Change the multibyte character set support so that

Kanji support is one case of multibyte character
support, rather than being a specific case in
single byte character support.

This allows us to add Big5 Chinese support (code page 950)
and Korean Hangul support (code page 949) at very little
cost. Also allows us to easily add future multibyte
code pages.

Makefile: Added codepages 949, 950 as we now support more multibyte
codepages.
asyncdns.c: Fixed problem with child being re-spawned when parent killed.
charcnv.c
charset.c
client.c
clitar.c
kanji.c
kanji.h
smb.h
util.c
loadparm.c: Generic multibyte codepage support (adding Big5 Chinese
            and Korean Hangul).
nmbd.c: Fixed problem with child being re-spawned when parent killed.
mangle.c: Modified str_checksum so that first 15 characters have more
          effect on outcome. This helps with short name mangling as
          most 'long' names are still shorter than 15 chars (bug was
          foobar_mng and foobar_sum would hash to the same value, with
          the modified code they hash differently.
Jeremy.
(This used to be commit 299016338cfb47f0c585875ef9b468121fcee97d)

15 files changed, 593 insertions, 218 deletions

diff --git a/source3/client/client.c b/source3/client/client.c
index 798dfe577e..988b6685a8 100644
--- a/source3/client/client.c
+++ b/source3/client/client.c
@@ -130,13 +130,8 @@ extern int Client;
 
 #define USENMB
 
-static BOOL setup_term_code(char *code)
-{
-	interpret_coding_system(code);
-	return True;
-}
-#define CNV_LANG(s) dos2unix_format(s,False)
-#define CNV_INPUT(s) unix2dos_format(s,True)
+#define CNV_LANG(s) dos_to_unix(s,False)
+#define CNV_INPUT(s) unix_to_dos(s,True)
 
 /****************************************************************************
 send an SMBclose on an SMB file handle
@@ -3756,15 +3751,7 @@ static void usage(char *pname)
 
   codepage_initialise(lp_client_code_page());
 
-  if(lp_client_code_page() == KANJI_CODEPAGE)
-  {
-        if (!setup_term_code (term_code))
-    {
-            DEBUG(0, ("%s: unknown terminal code name\n", optarg));
-            usage (pname);
-            exit (1);
-        }
-  }
+  interpret_coding_system(term_code);
 
   if (*workgroup == 0)
     strcpy(workgroup,lp_workgroup());
diff --git a/source3/client/clitar.c b/source3/client/clitar.c
index 6a811f41c7..ccaab486d0 100644
--- a/source3/client/clitar.c
+++ b/source3/client/clitar.c
@@ -322,33 +322,25 @@ static void fixtarname(char *tptr, char *fp, int l)
    * to lovely unix /'s :-} */
 
   *tptr++='.';
-  if(lp_client_code_page() == KANJI_CODEPAGE)
-  {
-    while (l > 0) {
-      if (is_shift_jis (*fp)) {
+
+  while (l > 0) {
+    int skip;
+    if((skip = skip_multibyte_char( *fp)) != 0) {
+      if (skip == 2) {
         *tptr++ = *fp++;
         *tptr++ = *fp++;
         l -= 2;
-      } else if (is_kana (*fp)) {
-        *tptr++ = *fp++;
-        l--;
-      } else if (*fp == '\\') {
-        *tptr++ = '/';
-        fp++;
-        l--;
-      } else {
+      } else if (skip == 1) {
         *tptr++ = *fp++;
         l--;
       }
-    }
-  }
-  else
-  {
-    while (l--)
-    {
-      *tptr=(*fp == '\\') ? '/' : *fp;
-      tptr++;
+    } else if (*fp == '\\') {
+      *tptr++ = '/';
       fp++;
+      l--;
+    } else {
+      *tptr++ = *fp++;
+      l--;
     }
   }
 }
@@ -1227,33 +1219,24 @@ static void unfixtarname(char *tptr, char *fp, int l)
   if (*fp == '.') fp++;
   if (*fp == '\\' || *fp == '/') fp++;
 
-  if(lp_client_code_page() == KANJI_CODEPAGE)
-  {
-    while (l > 0) {
-      if (is_shift_jis (*fp)) {
+  while (l > 0) {
+    int skip;
+    if(( skip = skip_multibyte_char( *fp )) != 0) {
+      if (skip == 2) {
         *tptr++ = *fp++;
         *tptr++ = *fp++;
         l -= 2;
-      } else if (is_kana (*fp)) {
-        *tptr++ = *fp++;
-        l--;
-      } else if (*fp == '/') {
-        *tptr++ = '\\';
-        fp++;
-        l--;
-      } else {
+      } else if (skip == 1) {
         *tptr++ = *fp++;
         l--;
       }
-    }
-  }
-  else
-  {
-    while (l--)
-    {
-      *tptr=(*fp == '/') ? '\\' : *fp;
-      tptr++;
+    } else if (*fp == '/') {
+      *tptr++ = '\\';
       fp++;
+      l--;
+    } else {
+      *tptr++ = *fp++;
+      l--;
     }
   }
 }
diff --git a/source3/client/smbmount.c b/source3/client/smbmount.c
index 847c4b1f3d..e5902ff0d9 100644
--- a/source3/client/smbmount.c
+++ b/source3/client/smbmount.c
@@ -136,13 +136,8 @@ extern int Client;
 
 #define USENMB
 
-static BOOL setup_term_code(char *code)
-{
-	interpret_coding_system(code);
-	return True;
-}
-#define CNV_LANG(s) dos2unix_format(s,False)
-#define CNV_INPUT(s) unix2dos_format(s,True)
+#define CNV_LANG(s) dos_to_unix(s,False)
+#define CNV_INPUT(s) unix_to_dos(s,True)
 
