1.1b9_00

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/v1_1r@109 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

10 files changed, 221 insertions, 79 deletions

diff --git a/ChangeLog b/ChangeLog
index 18a1f3afdd..41c1078314 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,5 +1,18 @@
+Mon Mar  9 16:53:51 1998  Yukihiro Matsumoto  <matz@netlab.co.jp>
+
+	* sprintf.c (f_sprintf): zero padding failed for negative
+	  integers.
+
+	* sprintf.c (remove_sign_bits): failed to remove some bits.
+
+Sat Mar  7 00:43:32 1998  Yukihiro Matsumoto  <matz@netlab.co.jp>
+
+	* parse.y (logop): short cut syntax tree.
+
 Fri Mar  6 17:23:07 1998  Yukihiro Matsumoto  <matz@netlab.co.jp>
 
+	* regex.c (mbcinit): table driven mbchar detection.
+
 	* object.c (obj_alloc): check for allocating instance for the
 	  primitive classes (mostly perfect).
 
diff --git a/eval.c b/eval.c
index 73faf7fb91..4cbe54a445 100644
--- a/eval.c
+++ b/eval.c
@@ -2165,8 +2165,8 @@ rb_eval(self, node)
 			   1, INT2FIX(node->nd_mid));
 	    }
 	    if (FL_TEST(the_class, FL_SINGLETON)) {
-		VALUE recv = rb_iv_get(the_class, "__attached__");
-		rb_funcall(recv, rb_intern("singleton_method_added"),
+		rb_funcall(rb_iv_get(the_class, "__attached__"),
+			   rb_intern("singleton_method_added"),
 			   1, INT2FIX(node->nd_mid));
 	    }
 	    else {
@@ -2215,8 +2215,20 @@ rb_eval(self, node)
 
 	    body = search_method(the_class, node->nd_mid, &origin);
 	    if (!body || !body->nd_body) {
-		NameError("undefined method `%s' for class `%s'",
-			  rb_id2name(node->nd_mid), rb_class2name(the_class));
+		char *s0 = " class";
+		VALUE klass = the_class;
+
+		if (FL_TEST(the_class, FL_SINGLETON)) {
+		    VALUE obj = rb_iv_get(the_class, "__attached__");
+		    switch (TYPE(obj)) {
+		      case T_MODULE:
+		      case T_CLASS:
+			klass = obj;
+			s0 = "";
+		    }
+		}
+		NameError("undefined method `%s' for%s `%s'",
+			  rb_id2name(node->nd_mid),s0,rb_class2name(klass));
 	    }
 	    rb_clear_cache_by_id(node->nd_mid);
 	    rb_add_method(the_class, node->nd_mid, 0, NOEX_PUBLIC);
diff --git a/ext/Setup b/ext/Setup
index 4258bd4873..9e3a2474c3 100644
--- a/ext/Setup
+++ b/ext/Setup
@@ -1,7 +1,7 @@
 #option nodynamic
 
 #GD
-curses
+#curses
 #dbm
 #etc
 #fcntl
diff --git a/ext/curses/curses.c b/ext/curses/curses.c
index 127f51a0cb..c502d3faf6 100644
--- a/ext/curses/curses.c
+++ b/ext/curses/curses.c
@@ -109,9 +109,11 @@ curses_close_screen()
 static void
 curses_finalize()
 {
+    if (stdscr
 #ifdef HAVE_ISENDWIN
-    if (!isendwin())
+	&& !isendwin()
 #endif
+	)
 	endwin();
 }
 
diff --git a/parse.y b/parse.y
index 3c4a8fc3a3..47c610a7ed 100644
--- a/parse.y
+++ b/parse.y
@@ -3649,6 +3649,14 @@ logop(type, left, right)
 {
     value_expr(left);
 
+    switch (nd_type(left)) {
+      case NODE_NIL:		/* always false */
+      case NODE_FALSE:
+	return type == NODE_OR ? right : left;
+      case NODE_LIT:		/* always true */
+	return type == NODE_AND ? right : left;
+    }
+
     return node_newnode(type, cond(left), cond(right));
 }
 
diff --git a/re.c b/re.c
index 5f2b981c68..9d01157ab8 100644
--- a/re.c
+++ b/re.c
@@ -139,35 +139,33 @@ kcode_set_option(reg)
 {
     if (!FL_TEST(reg, KCODE_FIXED)) return;
 
