う)(え)	あいうえ	y	$1-$2-$3-$4	あいう-あ-い-え
+[あ-おさ-そ][あ-えさ-せ]*	あえさす	y	$&	あえさす
+^あ(いう+|い[eh])g|.h$	あいh	y	$&-$1	いh-
+(いう+え$|ef*g.|h?i(j|k))	effgz	y	$&-$1-$2	effgz-effgz-
+(いう+え$|ef*g.|h?i(j|k))	ij	y	$&-$1-$2	ij-ij-j
+(いう+え$|ef*g.|h?i(j|k))	effg	n	-	-
+(いう+え$|ef*g.|h?i(j|k))	いうええ	n	-	-
+(いう+え$|ef*g.|h?i(j|k))	reffgz	y	$&-$1-$2	effgz-effgz-
+((((((((((あ))))))))))	あ	y	$10	あ
+((((((((((あ))))))))))\10	ああ	y	$&	ああ
+((((((((((あ))))))))))\41	ああ	n	-	-
+((((((((((あ))))))))))\41	あ!	y	$&	あ!
+(((((((((あ)))))))))	あ	y	$&	あ
+(.*)う(.*)	あいうえe	y	$&-$1-$2	あいうえe-あい-えe
+\((.*), (.*)\)	(あ, い)	y	($2, $1)	(い, あ)
+[k]	あい	n	-	-
+あいうえ	あいうえ	y	$&-\$&-\\$&	あいうえ-$&-\あいうえ
+あ(いう)え	あいうえ	y	$1-\$1-\\$1	いう-$1-\いう
+あ[-]?う	あう	y	$&	あう
+(あいう)\1	あいうあいう	y	$1	あいう
+([あ-う]*)\1	あいうあいう	y	$1	あいう
+'あいう'i	あいう	y	$&	あいう
+'あいう'i	Xいう	n	-	-
+'あいう'i	あXう	n	-	-
+'あいう'i	あいX	n	-	-
+'あいう'i	XあいうY	y	$&	あいう
+'あいう'i	あいあいう	y	$&	あいう
+'あい*う'i	あいう	y	$&	あいう
+'あい*いう'i	あいう	y	$&	あいう
+'あい*いう'i	あいいう	y	$&	あいいう
+'あい*?いう'i	あいいいいう	y	$&	あいいいいう
+'あい{0,}?いう'i	あいいいいう	y	$&	あいいいいう
+'あい+?いう'i	あいいう	y	$&	あいいう
+'あい+いう'i	あいう	n	-	-
+'あい+いう'i	あいQ	n	-	-
+'あい{1,}いう'i	あいQ	n	-	-
+'あい+いう'i	あいいいいう	y	$&	あいいいいう
+'あい{1,}?いう'i	あいいいいう	y	$&	あいいいいう
+'あい{1,3}?いう'i	あいいいいう	y	$&	あいいいいう
+'あい{3,4}?いう'i	あいいいいう	y	$&	あいいいいう
+'あい{4,5}?いう'i	あいいいいう	n	-	-
+'あい??いう'i	あいいう	y	$&	あいいう
+'あい??いう'i	あいう	y	$&	あいう
+'あい{0,1}?いう'i	あいう	y	$&	あいう
+'あい??いう'i	あいいいいう	n	-	-
+'あい??う'i	あいう	y	$&	あいう
+'あい{0,1}?う'i	あいう	y	$&	あいう
+'^あいう$'i	あいう	y	$&	あいう
+'^あいう$'i	あいうう	n	-	-
+'^あいう'i	あいうう	y	$&	あいう
+'^あいう$'i	ああいう	n	-	-
+'あいう$'i	ああいう	y	$&	あいう
+'$'i	あいう	y	$&	
+'あ.う'i	あいう	y	$&	あいう
+'あ.う'i	あXう	y	$&	あXう
+'あ.*?う'i	あXYZう	y	$&	あXYZう
+'あ[いう]え'i	あいう	n	-	-
+'あ[いう]え'i	あいえ	y	$&	あいえ
+'あ[い-え]'i	ああう	y	$&	あう
+'あ[-い]'i	あ-	y	$&	あ-
+'あ[い-]'i	あ-	y	$&	あ-
+'あ[い-あ]'i	-	c	-	-
+'あ[]い'i	-	c	-	-
+'あ['i	-	c	-	-
+'あ]'i	あ]	y	$&	あ]
+'あ[]]い'i	あ]い	y	$&	あ]い
+'あ[^いう]え'i	あそえ	y	$&	あそえ
+'あ[^いう]え'i	あいえ	n	-	-
+'あ[^-い]う'i	あえう	y	$&	あえう
+'あ[^-い]う'i	あ-う	n	-	-
+'あ[^]い]う'i	あ]う	n	-	-
+'あ[^]い]う'i	あえう	y	$&	あえう
+'あい|うえ'i	あいう	y	$&	あい
+'あい|うえ'i	あいうえ	y	$&	あい
+'*あ'i	-	c	-	-
+'(*)い'i	-	c	-	-
+'$い'i	い	n	-	-
+'あ\'i	-	c	-	-
+'あ\(い'i	あ(い	y	$&-$1	あ(い-
+'あ\(*い'i	あい	y	$&	あい
+'あ\(*い'i	あ((い	y	$&	あ((い
+'あいう)'i	-	c	-	-
+'(あいう'i	-	c	-	-
+'((あ))'i	あいう	y	$&-$1-$2	あ-あ-あ
+'(あ)い(う)'i	あいう	y	$&-$1-$2	あいう-あ-う