 /****************************************************************************
 check for existance of a dir
@@ -879,15 +874,7 @@ static void usage(char *pname)
 
   codepage_initialise(lp_client_code_page());
 
-  if(lp_client_code_page() == KANJI_CODEPAGE)
-  {
-        if (!setup_term_code (term_code))
-    {
-            DEBUG(0, ("%s: unknown terminal code name\n", optarg));
-            usage (pname);
-            exit (1);
-        }
-  }
+  interpret_coding_system(term_code);
 
   if (*workgroup == 0)
     strcpy(workgroup,lp_workgroup());
diff --git a/source3/include/kanji.h b/source3/include/kanji.h
index 101b98cfa3..302db13a27 100644
--- a/source3/include/kanji.h
+++ b/source3/include/kanji.h
@@ -103,10 +103,13 @@
 #define bin2hex(x)						      \
     ( (((int) (x)) >= 10)? (((int) (x))-10 + (int) 'a'): (((int) (x)) + (int) '0') )
 
-#else /* not _KANJI_C_ */
+/* For Hangul (Korean - code page 949). */
+#define is_hangul(c) ((0x81 <= ((unsigned char) (c)) && ((unsigned char) (c)) <= 0xfd))
 
-extern char *(*_dos_to_unix)(char *str, BOOL overwrite);
-extern char *(*_unix_to_dos)(char *str, BOOL overwrite);
+/* For traditional Chinese (known as Big5 encoding - code page 950). */
+#define is_big5_c1(c) ((0xa1 <= ((unsigned char) (c)) && ((unsigned char) (c)) <= 0xf9)) 
+
+#else /* not _KANJI_C_ */
 
 /*
  * The following is needed for AIX systems that have
@@ -130,12 +133,24 @@ extern char *(*_unix_to_dos)(char *str, BOOL overwrite);
 #undef strtok
 #endif /* strtok */
 
-/* Ensure we use our definitions. */
+/* Ensure we use our definitions in all other files than kanji.c. */
 
-#define strchr sj_strchr
-#define strrchr sj_strrchr
-#define strstr sj_strstr
-#define strtok sj_strtok
+/* Function pointers we will replace. */
+extern char *(*multibyte_strchr)(char *s, int c);
+extern char *(*multibyte_strrchr)(char *s, int c);
+extern char *(*multibyte_strstr)(char *s1, char *s2);
+extern char *(*multibyte_strtok)(char *s1, char *s2);
+extern char *(*_dos_to_unix)(char *str, BOOL overwrite);
+extern char *(*_unix_to_dos)(char *str, BOOL overwrite);
+extern BOOL (*is_multibyte_char)(char c);
+
+#define strchr(s1, c) ((*multibyte_strchr)((s1), (c)))
+#define strrchr(s1, c) ((*multibyte_strrchr)((s1), (c)))
+#define strstr(s1, s2) ((*multibyte_strstr)((s1), (s2)))
+#define strtok(s1, s2) ((*multibyte_strtok)((s1), (s2)))
+#define dos_to_unix(x,y) ((*_dos_to_unix)((x), (y)))
+#define unix_to_dos(x,y) ((*_unix_to_dos)((x), (y)))
+#define skip_multibyte_char(c) ((*is_multibyte_char)((c)))
 
 #endif /* _KANJI_C_ */
 
@@ -149,7 +164,4 @@ extern char *(*_unix_to_dos)(char *str, BOOL overwrite);
 #define CAP_CODE (6)
 #define DOSV_CODE SJIS_CODE
 
-#define unix_to_dos(x,y) unix2dos_format(x,y)
-#define dos_to_unix(x,y) dos2unix_format(x,y)
-
 #endif /* _KANJI_H_ */
diff --git a/source3/include/proto.h b/source3/include/proto.h
index 802d9973df..47ef5812ca 100644
--- a/source3/include/proto.h
+++ b/source3/include/proto.h
@@ -9,12 +9,14 @@ BOOL allow_access(char *deny_list,char *allow_list,char *cname,char *caddr);
 /*The following definitions come from  asyncdns.c  */
 
 int asyncdns_fd(void);
+void kill_async_dns_child();
 void start_async_dns(void);
 void run_dns_queue(void);
 BOOL queue_dns_query(struct packet_struct *p,struct nmb_name *question,
 		     struct name_record **n);
 BOOL queue_dns_query(struct packet_struct *p,struct nmb_name *question,
 		     struct name_record **n);
+void kill_async_dns_child();
 
 /*The following definitions come from  cgi.c  */
 
@@ -206,11 +208,8 @@ int reply_trans(char *inbuf,char *outbuf, int size, int bufsize);
 
 /*The following definitions come from  kanji.c  */
 
-char *sj_strtok(char *s1, char *s2);
-char *sj_strstr(char *s1, char *s2);
-char *sj_strchr (char *s, int c);
-char *sj_strrchr(char *s, int c);
-int interpret_coding_system(char *str);
+void interpret_coding_system(char *str);
+void initialize_multibyte_vectors( int client_codepage);
 
 /*The following definitions come from  loadparm.c  */
 
diff --git a/source3/include/smb.h b/source3/include/smb.h
index 77f4006c4a..f7a134d797 100644
--- a/source3/include/smb.h
+++ b/source3/include/smb.h
@@ -2099,7 +2099,10 @@ enum case_handling {CASE_LOWER,CASE_UPPER};
 #endif 
 
