* intern.h: prototypes for new functions; rb_cstr_to_inum(),

  rb_str_to_inum(), rb_cstr_to_dbl(), rb_str_to_dbl()

* bignum.c (rb_cstr_to_inum): changed from rb_cstr2inum(), and
  added argument badcheck to be consistent with parser. [new]

* bignum.c (rb_str_to_inum): ditto.

* bignum.c (rb_cstr2inum): wapper of rb_cstr_to_inum() now.

* bignum.c (rb_str2inum): ditto.

* object.c (rb_cstr_to_dbl): float number parser. [new]

* object.c (rb_str_to_dbl): ditto.

* object.c (rb_Float): use rb_cstr_to_dbl() for strict check.

* object.c (rb_Integer): use rb_str_to_inum() for strict check.

* string.c (rb_str_to_f): use rb_str_to_dbl() with less check.

* string.c (rb_str_to_i): use rb_str_to_inum() with less check.

* string.c (rb_str_hex): ditto.

* string.c (rb_str_oct): ditto.

* sprintf.c (rb_f_sprintf): ditto.

* time.c (obj2long): ditto.

* parse.y (yylex): use rb_cstr_to_inum() for strict check.


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@2041 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

8 files changed, 183 insertions, 91 deletions

diff --git a/ChangeLog b/ChangeLog
index 100d46cf5c..b138cdc72b 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,39 @@
+Fri Feb  1 19:10:04 2002  Nobuyoshi Nakada  <nobu.nakada@nifty.ne.jp>
+
+	* intern.h: prototypes for new functions; rb_cstr_to_inum(),
+	  rb_str_to_inum(), rb_cstr_to_dbl(), rb_str_to_dbl()
+
+	* bignum.c (rb_cstr_to_inum): changed from rb_cstr2inum(), and
+	  added argument badcheck to be consistent with parser. [new]
+
+	* bignum.c (rb_str_to_inum): ditto.
+
+	* bignum.c (rb_cstr2inum): wapper of rb_cstr_to_inum() now.
+
+	* bignum.c (rb_str2inum): ditto.
+
+	* object.c (rb_cstr_to_dbl): float number parser. [new]
+
+	* object.c (rb_str_to_dbl): ditto.
+
+	* object.c (rb_Float): use rb_cstr_to_dbl() for strict check.
+
+	* object.c (rb_Integer): use rb_str_to_inum() for strict check.
+
+	* string.c (rb_str_to_f): use rb_str_to_dbl() with less check.
+
+	* string.c (rb_str_to_i): use rb_str_to_inum() with less check.
+
+	* string.c (rb_str_hex): ditto.
+
+	* string.c (rb_str_oct): ditto.
+
+	* sprintf.c (rb_f_sprintf): ditto.
+
+	* time.c (obj2long): ditto.
+
+	* parse.y (yylex): use rb_cstr_to_inum() for strict check.
+
 Fri Feb  1 17:46:39 2002  Nobuyoshi Nakada  <nobu.nakada@nifty.ne.jp>
 
 	* regex.c (mbc_startpos): become macro.
diff --git a/bignum.c b/bignum.c
index f4add5f5f4..173c8fb968 100644
--- a/bignum.c
+++ b/bignum.c
@@ -186,21 +186,26 @@ rb_int2inum(n)
 }
 
 VALUE
-rb_cstr2inum(str, base)
+rb_cstr_to_inum(str, base, badcheck)
     const char *str;
     int base;
+    int badcheck;
 {
     const char *s = str;
     char *end;
-    int badcheck = (base==0)?1:0;
-    char sign = 1, c;
+    char sign = 1, c, nondigit = 0;
     BDIGIT_DBL num;
     long len, blen = 1;
     long i;
     VALUE z;
     BDIGIT *zds;
 
