Kanji fixes for upper/lower case conversion with sjis characters.

Code from Takashi Fujita <fujita@ainix.isac.co.jp>.
Jeremy (jallison@whistle.com)

diff --git a/source/include/kanji.h b/source/include/kanji.h
index ee3ba7e09a68c70a1cccab7ac0c03a37d4d8b97e..865ac3b2c90f2358d9846dd8661b4f27cff03e1b 100644 (file)
--- a/source/include/kanji.h
+++ b/source/include/kanji.h
@@ -22,6 +22,7 @@
      and extend coding system to EUC/SJIS/JIS/HEX at 1994.10.11
      and add all jis codes sequence at 1995.8.16
      Notes: Hexadecimal code by <ohki@gssm.otuka.tsukuba.ac.jp>
+     and add upper/lower case conversion 1997.8.21
 */
 #ifndef _KANJI_H_
 #define _KANJI_H_
@@ -37,6 +38,22 @@
     && ((unsigned char) (c)) != 0x7f)
 #define is_kana(c) ((0xa0 <= ((unsigned char) (c)) && ((unsigned char) (c)) <= 0xdf))
 
+/* case conversion */
+#define is_sj_upper2(c) \
+  ((0x60 <= (unsigned char) (c)) && ((unsigned char) (c) <= 0x79))
+#define is_sj_lower2(c) \
+  ((0x81 <= (unsigned char) (c)) && ((unsigned char) (c) <= 0x9A))
+#define sjis_alph 0x82
+#define is_sj_alph(c) (sjis_alph == (unsigned char) (c))
+#define is_sj_upper(c1, c2) (is_sj_alph (c1) && is_sj_upper2 (c2))
+#define is_sj_lower(c1, c2) (is_sj_alph (c1) && is_sj_lower2 (c2))
+#define sj_toupper2(c) \
+    (is_sj_lower2 (c) ? ((int) ((unsigned char) (c) - 0x81 + 0x60)) : \
+     ((int) (unsigned char) (c)))
+#define sj_tolower2(c) \
+    (is_sj_upper2 (c) ? ((int) ((unsigned char) (c) - 0x60 + 0x81)) : \
+     ((int) (unsigned char) (c)))
+
 #ifdef _KANJI_C_
 /* FOR EUC CODE */
 #define euc_kana (0x8e)

diff --git a/source/include/proto.h b/source/include/proto.h
index f716c1f9a92eb68c156be935cf107962d9113eeb..9f05e5bbffddf943fc8465c8bfa44728753282b0 100644 (file)
--- a/source/include/proto.h
+++ b/source/include/proto.h
@@ -794,6 +794,7 @@ int sys_chdir(char *dname);
 int sys_utime(char *fname,struct utimbuf *times);
 int sys_rename(char *from, char *to);
 int sys_chmod(char *fname,int mode);
+char *sys_getwd(char *s);
 int sys_chown(char *fname,int uid,int gid);
 int sys_chroot(char *dname);
 struct hostent *sys_gethostbyname(char *name);

diff --git a/source/lib/system.c b/source/lib/system.c
index ea86e9ccaabc239bc972a1f7f91d1838dfa01f11..39f845b30ebaa28acc7f9003d2a0130df4cc205f 100644 (file)
--- a/source/lib/system.c
+++ b/source/lib/system.c
@@ -331,6 +331,22 @@ int sys_chmod(char *fname,int mode)
   return(chmod(dos_to_unix(fname,False),mode));
 }
 