 /* Defines needed for multi-codepage support. */
+#define MSDOS_LATIN_1_CODEPAGE 850
 #define KANJI_CODEPAGE 932
+#define HANGUL_CODEPAGE 949
+#define BIG5_CODEPAGE 950
 
 #ifdef KANJI
 /* 
@@ -2110,7 +2113,7 @@ enum case_handling {CASE_LOWER,CASE_UPPER};
 /* 
  * Default client code page - 850 - Western European 
  */
-#define DEFAULT_CLIENT_CODE_PAGE 850
+#define DEFAULT_CLIENT_CODE_PAGE MSDOS_LATIN_1_CODEPAGE
 #endif /* KANJI */
 
 /* 
diff --git a/source3/lib/charcnv.c b/source3/lib/charcnv.c
index f02fcb2f92..20db58e4ab 100644
--- a/source3/lib/charcnv.c
+++ b/source3/lib/charcnv.c
@@ -164,9 +164,6 @@ char *unix2dos_format(char *str,BOOL overwrite)
 
     if (!mapsinited) initmaps();
 
-    if(lp_client_code_page() == KANJI_CODEPAGE)
-      return (*_unix_to_dos)(str, overwrite);
-    else {
       if (overwrite) {
           for (p = str; *p; p++) *p = unix2dos[(unsigned char)*p];
           return str;
@@ -175,7 +172,6 @@ char *unix2dos_format(char *str,BOOL overwrite)
           *dp = 0;
           return cvtbuf;
       }
-    }
 }
 
 /*
@@ -188,9 +184,6 @@ char *dos2unix_format(char *str, BOOL overwrite)
 
     if (!mapsinited) initmaps();
 
-    if(lp_client_code_page() == KANJI_CODEPAGE)
-      return (*_dos_to_unix)(str, overwrite);
-    else {
       if (overwrite) {
           for (p = str; *p; p++) *p = dos2unix[(unsigned char)*p];
           return str;
@@ -199,7 +192,6 @@ char *dos2unix_format(char *str, BOOL overwrite)
           *dp = 0;
           return cvtbuf;
       }
-    }
 }
 
 
diff --git a/source3/lib/charset.c b/source3/lib/charset.c
index 79a82f8587..fe170bdcf5 100644
--- a/source3/lib/charset.c
+++ b/source3/lib/charset.c
@@ -347,14 +347,21 @@ void codepage_initialise(int client_codepage)
 for code page %d failed. Using default client codepage 932\n", 
              CODEPAGEDIR, client_codepage, client_codepage));
     cp = cp_932;
+    client_codepage = KANJI_CODEPAGE;
 #else /* KANJI */
     DEBUG(6,("codepage_initialise: loading dynamic codepage file %s/codepage.%d \
 for code page %d failed. Using default client codepage 850\n", 
              CODEPAGEDIR, client_codepage, client_codepage));
     cp = cp_850;
+    client_codepage = MSDOS_LATIN_1_CODEPAGE;
 #endif /* KANJI */
   }
 
+  /*
+   * Setup the function pointers for the loaded codepage.
+   */
+  initialize_multibyte_vectors( client_codepage );
+
   if(cp)
   {
     for(i = 0; !((cp[i][0] == '\0') && (cp[i][1] == '\0')); i++)
diff --git a/source3/lib/kanji.c b/source3/lib/kanji.c
index d63798914e..9360405547 100644
--- a/source3/lib/kanji.c
+++ b/source3/lib/kanji.c
@@ -27,6 +27,55 @@
 #define _KANJI_C_
 #include "includes.h"
 
+/*
+ * Function pointers that get overridden when multi-byte code pages
+ * are loaded.
+ */
+
+char *(*multibyte_strchr)(char *, int ) = (char *(*)(char *, int )) strchr;
+char *(*multibyte_strrchr)(char *, int ) = (char *(*)(char *, int )) strrchr;
+char *(*multibyte_strstr)(char *, char *) = (char *(*)(char *, char *)) strstr;
+char *(*multibyte_strtok)(char *, char *) = (char *(*)(char *, char *)) strtok;
+
+/*
+ * Kanji is treated differently here due to historical accident of
+ * it being the first non-English codepage added to Samba.
+ * The define 'KANJI' is being overloaded to mean 'use kanji codepage
+ * by default' and also 'this is the filename-to-disk conversion 
+ * method to use'. This really should be removed and all control
+ * over this left in the smb.conf parameters 'client codepage'
+ * and 'coding system'.
+ */
+
+#ifndef KANJI
+
+/*
+ * Set the default conversion to be the functions in
+ * charcnv.c.
+ */
+
+static int not_multibyte_char(char);
+
+char *(*_dos_to_unix)(char *, BOOL) = dos2unix_format;
+char *(*_unix_to_dos)(char *, BOOL) = unix2dos_format;
+int (*is_multibyte_char)(char) = not_multibyte_char;
+
+#else /* KANJI */
+
+/*
+ * Set the default conversion to be the function
+ * sj_to_sj in this file.
+ */
+
+static char *sj_to_sj(char *from, BOOL overwrite);
+static int kanji_multibyte_char(char);
+
+char *(*_dos_to_unix)(char *, BOOL) = sj_to_sj;
+char *(*_unix_to_dos)(char *, BOOL) = sj_to_sj;
+int (*is_multibyte_char)(char) = kanji_multibyte_char;
+
+#endif /* KANJI */
+
 /* jis si/so sequence */
 static char jis_kso = JIS_KSO;
 static char jis_ksi = JIS_KSI;
@@ -37,13 +86,10 @@ static char hex_tag = HEXTAG;
 ********************************************************************/
 /*******************************************************************
  search token from S1 separated any char of S2
- S1 contain SHIFT JIS chars.
+ S1 contains SHIFT JIS chars.
 ********************************************************************/
-char *sj_strtok(char *s1, char *s2)
+static char *sj_strtok(char *s1, char *s2)
 {
-  if (lp_client_code_page() != KANJI_CODEPAGE) {
-   return strtok(s1, s2);
-  } else {
     static char *s = NULL;
     char *q;
     if (!s1) {
@@ -75,18 +121,14 @@ char *sj_strtok(char *s1, char *s2)
 	return q;
     }
     return NULL;
-  }
 }
 