+'あ+い+う'i	ああいいあいう	y	$&	あいう
+'あ{1,}い{1,}う'i	ああいいあいう	y	$&	あいう
+'あ**'i	-	c	-	-
+'あ.+?う'i	あいうあいう	y	$&	あいう
+'あ.*?う'i	あいうあいう	y	$&	あいう
+'あ.{0,5}?う'i	あいうあいう	y	$&	あいう
+'(あ+|い)*'i	あい	y	$&-$1	あい-い
+'(あ+|い){0,}'i	あい	y	$&-$1	あい-い
+'(あ+|い)+'i	あい	y	$&-$1	あい-い
+'(あ+|い){1,}'i	あい	y	$&-$1	あい-い
+'(あ+|い)?'i	あい	y	$&-$1	あ-あ
+'(あ+|い){0,1}'i	あい	y	$&-$1	あ-あ
+'(あ+|い){0,1}?'i	あい	y	$&-$1	-
+')('i	-	う	-	-
+'[^あい]*'i	うDE	y	$&	うDE
+'あいう'i		n	-	-
+'あ*'i		y	$&	
+'([あいう])*え'i	あいいいうえ	y	$&-$1	あいいいうえ-う
+'([あいう])*いうえ'i	あいうえ	y	$&-$1	あいうえ-あ
+'あ|い|う|え|e'i	E	y	$&	E
+'(あ|い|う|え|e)f'i	EF	y	$&-$1	EF-E
+'あいうd*eふg'i	あいうDEふG	y	$&	あいうDEふG
+'あい*'i	XあいYあいいいZ	y	$&	あい
+'あい*'i	XあYあいいいZ	y	$&	あ
+'(あい|うえ)e'i	あいうえE	y	$&-$1	うえE-うえ
+'[あいhgefえう]ij'i	HIJ	y	$&	HIJ
+'^(あい|うえ)e'i	あいうDE	n	x$1y	XY
+'(あいう|)ef'i	あいうDEF	y	$&-$1	EF-
+'(あ|い)う*D'i	あいうD	y	$&-$1	いうD-い
+'(あ|い)う*D'i	あいううD	y	$&-$1	いううD-い
+'(あい|あい*)いう'i	あいう	y	$&-$1	あいう-あ
+'あ([いう]*)う*'i	あいう	y	$&-$1	あいう-いう
+'あ([いう]*)(う*D)'i	あいうD	y	$&-$1-$2	あいうD-いう-D
+'あ([いう]+)(う*D)'i	あいうD	y	$&-$1-$2	あいうD-いう-D
+'あ([いう]*)(う+D)'i	あいうD	y	$&-$1-$2	あいうD-い-うD
+'あ[いうD]*DうDうDe'i	あDうDうDE	y	$&	あDうDうDE
+'あ[いうD]+DうDうDe'i	あDうDうDE	n	-	-
+'(あい|あ)い*う'i	あいう	y	$&-$1	あいう-あい
+'((あ)(い)う)(D)'i	あいうD	y	$1-$2-$3-$4	あいう-あ-い-D
+'[a-zA-Z_][a-zA-Z0-9_]*'i	ALPHa	y	$&	ALPHa
+'^あ(いう+|い[eh])g|.h$'i	あいH	y	$&-$1	いH-
+'(いう+え$|ef*g.|h?i(j|k))'i	EFFGZ	y	$&-$1-$2	EFFGZ-EFFGZ-
+'(いう+え$|ef*g.|h?i(j|k))'i	IJ	y	$&-$1-$2	IJ-IJ-J
+'(いう+え$|ef*g.|h?i(j|k))'i	EFFG	n	-	-
+'(いう+え$|ef*g.|h?i(j|k))'i	いうDD	n	-	-
+'(いう+え$|ef*g.|h?i(j|k))'i	REFFGZ	y	$&-$1-$2	EFFGZ-EFFGZ-
+'((((((((((あ))))))))))'i	あ	y	$10	あ
+'(((((((((あ)))))))))'i	あ	y	$&	あ
+'(?:(?:(?:(?:(?:(?:(?:(?:(?:(あ))))))))))'i	あ	y	$1	あ
+'(?:(?:(?:(?:(?:(?:(?:(?:(?:(あ|い|う))))))))))'i	う	y	$1	う
+'(.*)う(.*)'i	あいうDE	y	$&-$1-$2	あいうDE-あい-DE
+'\((.*), (.*)\)'i	(あ, い)	y	($2, $1)	(い, あ)
+'[k]'i	あい	n	-	-
+'あいうD'i	あいうD	y	$&-\$&-\\$&	あいうD-$&-\あいうD
+'あ(いう)D'i	あいうD	y	$1-\$1-\\$1	いう-$1-\いう
+'あ[-]?う'i	あう	y	$&	あう
+'(あいう)\1'i	あいうあいう	y	$1	あいう
+'([あ-う]*)\1'i	あいうあいう	y	$1	あいう
+あ(?!い).	あいあえ	y	$&	あえ
+あ(?=え).	あいあえ	y	$&	あえ
+あ(?=う|え).	あいあえ	y	$&	あえ
+あ(?:い|う|え)(.)	あうお	y	$1	お
+あ(?:い|う|え)*(.)	あうe	y	$1	e
+あ(?:い|う|え)+?(.)	あうお	y	$1	お
+あ(?:い|(う|e){1,2}?|え)+?(.)	あうe	y	$1$2	うe
+^(.+)?い	あい	y	$1	あ
diff -urN perl5.004_02/t/kanji/regexp.t jperl5.004_02/t/kanji/regexp.t