-    re_syntax_options &= ~RE_MBCTYPE_MASK;
     switch ((RBASIC(reg)->flags & KCODE_MASK)) {
       case KCODE_NONE:
+	mbcinit(MBCTYPE_ASCII);
 	break;
       case KCODE_EUC:
-	re_syntax_options |= RE_MBCTYPE_EUC;
+	mbcinit(MBCTYPE_EUC);
 	break;
       case KCODE_SJIS:
-	re_syntax_options |= RE_MBCTYPE_SJIS;
+	mbcinit(MBCTYPE_EUC);
 	break;
     }
-    re_set_syntax(re_syntax_options);
 }	  
 
 void
 kcode_reset_option()
 {
-    re_syntax_options &= ~RE_MBCTYPE_MASK;
     switch (reg_kcode) {
       case KCODE_NONE:
+	mbcinit(MBCTYPE_ASCII);
 	break;
       case KCODE_EUC:
-	re_syntax_options |= RE_MBCTYPE_EUC;
+	mbcinit(MBCTYPE_EUC);
 	break;
       case KCODE_SJIS:
-	re_syntax_options |= RE_MBCTYPE_SJIS;
+	mbcinit(MBCTYPE_SJIS);
 	break;
     }
-    re_set_syntax(re_syntax_options);
 }
 
 extern int rb_in_eval;
@@ -997,28 +995,29 @@ void
 rb_set_kcode(code)
     char *code;
 {
-    re_syntax_options &= ~RE_MBCTYPE_MASK;
     if (code == 0) goto set_no_conversion;
 
     switch (code[0]) {
       case 'E':
       case 'e':
 	reg_kcode = KCODE_EUC;
-	re_syntax_options |= RE_MBCTYPE_EUC;
+	mbcinit(MBCTYPE_EUC);
 	break;
       case 'S':
       case 's':
 	reg_kcode = KCODE_SJIS;
-	re_syntax_options |= RE_MBCTYPE_SJIS;
+	mbcinit(MBCTYPE_SJIS);
 	break;
       default:
       case 'N':
       case 'n':
+      case 'A':
+      case 'a':
       set_no_conversion:
 	reg_kcode = KCODE_NONE;
+	mbcinit(MBCTYPE_ASCII);
 	break;
     }
-    re_set_syntax(re_syntax_options);
 }
 
 static void
@@ -1056,11 +1055,7 @@ Init_Regexp()
 		  | RE_INTERVALS
 		  | RE_NO_BK_BRACES
 		  | RE_CONTEXTUAL_INVALID_OPS
-		  | RE_BACKSLASH_ESCAPE_IN_LISTS
-#ifdef DEFAULT_MBCTYPE
-		  | DEFAULT_MBCTYPE
-#endif
-	);
+		  | RE_BACKSLASH_ESCAPE_IN_LISTS);
 
     rb_define_virtual_variable("$~", match_getter, match_setter);
     rb_define_virtual_variable("$&", last_match_getter, 0);
diff --git a/regex.c b/regex.c
index 958f384c6c..00f40e8b1c 100644
--- a/regex.c
+++ b/regex.c
@@ -73,7 +73,11 @@ char *alloca();
 #endif
 
 #define RE_ALLOCATE alloca
+#ifdef C_ALLOCA
 #define FREE_VARIABLES() alloca(0)
+#else
+#define FREE_VARIABLES() 0
+#endif
 
 #define FREE_AND_RETURN_VOID(stackb)	return
 #define FREE_AND_RETURN(stackb,val)	return(val)
@@ -113,7 +117,7 @@ char *alloca();
 	stackp = stackx + (stackp - stackb);				\
 	stackb = stackx;						\
 	stacke = stackb + 2 * len;					\
-    } while (0);									\
+    } while (0)
 
 /* Get the interface, including the syntax bits.  */
 #include "regex.h"
@@ -176,11 +180,43 @@ init_syntax_once()
    done = 1;
 }
 