 /*******************************************************************
  search string S2 from S1
- S1 contain SHIFT JIS chars.
+ S1 contains SHIFT JIS chars.
 ********************************************************************/
-char *sj_strstr(char *s1, char *s2)
+static char *sj_strstr(char *s1, char *s2)
 {
-  if (lp_client_code_page() != KANJI_CODEPAGE) {
-    return strstr(s1, s2);
-  } else {
     int len = strlen ((char *) s2);
     if (!*s2) 
 	return (char *) s1;
@@ -102,18 +144,14 @@ char *sj_strstr(char *s1, char *s2)
 	}
     }
     return 0;
-  }
 }
 
 /*******************************************************************
  Search char C from beginning of S.
- S contain SHIFT JIS chars.
+ S contains SHIFT JIS chars.
 ********************************************************************/
-char *sj_strchr (char *s, int c)
+static char *sj_strchr (char *s, int c)
 {
-  if (lp_client_code_page() != KANJI_CODEPAGE) {
-    return strchr(s, c);
-  } else {
     for (; *s; ) {
 	if (*s == c)
 	    return (char *) s;
@@ -124,18 +162,14 @@ char *sj_strchr (char *s, int c)
 	}
     }
     return 0;
-  }
 }
 
 /*******************************************************************
  Search char C end of S.
- S contain SHIFT JIS chars.
+ S contains SHIFT JIS chars.
 ********************************************************************/
-char *sj_strrchr(char *s, int c)
+static char *sj_strrchr(char *s, int c)
 {
-  if (lp_client_code_page() != KANJI_CODEPAGE) {
-    return strrchr(s, c);
-  } else {
     char *q;
 
     for (q = 0; *s; ) {
@@ -149,7 +183,249 @@ char *sj_strrchr(char *s, int c)
 	}
     }
     return q;
+}
+
+/*******************************************************************
+ Kanji multibyte char function.
+*******************************************************************/
+   
+static int kanji_multibyte_char(char c)
+{
+  if(is_shift_jis(c)) {
+    return 2;
+  } else if (is_kana(c)) {
+    return 1;
+  }
+  return 0;
+}
+
+/*******************************************************************
+  Hangul (Korean - code page 949) functions
+********************************************************************/
+/*******************************************************************
+ search token from S1 separated any char of S2
+ S1 contains hangul chars.
+********************************************************************/
+static char *hangul_strtok(char *s1, char *s2)
+{
+    static char *s = NULL;
+    char *q;
+    if (!s1) {
+        if (!s) {
+            return NULL;
+        }
+        s1 = s;
+    }
+    for (q = s1; *s1; ) {
+        if (is_hangul (*s1)) {
+            s1 += 2;
+        } else {
+            char *p = strchr (s2, *s1);
+            if (p) {
+                if (s1 != q) {
+                    s = s1 + 1;
+                    *s1 = '\0';
+                    return q;
+                }
+                q = s1 + 1;
+            }
+            s1++;
+        }
+    }
+    s = NULL;
+    if (*q) {
+        return q;
+    }
+    return NULL;
+}
+
+/*******************************************************************
+ search string S2 from S1
+ S1 contains hangul chars.
+********************************************************************/
+static char *hangul_strstr(char *s1, char *s2)
+{
+    int len = strlen ((char *) s2);
+    if (!*s2)
+        return (char *) s1;
+    for (;*s1;) {
+        if (*s1 == *s2) {
+            if (strncmp (s1, s2, len) == 0)
+                return (char *) s1;
+        }
+        if (is_hangul (*s1)) {
+            s1 += 2;
+        } else {
+            s1++;
+        }
+    }
+    return 0;
+}
+
+/*******************************************************************
+ Search char C from beginning of S.
+ S contains hangul chars.
+********************************************************************/
+static char *hangul_strchr (char *s, int c)
+{
+    for (; *s; ) {
+        if (*s == c)
+            return (char *) s;
+        if (is_hangul (*s)) {
+            s += 2;
+        } else {
+            s++;
+        }
+    }
+    return 0;
+}
+
+/*******************************************************************
+ Search char C end of S.
+ S contains hangul chars.
+********************************************************************/
+static char *hangul_strrchr(char *s, int c)
+{
+    char *q;
+ 
+    for (q = 0; *s; ) {
+        if (*s == c) {
+            q = (char *) s;
+        }
+        if (is_hangul (*s)) {
+            s += 2;
+        } else {
+            s++;
+        }
+    }
+    return q;
+}
+
+/*******************************************************************
+ Hangul multibyte char function.
+*******************************************************************/
+
+static int hangul_multibyte_char(char c)
+{
+  if( is_hangul(c)) {
+    return 2;
+  }
+  return 0;
+}
+
+/*******************************************************************
+  Big5 Traditional Chinese (code page 950) functions
+********************************************************************/
+
+/*******************************************************************
+ search token from S1 separated any char of S2
+ S1 contains big5 chars.
+********************************************************************/
+static char *big5_strtok(char *s1, char *s2)
+{
+    static char *s = NULL;
+    char *q;
+    if (!s1) {
+        if (!s) {
+            return NULL;
+        }
+        s1 = s;
+    }
+    for (q = s1; *s1; ) {
+        if (is_big5_c1 (*s1)) {
+            s1 += 2;
+        } else {
+            char *p = strchr (s2, *s1);
+            if (p) {
+                if (s1 != q) {
+                    s = s1 + 1;
+                    *s1 = '\0';
+                    return q;
+                }
+                q = s1 + 1;
+            }
+            s1++;
+        }
+    }
+    s = NULL;
+    if (*q) {
+        return q;
+    }
+    return NULL;
+}
+
+/*******************************************************************
+ search string S2 from S1
+ S1 contains big5 chars.
+********************************************************************/
+static char *big5_strstr(char *s1, char *s2)
+{
+    int len = strlen ((char *) s2);
+    if (!*s2)
+        return (char *) s1;
+    for (;*s1;) {
+        if (*s1 == *s2) {
+            if (strncmp (s1, s2, len) == 0)
+                return (char *) s1;
+        }
+        if (is_big5_c1 (*s1)) {
+            s1 += 2;
+        } else {
+            s1++;
+        }
+    }
+    return 0;
+}
+
+/*******************************************************************
+ Search char C from beginning of S.
+ S contains big5 chars.
+********************************************************************/
+static char *big5_strchr (char *s, int c)
+{
+    for (; *s; ) {
+        if (*s == c)
+            return (char *) s;
+        if (is_big5_c1 (*s)) {
+            s += 2;
+        } else {
+            s++;
+        }
+    }
+    return 0;
+}
+
+/*******************************************************************
+ Search char C end of S.
+ S contains big5 chars.
+********************************************************************/
+static char *big5_strrchr(char *s, int c)
+{
+    char *q;
+ 
+    for (q = 0; *s; ) {
+        if (*s == c) {
+            q = (char *) s;
+        }
+        if (is_big5_c1 (*s)) {
+            s += 2;
+        } else {
+            s++;
+        }
+    }
+    return q;
+}
+
+/*******************************************************************
+ Big5 multibyte char function.
+*******************************************************************/
+
+static int big5_multibyte_char(char c)
+{
+  if( is_big5_c1(c)) {
+    return 2;
   }
+  return 0;
 }
 