--- perl5.004_02/t/kanji/regexp.t	Thu Jan  1 09:00:00 1970
+++ jperl5.004_02/t/kanji/regexp.t	Fri Aug  8 10:07:23 1997
@@ -0,0 +1,37 @@
+#!./perl
+# $RCSfile: regexp.t,v $$Revision: 4.1 $$Date: 92/08/07 18:28:20 $
+BEGIN {
+    $^H |= 0xf000;
+}
+
+open(TESTS,'kanji/re_tests') || open(TESTS,'t/op/re_tests')
+    || die "Can't open re_tests";
+while (<TESTS>) { }
+$numtests = $.;
+close(TESTS);
+
+print "1..$numtests\n";
+open(TESTS,'kanji/re_tests') || open(TESTS,'t/kanji/re_tests')
+    || die "Can't open re_tests";
+$| = 1;
+while (<TESTS>) {
+    ($pat, $subject, $result, $repl, $expect) = split(/[\t\n]/,$_);
+    $input = join(':',$pat,$subject,$result,$repl,$expect);
+    $pat = "'$pat'" unless $pat =~ /^'/;
+    eval "\$match = (\$subject =~ m$pat); \$got = \"$repl\";";
+    if ($result eq 'c') {
+	if ($@ ne '') {print "ok $.\n";} else {print "not ok $.\n";}
+    }
+    elsif ($result eq 'n') {
+	if (!$match) {print "ok $.\n";} else {print "not ok $. $input => $got\n";}
+    }
+    else {
+	if ($match && $got eq $expect) {
+	    print "ok $.\n";
+	}
+	else {
+	    print "not ok $. $input => $got\n";
+	}
+    }
+}
+close(TESTS);
diff -urN perl5.004_02/t/kanji/sjis.t jperl5.004_02/t/kanji/sjis.t
--- perl5.004_02/t/kanji/sjis.t	Thu Jan  1 09:00:00 1970
+++ jperl5.004_02/t/kanji/sjis.t	Fri Aug  8 10:07:23 1997
@@ -0,0 +1,3 @@
+#!./perl
+print "1..1\n";
+print "ok 1\n";
diff -urN perl5.004_02/toke.c jperl5.004_02/toke.c
--- perl5.004_02/toke.c	Tue Aug  5 22:26:01 1997
+++ jperl5.004_02/toke.c	Fri Aug  8 10:07:23 1997
@@ -766,18 +766,33 @@
 	if (lex_inwhat == OP_TRANS) {
 	    if (dorange) {
 		I32 i;
-		I32 max;
+		I32 max, min;
 		i = d - SvPVX(sv);
-		SvGROW(sv, SvLEN(sv) + 256);
-		d = SvPVX(sv) + i;
-		d -= 2;
-		max = (U8)d[1];
-		for (i = (U8)*d; i <= max; i++)
-		    *d++ = i;
+		
+		if ((hints & HINT_KANJI_TR)
+		    && kpart(SvPVX(sv), d-1) == KPART_KANJI_2
+		    && kpart(SvPVX(sv), d-3) == KPART_KANJI_2) {
+		    /* We have a kanji-range here */
+		    max = twochar_to_int(d[-2], d[-1]);
+		    min = twochar_to_int(d[-4], d[-3]);
+		    SvGROW(sv, SvLEN(sv) + (max-min)*2);
+		    d = SvPVX(sv) + i;
+		    d -= 4;
+		    for (i = min; i <= max; i = jnextcode(i)) {