-    while (*str && ISSPACE(*str)) str++;
+    if (badcheck) {
+	while (ISSPACE(*str)) str++;
+    }
+    else {
+	while (ISSPACE(*str) || *str == '_') str++;
+    }
 
     if (str[0] == '+') {
 	str++;
@@ -213,7 +218,7 @@ rb_cstr2inum(str, base)
 	if (badcheck) goto bad;
 	return INT2FIX(0);
     }
-    if (base == 0) {
+    if (base <= 0) {
 	if (str[0] == '0') {
 	    if (str[1] == 'x' || str[1] == 'X') {
 		base = 16;
@@ -225,6 +230,9 @@ rb_cstr2inum(str, base)
 		base = 8;
 	    }
 	}
+	else if (base < -1) {
+	    base = -base;
+	}
 	else {
 	    base = 10;
 	}
@@ -290,18 +298,27 @@ rb_cstr2inum(str, base)
 	  case '0': case '1': case '2': case '3': case '4':
 	  case '5': case '6': case '7': 
 	    c = c - '0';
+	    nondigit = 0;
 	    break;
 	  case 'a': case 'b': case 'c':
 	  case 'd': case 'e': case 'f':
-	    if (base != 16) c = base;
-	    else c = c - 'a' + 10;
-	    break;
+	    c -= 'a' - 'A';
 	  case 'A': case 'B': case 'C':
 	  case 'D': case 'E': case 'F':
-	    if (base != 16) c = base;
-	    else c = c - 'A' + 10;
+	    if (base != 16) {
+		nondigit = c;
+		c = base;
+	    }
+	    else {
+		c = c - 'A' + 10;
+		nondigit = 0;
+	    }
 	    break;
 	  case '_':
+	    if (badcheck) {
+		if (nondigit) goto bad;
+		nondigit = c;
+	    }
 	    continue;
 	  default:
 	    c = base;
@@ -334,9 +351,10 @@ rb_cstr2inum(str, base)
 }
 
 VALUE
-rb_str2inum(str, base)
+rb_str_to_inum(str, base, badcheck)
     VALUE str;
     int base;
+    int badcheck;
 {
     char *s;
     int len;
@@ -351,10 +369,26 @@ rb_str2inum(str, base)
 	p[len] = '\0';
 	s = p;
     }
-    if (base == 0 && len != strlen(s)) {
+    if (badcheck && len != strlen(s)) {
 	rb_raise(rb_eArgError, "string for Integer contains null byte");
     }
-    return rb_cstr2inum(s, base); 
+    return rb_cstr_to_inum(s, base, badcheck); 
+}
+
+VALUE
+rb_cstr2inum(str, base)
+    const char *str;
+    int base;
+{
+    return rb_cstr_to_inum(str, base, base==0);
+}
+
+VALUE
+rb_str2inum(str, base)
+    VALUE str;
+    int base;
+{
+    return rb_str_to_inum(str, base, base==0);
 }
 
 static char hexmap[] = "0123456789abcdef";
diff --git a/intern.h b/intern.h
index f3c6b62355..d41086dfd0 100644
--- a/intern.h
+++ b/intern.h
@@ -61,6 +61,8 @@ VALUE rb_uint2big _((unsigned long));
 VALUE rb_int2big _((long));
 VALUE rb_uint2inum _((unsigned long));
 VALUE rb_int2inum _((long));
+VALUE rb_cstr_to_inum _((const char*, int, int));
+VALUE rb_str_to_inum _((VALUE, int, int));
 VALUE rb_cstr2inum _((const char*, int));
 VALUE rb_str2inum _((VALUE, int));
 VALUE rb_big2str _((VALUE, int));
@@ -263,6 +265,8 @@ VALUE rb_Integer _((VALUE));
 VALUE rb_Float _((VALUE));
 VALUE rb_String _((VALUE));
 VALUE rb_Array _((VALUE));
+double rb_cstr_to_dbl _((const char *, int));
+double rb_str_to_dbl _((VALUE, int));
 /* parse.y */
 EXTERN int   ruby_sourceline;
 EXTERN char *ruby_sourcefile;
diff --git a/object.c b/object.c
index f4558139e8..35517e963a 100644
--- a/object.c
+++ b/object.c
@@ -945,7 +945,7 @@ rb_Integer(val)
 	return val;
 
       case T_STRING:
-	return rb_str2inum(val, 0);
+	return rb_str_to_inum(val, 0, Qtrue);
 
       case T_FIXNUM:
 	return val;
@@ -963,6 +963,90 @@ rb_f_integer(obj, arg)
     return rb_Integer(arg);
 }
 