 /*******************************************************************
@@ -770,17 +1046,17 @@ static char *sj_to_sj(char *from, BOOL overwrite)
  _dos_to_unix		_unix_to_dos
 ************************************************************************/
 
-char *(*_dos_to_unix)(char *str, BOOL overwrite) = sj_to_sj;
-char *(*_unix_to_dos)(char *str, BOOL overwrite) = sj_to_sj;
-
-static int setup_string_function(int codes)
+static void setup_string_function(int codes)
 {
     switch (codes) {
     default:
+        _dos_to_unix = dos2unix_format;
+        _unix_to_dos = unix2dos_format;
+        break;
+
     case SJIS_CODE:
 	_dos_to_unix = sj_to_sj;
 	_unix_to_dos = sj_to_sj;
-
 	break;
 	
     case EUC_CODE:
@@ -813,13 +1089,12 @@ static int setup_string_function(int codes)
 	_unix_to_dos = cap_to_sj;
 	break;
     }
-    return codes;
 }
 
 /*
  * Interpret coding system.
  */
-int interpret_coding_system(char *str)
+void interpret_coding_system(char *str)
 {
     int codes = UNKNOWN_CODE;
     
@@ -909,5 +1184,58 @@ int interpret_coding_system(char *str)
 	jis_kso = '@';
 	jis_ksi = 'H';
     }	
-    return setup_string_function (codes);
+    setup_string_function (codes);
+}
+
+/*******************************************************************
+ Non multibyte char function.
+*******************************************************************/
+   
+static int not_multibyte_char(char c)
+{
+  return 0;
+}
+
+/*******************************************************************
+ Setup the function pointers for the functions that are replaced
+ when multi-byte codepages are used.
+
+ The dos_to_unix and unix_to_dos function pointers are only
+ replaced by setup_string_function called by interpret_coding_system
+ above.
+*******************************************************************/
+
+void initialize_multibyte_vectors( int client_codepage)
+{
+  switch( client_codepage )
+  {
+  case KANJI_CODEPAGE:
+    multibyte_strchr = (char *(*)(char *, int )) sj_strchr;
+    multibyte_strrchr = (char *(*)(char *, int )) sj_strrchr;
+    multibyte_strstr = (char *(*)(char *, char *)) sj_strstr;
+    multibyte_strtok = (char *(*)(char *, char *)) sj_strtok;
+    is_multibyte_char = kanji_multibyte_char;
+    break;
+  case HANGUL_CODEPAGE:
+    multibyte_strchr = (char *(*)(char *, int )) hangul_strchr;
+    multibyte_strrchr = (char *(*)(char *, int )) hangul_strrchr;
+    multibyte_strstr = (char *(*)(char *, char *)) hangul_strstr;
+    multibyte_strtok = (char *(*)(char *, char *)) hangul_strtok;
+    is_multibyte_char = hangul_multibyte_char;
+    break;
+  case BIG5_CODEPAGE:
+    multibyte_strchr = (char *(*)(char *, int )) big5_strchr;
+    multibyte_strrchr = (char *(*)(char *, int )) big5_strrchr;
+    multibyte_strstr = (char *(*)(char *, char *)) big5_strstr;
+    multibyte_strtok = (char *(*)(char *, char *)) big5_strtok;
+    is_multibyte_char = big5_multibyte_char;
+    break;
+  default:
+    multibyte_strchr = (char *(*)(char *, int )) strchr;
+    multibyte_strrchr = (char *(*)(char *, int )) strrchr;
+    multibyte_strstr = (char *(*)(char *, char *)) strstr;
+    multibyte_strtok = (char *(*)(char *, char *)) strtok;
+    is_multibyte_char = not_multibyte_char;
+    break; 
+  }
 }
diff --git a/source3/lib/util.c b/source3/lib/util.c
index 18614caeed..5af41cc06c 100644
--- a/source3/lib/util.c
+++ b/source3/lib/util.c
@@ -887,6 +887,15 @@ int StrCaseCmp(char *s, char *t)
      asynchronous upper to lower mapping.
    */
 #if !defined(KANJI_WIN95_COMPATIBILITY)
+  /*
+   * For completeness we should put in equivalent code for code pages
+   * 949 (Korean hangul) and 950 (Big5 Traditional Chinese) here - but
+   * doubt anyone wants Samba to behave differently from Win95 and WinNT
+   * here. They both treat full width ascii characters as case senstive
+   * filenames (ie. they don't do the work we do here).
+   * JRA.
+   */
+