+			*d++ = int_to_1stbyte(i);
+			*d++ = int_to_2ndbyte(i);
+		    }
+		} else {
+		    SvGROW(sv, SvLEN(sv) + 256);
+		    d = SvPVX(sv) + i;
+		    d -= 2;
+		    max = d[1] & 0377;
+		    for (i = (*d & 0377); i <= max; i++)
+		      *d++ = i;
+		}
 		dorange = FALSE;
 		continue;
-	    }
-	    else if (*s == '-' && s+1 < send  && s != start) {
+	    } else if (*s == '-' && s+1 < send  && s != start) {
 		dorange = TRUE;
 		s++;
 	    }
@@ -832,10 +847,14 @@
 	    case '4': case '5': case '6': case '7':
 		*d++ = scan_oct(s, 3, &len);
 		s += len;
+		if ((hints & HINT_KANJI_TR) && lex_inwhat == OP_TRANS)
+		    yylval.opval->op_private &= ~OPpTRANS_KANJI;
 		continue;
 	    case 'x':
 		*d++ = scan_hex(++s, 2, &len);
 		s += len;
+		if ((hints & HINT_KANJI_TR) && lex_inwhat == OP_TRANS)
+		    yylval.opval->op_private &= ~OPpTRANS_KANJI;
 		continue;
 	    case 'c':
 		s++;
@@ -867,6 +886,8 @@
 	    s++;
 	    continue;
 	}
+	if ((hints & HINT_KANJI_STRING) && iskanji(*s) && s < send)
+	    *d++ = *s++;
 	*d++ = *s++;
     }
     *d = '\0';
@@ -4732,8 +4753,8 @@
 	croak("Translation replacement not terminated");
     }
 
-    New(803,tbl,256,short);
-    op = newPVOP(OP_TRANS, 0, (char*)tbl);
+    op = newPVOP(OP_TRANS, 0, 0);
+    /* tbl is allocated at pmtrans since it's size is is no longer fixed */
 
     complement = delete = squash = 0;
     while (*s == 'c' || *s == 'd' || *s == 's') {
@@ -4741,12 +4762,13 @@
 	    complement = OPpTRANS_COMPLEMENT;
 	else if (*s == 'd')
 	    delete = OPpTRANS_DELETE;
-	else
+	else 
 	    squash = OPpTRANS_SQUASH;
 	s++;
     }
     op->op_private = delete|squash|complement;
-
+    if (hints & HINT_KANJI_TR)
+	op->op_private |= OPpTRANS_KANJI;
     lex_op = op;
     yylval.ival = OP_TRANS;
     return s;
@@ -4978,9 +5000,11 @@
 			s++;
 		    else
 			*to++ = *s++;
-		}
-		else if (*s == term)
+		} else if (*s == term) {
 		    break;
+		} else if ((hints & HINT_KANJI_STRING) && iskanji(*s) && s < bufend-1) {
+		    *to++ = *s++;
+		}
 		*to = *s;
 	    }
 	}
@@ -4998,6 +5022,8 @@
 		    break;
 		else if (*s == multi_open)
 		    brackets++;
+		else if ((hints & HINT_KANJI_STRING) && iskanji(*s) && s < bufend-1) 