+/*******************************************************************
+for getwd
+********************************************************************/
+char *sys_getwd(char *s)
+{
+  char *wd;
+#ifdef USE_GETCWD
+  wd = (char *) getcwd (s, sizeof (pstring));
+#else
+  wd = (char *) getwd (s);
+#endif
+  if (wd)
+    unix_to_dos (wd, True);
+  return wd;
+}
+
 /*******************************************************************
 chown isn't used much but OS/2 doesn't have it
 ********************************************************************/

diff --git a/source/lib/util.c b/source/lib/util.c
index a048c8b3a7555c67048e7cc252aafc44cce5035f..6f6e03fbd60f0b813ffc16f9d27a3e5f7ba09e9f 100644 (file)
--- a/source/lib/util.c
+++ b/source/lib/util.c
@@ -808,12 +808,52 @@ int StrCaseCmp(const char *s, const char *t)
   /* We *must* use toupper rather than tolower here due to the
      asynchronous upper to lower mapping.
    */
+#ifdef KANJI
+  int diff;
+  for (;;)
+  {
+    if (!*s || !*t)
+         return toupper (*s) - toupper (*t);
+    else if (is_sj_alph (*s) && is_sj_alph (*t))
+       {
+         diff = sj_toupper2 (*(s+1)) - sj_toupper2 (*(t+1));
+         if (diff)
+           return diff;
+         s += 2;
+         t += 2;
+       }
+    else if (is_shift_jis (*s) && is_shift_jis (*t))
+       {
+         diff = ((int) (unsigned char) *s) - ((int) (unsigned char) *t);
+         if (diff)
+           return diff;
+         diff = ((int) (unsigned char) *(s+1)) - ((int) (unsigned char) *(t+1));
+         if (diff)
+           return diff;
+         s += 2;
+         t += 2;
+       }
+    else if (is_shift_jis (*s))
+         return 1;
+    else if (is_shift_jis (*t))
+         return -1;
+    else 
+       {
+         diff = toupper (*s) - toupper (*t);
+         if (diff)
+           return diff;
+         s++;
+         t++;
+       }
+  }
+#else /* KANJI */
   while (*s && *t && toupper(*s) == toupper(*t))
   {
     s++; t++;
   }
 
   return(toupper(*s) - toupper(*t));
+#endif /* KANJI */
 }
 
 /*******************************************************************
@@ -825,6 +865,49 @@ int StrnCaseCmp(const char *s, const char *t, int n)
   /* We *must* use toupper rather than tolower here due to the
      asynchronous upper to lower mapping.
    */
+#ifdef KANJI
+  int diff;
+  for (;n > 0;)
+  {
+    if (!*s || !*t)
+         return toupper (*s) - toupper (*t);
+    else if (is_sj_alph (*s) && is_sj_alph (*t))
+       {
+         diff = sj_toupper2 (*(s+1)) - sj_toupper2 (*(t+1));
+         if (diff)
+           return diff;
+         s += 2;
+         t += 2;
+         n -= 2;
+       }
+    else if (is_shift_jis (*s) && is_shift_jis (*t))
+       {
+         diff = ((int) (unsigned char) *s) - ((int) (unsigned char) *t);
+         if (diff)
+           return diff;
+         diff = ((int) (unsigned char) *(s+1)) - ((int) (unsigned char) *(t+1));
+         if (diff)
+           return diff;
+         s += 2;
+         t += 2;
+         n -= 2;
+       }
+    else if (is_shift_jis (*s))
+         return 1;
+    else if (is_shift_jis (*t))
+         return -1;
+    else 
+       {
+         diff = toupper (*s) - toupper (*t);
+         if (diff)
+           return diff;
+         s++;
+         t++;
+         n--;
+       }
+  }
+  return 0;
+#else /* KANJI */
   while (n-- && *s && *t && toupper(*s) == toupper(*t))
   {
     s++; t++;
@@ -835,6 +918,7 @@ int StrnCaseCmp(const char *s, const char *t, int n)
 
   /* identical up to where we run out of chars, and strings are same length */
   return(0);
+#endif /* KANJI */
 }
 