   if(lp_client_code_page() == KANJI_CODEPAGE)
   {
     /* Win95 treats full width ascii characters as case sensitive. */
@@ -951,6 +960,15 @@ int StrnCaseCmp(char *s, char *t, int n)
      asynchronous upper to lower mapping.
    */
 #if !defined(KANJI_WIN95_COMPATIBILITY)
+  /*
+   * For completeness we should put in equivalent code for code pages
+   * 949 (Korean hangul) and 950 (Big5 Traditional Chinese) here - but
+   * doubt anyone wants Samba to behave differently from Win95 and WinNT
+   * here. They both treat full width ascii characters as case senstive
+   * filenames (ie. they don't do the work we do here).
+   * JRA. 
+   */
+
   if(lp_client_code_page() == KANJI_CODEPAGE)
   {
     /* Win95 treats full width ascii characters as case sensitive. */
@@ -1058,6 +1076,15 @@ void strlower(char *s)
   while (*s)
   {
 #if !defined(KANJI_WIN95_COMPATIBILITY)
+  /*
+   * For completeness we should put in equivalent code for code pages
+   * 949 (Korean hangul) and 950 (Big5 Traditional Chinese) here - but
+   * doubt anyone wants Samba to behave differently from Win95 and WinNT
+   * here. They both treat full width ascii characters as case senstive
+   * filenames (ie. they don't do the work we do here).
+   * JRA. 
+   */
+
     if(lp_client_code_page() == KANJI_CODEPAGE)
     {
       /* Win95 treats full width ascii characters as case sensitive. */
@@ -1096,6 +1123,15 @@ void strupper(char *s)
   while (*s)
   {
 #if !defined(KANJI_WIN95_COMPATIBILITY)
+  /*
+   * For completeness we should put in equivalent code for code pages
+   * 949 (Korean hangul) and 950 (Big5 Traditional Chinese) here - but
+   * doubt anyone wants Samba to behave differently from Win95 and WinNT
+   * here. They both treat full width ascii characters as case senstive
+   * filenames (ie. they don't do the work we do here).
+   * JRA. 
+   */
+
     if(lp_client_code_page() == KANJI_CODEPAGE)
     {
       /* Win95 treats full width ascii characters as case sensitive. */
@@ -1157,6 +1193,15 @@ void string_replace(char *s,char oldc,char newc)
   while (*s)
   {
 #if !defined(KANJI_WIN95_COMPATIBILITY)
+  /*
+   * For completeness we should put in equivalent code for code pages
+   * 949 (Korean hangul) and 950 (Big5 Traditional Chinese) here - but
+   * doubt anyone wants Samba to behave differently from Win95 and WinNT
+   * here. They both treat full width ascii characters as case senstive
+   * filenames (ie. they don't do the work we do here).
+   * JRA. 
+   */
+
     if(lp_client_code_page() == KANJI_CODEPAGE)
     {
       /* Win95 treats full width ascii characters as case sensitive. */
@@ -1783,6 +1828,15 @@ BOOL strhasupper(char *s)
   while (*s) 
   {
 #if !defined(KANJI_WIN95_COMPATIBILITY)
+  /*
+   * For completeness we should put in equivalent code for code pages
+   * 949 (Korean hangul) and 950 (Big5 Traditional Chinese) here - but
+   * doubt anyone wants Samba to behave differently from Win95 and WinNT
+   * here. They both treat full width ascii characters as case senstive
+   * filenames (ie. they don't do the work we do here).
+   * JRA. 
+   */
+
     if(lp_client_code_page() == KANJI_CODEPAGE)
     {
       /* Win95 treats full width ascii characters as case sensitive. */
@@ -1816,6 +1870,15 @@ BOOL strhaslower(char *s)
   while (*s) 
   {
 #if !defined(KANJI_WIN95_COMPATIBILITY)
+  /*
+   * For completeness we should put in equivalent code for code pages
+   * 949 (Korean hangul) and 950 (Big5 Traditional Chinese) here - but
+   * doubt anyone wants Samba to behave differently from Win95 and WinNT
+   * here. They both treat full width ascii characters as case senstive
+   * filenames (ie. they don't do the work we do here).
+   * JRA. 
+   */
+
     if(lp_client_code_page() == KANJI_CODEPAGE)
     {
       /* Win95 treats full width ascii characters as case sensitive. */
@@ -1857,6 +1920,15 @@ int count_chars(char *s,char c)
   int count=0;
 