-/* Sequents are missing isgraph.  */
-#ifndef isgraph
-#define isgraph(c) (isprint((c)) && !isspace((c)))
+/* Jim Meyering writes:
+
+   "... Some ctype macros are valid only for character codes that
+   isascii says are ASCII (SGI's IRIX-4.0.5 is one such system --when
+   using /bin/cc or gcc but without giving an ansi option).  So, all
+   ctype uses should be through macros like ISPRINT...  If
+   STDC_HEADERS is defined, then autoconf has verified that the ctype
+   macros don't need to be guarded with references to isascii. ...
+   Defining isascii to 1 should let any compiler worth its salt
+   eliminate the && through constant folding."  */
+#if ! defined (isascii) || defined (STDC_HEADERS)
+#undef isascii
+#define isascii(c) 1
+#endif
+
+#ifdef isblank
+#define ISBLANK(c) (isascii (c) && isblank (c))
+#else
+#define ISBLANK(c) ((c) == ' ' || (c) == '\t')
+#endif
+#ifdef isgraph
+#define ISGRAPH(c) (isascii (c) && isgraph (c))
+#else
+#define ISGRAPH(c) (isascii (c) && isprint (c) && !isspace (c))
 #endif
 
+#define ISPRINT(c) (isascii (c) && isprint (c))
+#define ISDIGIT(c) (isascii (c) && isdigit (c))
+#define ISALNUM(c) (isascii (c) && isalnum (c))
+#define ISALPHA(c) (isascii (c) && isalpha (c))
+#define ISCNTRL(c) (isascii (c) && iscntrl (c))
+#define ISLOWER(c) (isascii (c) && islower (c))
+#define ISPUNCT(c) (isascii (c) && ispunct (c))
+#define ISSPACE(c) (isascii (c) && isspace (c))
+#define ISUPPER(c) (isascii (c) && isupper (c))
+#define ISXDIGIT(c) (isascii (c) && isxdigit (c))
+
 /* These are the command codes that appear in compiled regular
    expressions, one per byte.  Some command codes are followed by
    argument bytes.  A command code can specify any interpretation
@@ -328,7 +364,7 @@ re_set_syntax(syntax)
 }
 
 /* Set by re_set_syntax to the current regexp syntax to recognize.  */
-long re_syntax_options = DEFAULT_MBCTYPE;
+long re_syntax_options = 0;
 
 
 /* Macros for re_compile_pattern, which is found below these definitions.  */
@@ -404,7 +440,7 @@ long re_syntax_options = DEFAULT_MBCTYPE;
   { if (p != pend) 							\
       { 								\
         PATFETCH(c); 							\
-	while (isdigit(c)) 						\
+	while (ISDIGIT(c)) 						\
 	  { 								\
 	    if (num < 0) 						\
 	       num = 0; 						\
@@ -1205,7 +1241,7 @@ re_compile_pattern(pattern, size, bufp)
 		      for (c = 0; c < (1 << BYTEWIDTH); c++)
 		          if (SYNTAX(c) != Sword)
 			      SET_LIST_BIT(c);
-		      if (re_syntax_options & RE_MBCTYPE_MASK) {
+		      if (current_mbctype) {
 			  set_list_bits(0x8000, 0xffff, (unsigned char*)b);
 		      }
 		      last = -1;
@@ -1213,16 +1249,16 @@ re_compile_pattern(pattern, size, bufp)
 
 		    case 's':
 		      for (c = 0; c < 256; c++)
-			  if (isspace(c))
+			  if (ISSPACE(c))
 			      SET_LIST_BIT(c);
 		      last = -1;
 		      continue;
 
 		    case 'S':
 		      for (c = 0; c < 256; c++)
-			  if (!isspace(c))
+			  if (!ISSPACE(c))
 			      SET_LIST_BIT(c);
-		      if (re_syntax_options & RE_MBCTYPE_MASK) {
+		      if (current_mbctype) {
 			  set_list_bits(0x8000, 0xffff, (unsigned char*)b);
 		      }
 		      last = -1;
@@ -1236,9 +1272,9 @@ re_compile_pattern(pattern, size, bufp)
 