+double
+rb_cstr_to_dbl(p, badcheck)
+    const char *p;
+    int badcheck;
+{
+    const char *q;
+    char *end;
+    double d;
+
+    q = p;
+    if (badcheck) {
+	while (ISSPACE(*p)) p++;
+    }
+    else {
+	while (ISSPACE(*p) || *p == '_') p++;
+    }
+    d = strtod(p, &end);
+    if (p == end) {
+	if (badcheck) {
+      bad:
+	    rb_invalid_str(q, "Float()");
+	}
+	return d;
+    }
+    if (*end) {
+	char *buf = ALLOCA_N(char, strlen(p));
+	char *n = buf;
+
+	while (p < end) *n++ = *p++;
+	while (*p) {
+	    if (*p == '_') {
+		/* remove underscores between digits */
+		if (badcheck) {
+		    if (n == buf || !ISDIGIT(n[-1])) goto bad;
+		    ++p;
+		    if (!ISDIGIT(*p)) goto bad;
+		}
+		else {
+		    while (*++p == '_');
+		    continue;
+		}
+	    }
+	    *n++ = *p++;
+	}
+	*n = '\0';
+	p = buf;
+	d = strtod(p, &end);
+	if (badcheck) {
+	    if (p == end) goto bad;
+	    while (*end && ISSPACE(*end)) end++;
+	    if (*end) goto bad;
+	}
+    }
+    if (errno == ERANGE) {
+	errno = 0;
+	rb_raise(rb_eArgError, "Float %s out of range", q);
+    }
+    return d;
+}
+
+double
+rb_str_to_dbl(str, badcheck)
+    VALUE str;
+    int badcheck;
+{
+    char *s;
+    int len;
+
+    StringValue(str);
+    s = RSTRING(str)->ptr;
+    len = RSTRING(str)->len;
+    if (s[len]) {		/* no sentinel somehow */
+	char *p = ALLOCA_N(char, len+1);
+
+	MEMCPY(p, s, char, len);
+	p[len] = '\0';
+	s = p;
+    }
+    if (badcheck && len != strlen(s)) {
+	rb_raise(rb_eArgError, "string for Float contains null byte");
+    }
+    return rb_cstr_to_dbl(s, badcheck);
+}
+
 VALUE
 rb_Float(val)
     VALUE val;
@@ -978,53 +1062,7 @@ rb_Float(val)
 	return rb_float_new(rb_big2dbl(val));
 
       case T_STRING:
-        {
-	    char *q, *p, *end;
-	    double d;
-
-	    q = p = StringValuePtr(val);
-	    while (*p && ISSPACE(*p)) p++;
-	    d = strtod(p, &end);
-	    if (p == end) {
-	      bad:
-		rb_invalid_str(q, "Float()");
-	    }
-	    if (*end) {
-		if (*end == '_') {
-		    char *buf = ALLOCA_N(char, strlen(p));
-		    char *n = buf, *last = p;
-
-		    while (p < end) *n++ = *p++;
-		    while (*p) {
-			if (*p == '_' && (n > buf && ISDIGIT(n[-1]))) {
-			    /* remove underscores between digits */
-			    last = ++p;
-			    while (*p == '_') ++p;
-			    if (!ISDIGIT(*p)) {
-				while (last < p) *n++ = *last++;
-				continue;
-			    }
-			    last = p;
-			}
-			*n++ = *p++;
-		    }
-		    while (*last && (*last == '_' || ISSPACE(*last)))
-			last++;
-		    if (!*last) goto bad;
-		    *n = '\0';
-		    p = buf;
-		    d = strtod(p, &end);
-		    if (p == end) goto bad;
-		}
-		while (*end && ISSPACE(*end)) end++;
-		if (*end) goto bad;
-	    }
-	    if (errno == ERANGE) {
-		errno = 0;
-		rb_raise(rb_eArgError, "Float %s out of range", p);
-	    }
-	    return rb_float_new(d);
-	}
+	return rb_float_new(rb_str_to_dbl(val, Qtrue));
 