 #if !defined(KANJI_WIN95_COMPATIBILITY)
+  /*
+   * For completeness we should put in equivalent code for code pages
+   * 949 (Korean hangul) and 950 (Big5 Traditional Chinese) here - but
+   * doubt anyone wants Samba to behave differently from Win95 and WinNT
+   * here. They both treat full width ascii characters as case senstive
+   * filenames (ie. they don't do the work we do here).
+   * JRA. 
+   */
+
   if(lp_client_code_page() == KANJI_CODEPAGE)
   {
     /* Win95 treats full width ascii characters as case sensitive. */
diff --git a/source3/nmbd/asyncdns.c b/source3/nmbd/asyncdns.c
index 3b71369d67..3fb16a08e9 100644
--- a/source3/nmbd/asyncdns.c
+++ b/source3/nmbd/asyncdns.c
@@ -103,7 +103,8 @@ static void asyncdns_process(void)
 }
 
 /**************************************************************************** **
-  catch a sigterm
+  catch a sigterm (in the child process - the parent has a different handler
+  see nmbd.c for details).
   We need a separate term handler here so we don't release any 
   names that our parent is going to release, or overwrite a 
   WINS db that our parent is going to write.
@@ -117,6 +118,17 @@ static int sig_term()
 }
 
 /***************************************************************************
+ Called by the parent process when it receives a SIGTERM - also kills the
+ child so we don't get child async dns processes lying around, causing trouble.
+  ****************************************************************************/
+
+void kill_async_dns_child()
+{
+  if(child_pid != 0 && child_pid != -1)
+    kill(child_pid, SIGTERM);
+}
+
+/***************************************************************************
   create a child process to handle DNS lookups
   ****************************************************************************/
 void start_async_dns(void)
@@ -326,4 +338,12 @@ BOOL queue_dns_query(struct packet_struct *p,struct nmb_name *question,
           send_wins_name_query_response(0, p, *n);
 	return False;
 }
+
+/***************************************************************************
+ With sync dns there is no child to kill on SIGTERM.
+  ****************************************************************************/
+void kill_async_dns_child()
+{
+  return;
+}
 #endif
diff --git a/source3/nmbd/nmbd.c b/source3/nmbd/nmbd.c
index d42580bcbd..1a12e0eec0 100644
--- a/source3/nmbd/nmbd.c
+++ b/source3/nmbd/nmbd.c
@@ -74,6 +74,9 @@ static int sig_term()
   /* Announce all server entries as 0 time-to-live, 0 type. */
   announce_my_servers_removed();
 
+  /* If there was an async dns child - kill it. */
+  kill_async_dns_child();
+
   exit(0);
 
   /* Keep compiler happy.. */
diff --git a/source3/param/loadparm.c b/source3/param/loadparm.c
index 0bb2056745..c0f54860c1 100644
--- a/source3/param/loadparm.c
+++ b/source3/param/loadparm.c
@@ -50,6 +50,17 @@
 #include "includes.h"
 
 /* Set default coding system for KANJI if none specified in Makefile. */
+/* 
+ * We treat KANJI specially due to historical precedent (it was the
+ * first non-english codepage added to Samba). With the new dynamic
+ * codepage support this is not needed anymore.
+ *
+ * The define 'KANJI' is being overloaded to mean 'use kanji codepage
+ * by default' and also 'this is the filename-to-disk conversion 
+ * method to use'. This really should be removed and all control
+ * over this left in the smb.conf parameters 'client codepage'
+ * and 'coding system'.
+ */
 #ifndef KANJI
 #define KANJI "sjis"
 #endif /* KANJI */
@@ -721,7 +732,6 @@ static void init_globals(void)
   Globals.bNISHomeMap = False;
   string_set(&Globals.szNISHomeMapName, "auto.home");
 #endif
-  interpret_coding_system(KANJI);
   Globals.client_code_page = DEFAULT_CLIENT_CODE_PAGE;
   Globals.bTimeServer = False;
   Globals.bBindInterfacesOnly = False;
@@ -748,6 +758,13 @@ static void init_globals(void)
   Globals.bWINSproxy = False;
 
   Globals.bDNSproxy = True;
+
+  /*
+   * This must be done last as it checks the value in 
+   * client_code_page.
+   */
+
+  interpret_coding_system(KANJI);
 }
 
 /***************************************************************************
diff --git a/source3/smbd/mangle.c b/source3/smbd/mangle.c
index 7b6a563511..d0d5ada247 100644
--- a/source3/smbd/mangle.c
+++ b/source3/smbd/mangle.c
@@ -42,7 +42,7 @@ int str_checksum(char *s)
   while( *s )
     {
     c = *s;
-    res ^= (c << (i % 15)) ^ (c >> (15-(i%15)));
+    res ^= (c << (15-(i%15))) ^ (c >> (i % 15));
     s++; i++;
     }
   return(res);
@@ -130,38 +130,25 @@ BOOL is_8_3(char *fname, BOOL check_case)
 
   dot_pos = strchr(fname,'.');
 
-    {
+  {
     char *p = fname;
+    int skip;
 
-    if(lp_client_code_page() == KANJI_CODEPAGE)
-      {
-      dot_pos = 0;
-      while (*p)
-        {
-        if (is_shift_jis (*p)) 
-          p += 2;
-        else if (is_kana (*p)) 
-          p ++;
-        else 
-          {
-          if (*p == '.' && !dot_pos)
-            dot_pos = (char *) p;
-          if (!isdoschar(*p))
-            return(False);
-          p++;
-          }
-        }
-      }      
-    else
+    dot_pos = 0;
+    while (*p)
+    {
+      if((skip = skip_multibyte_char( *p )) != 0)
+        p += skip;
+      else 
       {
-      while (*p)
-        {
+        if (*p == '.' && !dot_pos)
+          dot_pos = (char *) p;
         if (!isdoschar(*p))
           return(False);
         p++;
-        }      
       }
     }
+  }      
 