 		    case 'D':
 		      for (c = 0; c < 256; c++)
-			  if (!isdigit(c))
+			  if (!ISDIGIT(c))
 			      SET_LIST_BIT(c);
-		      if (re_syntax_options & RE_MBCTYPE_MASK) {
+		      if (current_mbctype) {
 			  set_list_bits(0x8000, 0xffff, (unsigned char*)b);
 		      }
 		      last = -1;
@@ -1246,7 +1282,7 @@ re_compile_pattern(pattern, size, bufp)
 
 		    case 'x':
 		      c = scan_hex(p, 2, &numlen);
-		      if ((re_syntax_options & RE_MBCTYPE_MASK) && c > 0x7f)
+		      if (current_mbctype && c > 0x7f)
 			  c = 0xff00 | c;
 		      p += numlen;
 		      break;
@@ -1255,8 +1291,8 @@ re_compile_pattern(pattern, size, bufp)
 		    case '5': case '6': case '7': case '8': case '9':
 		      PATUNFETCH;
 		      c = scan_oct(p, 3, &numlen);
-		      if ((re_syntax_options & RE_MBCTYPE_MASK) && ismbchar(c))
-			  c = 0xff00 | c;
+		      if (ismbchar(c))
+			  c |= 0xff00;
 		      p += numlen;
 		      break;
 
@@ -1693,7 +1729,7 @@ re_compile_pattern(pattern, size, bufp)
 	      c1 = 0;
 	      c = scan_hex(p, 2, &numlen);
 	      p += numlen;
-	      if ((re_syntax_options & RE_MBCTYPE_MASK) && c > 0x7f)
+	      if (current_mbctype && c > 0x7f)
 		  c1 = 0xff;
 	      goto numeric_char;
 
@@ -1702,7 +1738,7 @@ re_compile_pattern(pattern, size, bufp)
 	      c1 = 0;
 	      c = scan_oct(p, 3, &numlen);
 	      p += numlen;
-	      if ((re_syntax_options & RE_MBCTYPE_MASK) && c > 0x7f)
+	      if (current_mbctype && c > 0x7f)
 		  c1 = 0xff;
 	      goto numeric_char;
 
@@ -1726,7 +1762,7 @@ re_compile_pattern(pattern, size, bufp)
 		      c = scan_oct(p_save, 3, &numlen);
 		      p = p_save + numlen;
 		      c1 = 0;
-		      if ((re_syntax_options & RE_MBCTYPE_MASK) && c > 0x7f)
+		      if (current_mbctype && c > 0x7f)
 			  c1 = 0xff;
 		      goto numeric_char;
 		  }
@@ -3394,3 +3430,100 @@ re_free_registers(regs)
     if (regs->beg) free(regs->beg);
     if (regs->end) free(regs->end);
 }
+
+/* Functions for multi-byte support.
+   Created for grep multi-byte extension Jul., 1993 by t^2 (Takahiro Tanimoto)
+   Last change: Jul. 9, 1993 by t^2  */
+static const unsigned char mbctab_ascii[] = {
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+};
+
+static const unsigned char mbctab_euc[] = {
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
+};
+
+static const unsigned char mbctab_sjis[] = {
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
+};
+
+#ifdef EUC
+const unsigned char *mbctab = mbctab_euc;
+int current_mbctype = MBCTYPE_EUC;
+#else
+#ifdef SJIS
+const unsigned char *mbctab = mbctab_sjis;
+int current_mbctype = MBCTYPE_SJIS;
+#else
+const unsigned char *mbctab = mbctab_ascii;
+int current_mbctype = MBCTYPE_ASCII;
+#endif
+#endif
+
+void
+#ifdef __STDC__
+mbcinit(int mbctype)
+#else
+mbcinit(mbctype)
+     int mbctype;
+#endif
+{
+  switch (mbctype) {
+  case MBCTYPE_ASCII:
+    mbctab = mbctab_ascii;
+    current_mbctype = MBCTYPE_ASCII;
+    break;
+  case MBCTYPE_EUC:
+    mbctab = mbctab_euc;
+    current_mbctype = MBCTYPE_EUC;
+    break;
+  case MBCTYPE_SJIS:
+    mbctab = mbctab_sjis;
+    current_mbctype = MBCTYPE_SJIS;
+    break;
+  }
+}
diff --git a/regex.h b/regex.h
index bf35589085..1bca032024 100644
--- a/regex.h
+++ b/regex.h
@@ -165,43 +165,17 @@ extern long re_syntax_options;
                         | RE_NO_HYPHEN_RANGE_END)
 