+		    *to++ = *s++;
 		*to = *s;
 	    }
 	}
@@ -5190,6 +5216,11 @@
 	    eol = bufend = SvPVX(linestr) + SvCUR(linestr);
 	if (*s != '#') {
 	    for (t = s; t < eol; t++) {
+		if ((hints & HINT_KANJI_STRING) && iskanji(*t)) {
+		    t++;
+		    continue;
+		}
+
 		if (*t == '~' && t[1] == '~' && SvCUR(stuff)) {
 		    needargs = FALSE;
 		    goto enough;	/* ~~ must be first line in formline */
@@ -5223,6 +5254,8 @@
 	else
 	    lex_state = LEX_FORMLINE;
 	nextval[nexttoke].opval = (OP*)newSVOP(OP_CONST, 0, stuff);
+	nextval[nexttoke].opval->op_private |=
+	  ((hints & HINT_KANJI_FORMAT) ? 8 : 0);
 	force_next(THING);
 	nextval[nexttoke].ival = OP_FORMLINE;
 	force_next(LSTOP);
diff -urN perl5.004_02/util.c jperl5.004_02/util.c
--- perl5.004_02/util.c	Tue Jul 29 09:44:50 1997
+++ jperl5.004_02/util.c	Fri Aug  8 10:07:23 1997
@@ -323,6 +323,7 @@
 {
     register char *s, *x;
     register I32 first;
+    char *tops = big;
 
     if (!little)
 	return big;
@@ -330,6 +331,10 @@
     if (!first)
 	return big;
     while (*big) {
+	if ((hints & HINT_KANJI_STRING) && kpart(tops, big) == KPART_KANJI_2) {
+	    big++;
+	    continue;
+	}
 	if (*big++ != first)
 	    continue;
 	for (x=big,s=little; *s; /**/ ) {
@@ -358,6 +363,7 @@
     register char *s, *x;
     register I32 first = *little;
     register char *littleend = lend;
+    char *tops = big;
 
     if (!first && little >= littleend)
 	return big;
@@ -365,6 +371,10 @@
 	return Nullch;
     bigend -= littleend - little++;
     while (big <= bigend) {
+	if ((hints & HINT_KANJI_STRING) && kpart(tops, big) == KPART_KANJI_2) {
+	    big++;
+	    continue;
+	}
 	if (*big++ != first)
 	    continue;
 	for (x=big,s=little; s < littleend; /**/ ) {
@@ -398,6 +408,10 @@
     bigbeg = big;
     big = bigend - (littleend - little++);
     while (big >= bigbeg) {
+	if ((hints & HINT_KANJI_STRING) && kpart(bigbeg, big) == KPART_KANJI_2) {
+	    big--;
+	    continue;
+	}
 	if (*big-- != first)
 	    continue;
 	for (x=big+2,s=little; s < littleend; /**/ ) {
@@ -895,6 +909,7 @@
     register unsigned char *table;
     register unsigned char *olds;
     register unsigned char *oldlittle;
+    unsigned char *tops = big;
 