 /*******************************************************************
@@ -880,6 +964,9 @@ void strlower(char *s)
     {
 #ifdef KANJI
        if (is_shift_jis (*s)) {
+           if (is_sj_upper (s[0], s[1])) {
+                 s[1] = sj_tolower2 (s[1]);
+           }
            s += 2;
        } else if (is_kana (*s)) {
            s++;
@@ -888,7 +975,7 @@ void strlower(char *s)
                *s = tolower(*s);
            s++;
        }
-#else
+#else /* KANJI */
       if (isupper(*s))
          *s = tolower(*s);
       s++;
@@ -905,6 +992,9 @@ void strupper(char *s)
     {
 #ifdef KANJI
        if (is_shift_jis (*s)) {
+           if (is_sj_lower (s[0], s[1])) {
+                 s[1] = sj_toupper2 (s[1]);
+           }
            s += 2;
        } else if (is_kana (*s)) {
            s++;
@@ -913,11 +1003,11 @@ void strupper(char *s)
                *s = toupper(*s);
            s++;
        }
-#else
+#else /* KANJI */
       if (islower(*s))
        *s = toupper(*s);
       s++;
-#endif
+#endif /* KANJI */
     }
 }
 
@@ -961,7 +1051,7 @@ void string_replace(char *s,char oldc,char newc)
                *s = newc;
            s++;
        }
-#else
+#else /* KANJI */
       if (oldc == *s)
        *s = newc;
       s++;
@@ -1259,27 +1349,6 @@ int ChDir(char *path)
   return(res);
 }
 
-
-/*******************************************************************
-  return the absolute current directory path. A dumb version.
-********************************************************************/
-static char *Dumb_GetWd(char *s)
-{
-  char *p;
-#ifdef USE_GETCWD
-  p = (char *)getcwd(s,sizeof(pstring));
-#else
-  p = (char *)getwd(s));
-#endif
-  if(!p)
-    return NULL;
-
-  /* Ensure we always return in dos format. */
-  unix_to_dos(p,True);
-  return p;
-}
-
-
 /* number of list structures for a caching GetWd function. */
 #define MAX_GETWDCACHE (50)
 
@@ -1306,7 +1375,7 @@ char *GetWd(char *str)
   *s = 0;
 
   if (!use_getwd_cache)
-    return(Dumb_GetWd(str));
+    return(sys_getwd(str));
 
   /* init the cache */
   if (!getwd_cache_init)
@@ -1325,7 +1394,7 @@ char *GetWd(char *str)
   if (stat(".",&st) == -1) 
     {
       DEBUG(0,("Very strange, couldn't stat \".\"\n"));
-      return(Dumb_GetWd(str));
+      return(sys_getwd(str));
     }
 
 
@@ -1369,7 +1438,7 @@ char *GetWd(char *str)
       The very slow getcwd, which spawns a process on some systems, or the
       not quite so bad getwd. */
 
-  if (!Dumb_GetWd(s))
+  if (!sys_getwd(s))
     {
       DEBUG(0,("Getwd failed, errno %d\n",errno));
       return (NULL);
@@ -1619,7 +1688,7 @@ BOOL strhasupper(char *s)
            if (isupper(*s)) return(True);
            s++;
        }
-#else 
+#else /* KANJI */
       if (isupper(*s)) return(True);
       s++;
 #endif /* KANJI */
@@ -1636,6 +1705,8 @@ BOOL strhaslower(char *s)
     {
 #ifdef KANJI
        if (is_shift_jis (*s)) {
+           if (is_sj_upper (s[0], s[1])) return(True);
+           if (is_sj_lower (s[0], s[1])) return (True);
            s += 2;
        } else if (is_kana (*s)) {
            s++;
@@ -1643,7 +1714,7 @@ BOOL strhaslower(char *s)
            if (islower(*s)) return(True);
            s++;
        }
-#else 
+#else /* KANJI */
       if (islower(*s)) return(True);
       s++;
 #endif /* KANJI */
@@ -1657,12 +1728,26 @@ find the number of chars in a string
 int count_chars(char *s,char c)
 {
   int count=0;
+#ifdef KANJI
+  while (*s)
+  {
+    if (is_shift_jis (*s))
+         s += 2;
+    else 
+       {
+         if (*s == c)
+           count++;
+         s++;
+       }
+  }
+#else /* KANJI */
   while (*s) 
     {
       if (*s == c)
        count++;
       s++;
     }
+#endif /* KANJI */
   return(count);
 }