 /* For multi-byte char support */
-#define RE_MBCTYPE_EUC (1L << 20)
-#define RE_MBCTYPE_SJIS (1L << 21)
-#define RE_MBCTYPE_MASK (RE_MBCTYPE_EUC | RE_MBCTYPE_SJIS)
-
-#ifdef EUC
-#define DEFAULT_MBCTYPE RE_MBCTYPE_EUC
-#else
-#ifdef SJIS
-#define DEFAULT_MBCTYPE RE_MBCTYPE_SJIS
-#else
-#define DEFAULT_MBCTYPE 0
-#endif
-#endif
+#define MBCTYPE_ASCII 0
+#define MBCTYPE_EUC 1
+#define MBCTYPE_SJIS 2
+
+extern const unsigned char *mbctab;
+extern int current_mbctype;
+
+void mbcinit (int);
 
 #undef ismbchar
-#define ismbchar(c) \
-  (re_syntax_options & RE_MBCTYPE_EUC		\
-   ? (   0xa1 <= (unsigned char) (c)		\
-      && (unsigned char) (c) <= 0xfe)		\
-   : (re_syntax_options & RE_MBCTYPE_SJIS	\
-      ? ((   0x81 <= (unsigned char) (c)	\
-	  && (unsigned char) (c) <= 0x9f)	\
-	||  ((0xe0 <= (unsigned char) (c))	\
-	  && (unsigned char) (c) <= 0xef))	\
-      : 0))
-
-#undef ismbchar2
-#define ismbchar2(c) \
-  (re_syntax_options & RE_MBCTYPE_EUC		\
-   ? (   0xa1 <= (unsigned char) (c)		\
-      && (unsigned char) (c) <= 0xfe)		\
-   : (re_syntax_options & RE_MBCTYPE_SJIS	\
-      ? ((   0x40 <= (unsigned char) (c)	\
-	  && (unsigned char) (c) <= 0x7e)	\
-	||  ((0x80 <= (unsigned char) (c))	\
-	  && (unsigned char) (c) <= 0xfc))	\
-      : 0))
+#define ismbchar(c) mbctab[(unsigned char)(c)]
 
 /* This data structure is used to represent a compiled pattern.  */
 
diff --git a/sprintf.c b/sprintf.c
index 57ce682b5e..d8de7d5126 100644
--- a/sprintf.c
+++ b/sprintf.c
@@ -28,10 +28,10 @@ remove_sign_bits(str, base)
       x_retry:
 	switch (*t) {
 	  case 'c':
-	    *t = '8';
+	    *t = '4';
 	    break;
 	  case 'd':
-	    *t = '9';
+	    *t = '5';
 	    break;
 	  case 'e':
 	    *t = '2';
@@ -418,6 +418,7 @@ f_sprintf(argc, argv)
 		    int n = slen-len;
 		    char d = ' ';
 		    if (flags & FZERO) d = '0';
+		    if (s[0] == '.') d = '.';
 		    CHECK(n);
 		    while (n--) {
 			buf[blen++] = d;
diff --git a/string.c b/string.c
index 2b6295cfbc..6bd63c096c 100644
--- a/string.c
+++ b/string.c
@@ -1375,11 +1375,15 @@ str_inspect(str)
 
     while (p < pend) {
 	UCHAR c = *p++;
-	if (ismbchar(c) && p < pend && ismbchar2(*p)) {
+	if (ismbchar(c) && p < pend) {
 	    CHECK(2);
 	    *b++ = c;
 	    *b++ = *p++;
 	}
+	if (c & 0x80) {
+	    CHECK(1);
+	    *b++ = c;
+	}
 	else if (c == '"') {
 	    CHECK(2);
 	    *b++ = '\\';