   /* no dot and less than 9 means OK */
   if (!dot_pos)
@@ -542,6 +529,7 @@ void mangle_name_83(char *s)
   char base[9];
   int baselen = 0;
   int extlen = 0;
+  int skip;
 
   extension[0]=0;
   base[0]=0;
@@ -572,40 +560,29 @@ void mangle_name_83(char *s)
       *p++ = 0;
       while (*p && extlen < 3)
         {
-        if(lp_client_code_page() == KANJI_CODEPAGE)
+        skip = skip_multibyte_char(*p);
+        if (skip == 2)
           {
-          if (is_shift_jis (*p))
+          if (extlen < 2)
             {
-            if (extlen < 2)
-              {
-              extension[extlen++] = p[0];
-              extension[extlen++] = p[1];
-              }
-            else 
-              {
-              extension[extlen++] = base36 (((unsigned char) *p) % 36);
-              }
-            p += 2;
+            extension[extlen++] = p[0];
+            extension[extlen++] = p[1];
             }
-          else
+          else 
             {
-            if( is_kana (*p) )
-              {
-              extension[extlen++] = p[0];
-              p++;
-              }
-            else 
-              {
-              if (isdoschar (*p) && *p != '.')
-                extension[extlen++] = p[0];
-              p++;
-              }
+            extension[extlen++] = base36 (((unsigned char) *p) % 36);
             }
+          p += 2;
           }
-        else
+        else if( skip == 1 )
           {
-          if (isdoschar(*p) && *p != '.')
-            extension[extlen++] = *p;
+          extension[extlen++] = p[0];
+          p++;
+          }
+        else 
+          {
+          if (isdoschar (*p) && *p != '.')
+            extension[extlen++] = p[0];
           p++;
           }
         }
@@ -617,9 +594,8 @@ void mangle_name_83(char *s)
 
   while (*p && baselen < 5)
     {
-    if(lp_client_code_page() == KANJI_CODEPAGE)
-      {
-      if (is_shift_jis (*p))
+      skip = skip_multibyte_char(*p);
+      if (skip == 2)
         {
         if (baselen < 4)
           {
@@ -632,27 +608,17 @@ void mangle_name_83(char *s)
           }
         p += 2;
         }
-      else
+      else if( skip == 1)
         {
-        if( is_kana (*p) )
-          {
+        base[baselen++] = p[0];
+        p++;
+        }
+      else 
+        {
+        if (isdoschar (*p) && *p != '.')
           base[baselen++] = p[0];
-          p++;
-          }
-        else 
-          {
-          if (isdoschar (*p) && *p != '.')
-            base[baselen++] = p[0];
-          p++;
-          }
+        p++;
         }
-      }
-    else
-      {
-      if (isdoschar(*p) && *p != '.')
-        base[baselen++] = *p;
-      p++;
-      }
     }
   base[baselen] = 0;
 
@@ -679,6 +645,7 @@ static BOOL illegal_name(char *name)
   static unsigned char illegal[256];
   static BOOL initialised=False;
   unsigned char *s;
+  int skip;
 
   if( !initialised )
     {
@@ -690,26 +657,19 @@ static BOOL illegal_name(char *name)
       illegal[*s] = True;
     }
 
-  if(lp_client_code_page() == KANJI_CODEPAGE)
+  for (s = (unsigned char *)name; *s;)
     {
-    for (s = (unsigned char *)name; *s;)
+    skip = skip_multibyte_char( *s );
+    if (skip != 0)
+      s += skip;
+    else
       {
-      if (is_shift_jis (*s))
-        s += 2;
+      if (illegal[*s])
+        return(True);
       else
-        {
-        if (illegal[*s])
-          return(True);
-        else
-          s++;
-        }
+        s++;
       }
     }
-  else
-    {
-    for (s = (unsigned char *)name;*s;s++)
-      if (illegal[*s]) return(True);
-    }
 
   return(False);
   } /* illegal_name */
diff --git a/source3/smbd/trans2.c b/source3/smbd/trans2.c
index 35272ae3d9..3cd6c138c8 100644
--- a/source3/smbd/trans2.c
+++ b/source3/smbd/trans2.c
@@ -908,11 +908,14 @@ resume_key = %d resume name = %s continue=%d level = %d\n",
     start_pos = TellDir(dirptr);
     for(current_pos = start_pos; current_pos >= 0; current_pos--)
     {
+      DEBUG(7,("call_trans2findnext: seeking to pos %d\n", current_pos));
+
       SeekDir(dirptr, current_pos);
       dname = ReadDirName(dirptr);
       if(dname && strcsequal( resume_name, dname))
       {
         SeekDir(dirptr, current_pos+1);
+        DEBUG(7,("call_trans2findnext: got match at pos %d\n", current_pos+1 ));
         break;
       }
     }
@@ -923,12 +926,14 @@ resume_key = %d resume name = %s continue=%d level = %d\n",
 
     if(current_pos < 0)
     {
+      DEBUG(7,("call_trans2findnext: notfound: seeking to pos %d\n", start_pos));
       SeekDir(dirptr, start_pos);
       for(current_pos = start_pos; (dname = ReadDirName(dirptr)) != NULL; SeekDir(dirptr,++current_pos))
       {
         if(strcsequal( resume_name, dname))
         {
           SeekDir(dirptr, current_pos+1);
+          DEBUG(7,("call_trans2findnext: got match at pos %d\n", current_pos+1 ));
           break;
         }
       } /* end for */