     if (SvTYPE(littlestr) != SVt_PVBM || !SvVALID(littlestr)) {
 	STRLEN len;
@@ -910,12 +925,18 @@
 	    return Nullch;
 	little = (unsigned char*)SvPVX(littlestr);
 	s = bigend - littlelen;
-	if (*s == *little && memEQ((char*)s,(char*)little,littlelen))
+	if (*s == *little
+	    && (!(hints & HINT_KANJI_STRING)
+		|| kpart((char *)tops,(char *)s)!=KPART_KANJI_2)
+	    && memEQ((char*)s,(char*)little,littlelen))
 	    return (char*)s;		/* how sweet it is */
 	else if (bigend[-1] == '\n' && little[littlelen-1] != '\n'
 		 && s > big) {
 	    s--;
-	    if (*s == *little && memEQ((char*)s,(char*)little,littlelen))
+	    if (*s == *little
+		&& (!(hints & HINT_KANJI_STRING)
+		    || kpart((char *)tops,(char *)s)!=KPART_KANJI_2)
+		&& memEQ((char*)s,(char*)little,littlelen))
 		return (char*)s;
 	}
 	return Nullch;
@@ -952,6 +973,14 @@
 		    goto top2;
 		return Nullch;
 	    }
+	    if ((hints & HINT_KANJI_STRING)
+		&& kpart((char *)tops,(char *)s) == 2) {
+		    s = olds + 1;
+		    little = oldlittle;
+		    if (s < bigend)
+			    goto top2;
+		    return Nullch;
+	    }
 	    return (char *)s;
 	}
     }
@@ -1019,6 +1048,8 @@
     return Nullch;
 }
 
+#define ISKANJI(c) ((hints & HINT_KANJI_REGEXP) && iskanji(c))
+
 I32
 ibcmp(s1, s2, len)
 char *s1, *s2;
@@ -1029,6 +1060,12 @@
     while (len--) {
 	if (*a != *b && *a != fold[*b])
 	    return 1;
+	if (ISKANJI(*a) && a[1] && b[1]) {
+	    len--;
+	    a++,b++;
+	    if (*a != *b)
+		return 1;
+	}
 	a++,b++;
     }
     return 0;
@@ -1044,6 +1081,12 @@
     while (len--) {
 	if (*a != *b && *a != fold_locale[*b])
 	    return 1;
+	if (ISKANJI(*a) && a[1] && b[1]) {
+	    len--;
+	    a++,b++;
+	    if (*a != *b)
+		return 1;
+	}
 	a++,b++;
     }
     return 0;
diff -urN perl5.004_02/win32/Makefile jperl5.004_02/win32/Makefile
--- perl5.004_02/win32/Makefile	Tue Jul 29 09:02:36 1997
+++ jperl5.004_02/win32/Makefile	Fri Aug  8 10:07:23 1997
@@ -163,7 +163,8 @@
 	..\taint.c	\
 	..\toke.c	\
 	..\universal.c	\
-	..\util.c
+	..\util.c	\
+	..\kanji.c
 
 CORE_OBJ= ..\av.obj	\
 	..\deb.obj	\
@@ -190,7 +191,8 @@
 	..\taint.obj	\
 	..\toke.obj	\
 	..\universal.obj\
-	..\util.obj
+	..\util.obj	\
+	..\kanji.obj
 
 WIN32_C = perllib.c \
 	win32.c \