       case T_NIL:
 	return rb_float_new(0.0);
diff --git a/parse.y b/parse.y
index 481927ec42..0081e9fa81 100644
--- a/parse.y
+++ b/parse.y
@@ -3345,7 +3345,7 @@ yylex()
 			yyerror("hexadecimal number without hex-digits");
 		    }
 		    else if (nondigit) goto trailing_uc;
-		    yylval.val = rb_cstr2inum(tok(), 16);
+		    yylval.val = rb_cstr_to_inum(tok(), 16, Qfalse);
 		    return tINTEGER;
 		}
 		if (c == 'b' || c == 'B') {
@@ -3369,7 +3369,7 @@ yylex()
 			yyerror("numeric literal without digits");
 		    }
 		    else if (nondigit) goto trailing_uc;
-		    yylval.val = rb_cstr2inum(tok(), 2);
+		    yylval.val = rb_cstr_to_inum(tok(), 2, Qfalse);
 		    return tINTEGER;
 		}
 		if (c >= '0' && c <= '7' || c == '_') {
@@ -3388,7 +3388,7 @@ yylex()
 			pushback(c);
 			tokfix();
 			if (nondigit) goto trailing_uc;
-			yylval.val = rb_cstr2inum(tok(), 8);
+			yylval.val = rb_cstr_to_inum(tok(), 8, Qfalse);
 			return tINTEGER;
 		    }
 		}
@@ -3482,7 +3482,7 @@ yylex()
 		yylval.val = rb_float_new(d);
 		return tFLOAT;
 	    }
-	    yylval.val = rb_cstr2inum(tok(), 10);
+	    yylval.val = rb_cstr_to_inum(tok(), 10, Qfalse);
 	    return tINTEGER;
 	}
 
diff --git a/sprintf.c b/sprintf.c
index fe53d6b27c..2355a65f9c 100644
--- a/sprintf.c
+++ b/sprintf.c
@@ -18,10 +18,6 @@
 
 #define BIT_DIGITS(N)   (((N)*146)/485 + 1)  /* log2(10) =~ 146/485 */
 
-#if !defined(atof) && !defined(HAVE_STDLIB_H)
-double strtod();
-#endif
-
 static void fmt_setup _((char*,int,int,int,int));
 
 static char*
@@ -397,7 +393,7 @@ rb_f_sprintf(argc, argv)
 		    bignum = 1;
 		    break;
 		  case T_STRING:
-		    val = rb_str2inum(val, 0);
+		    val = rb_str_to_inum(val, 0, Qtrue);
 		    goto bin_retry;
 		  case T_BIGNUM:
 		    bignum = 1;
diff --git a/string.c b/string.c
index 87411a5284..25fd851d08 100644
--- a/string.c
+++ b/string.c
@@ -1766,16 +1766,14 @@ rb_str_to_i(argc, argv, str)
       default:
 	rb_raise(rb_eArgError, "illegal radix %d", base);
     }
-    return rb_str2inum(str, base);
+    return rb_str_to_inum(str, base, Qfalse);
 }
 
 static VALUE
 rb_str_to_f(str)
     VALUE str;
 {
-    double f = strtod(RSTRING(str)->ptr, 0);
-
-    return rb_float_new(f);
+    return rb_float_new(rb_str_to_dbl(str, Qfalse));
 }
 
 static VALUE
@@ -2963,28 +2961,14 @@ static VALUE
 rb_str_hex(str)
     VALUE str;
 {
-    return rb_str2inum(str, 16);
+    return rb_str_to_inum(str, 16, Qfalse);
 }
 
 static VALUE
 rb_str_oct(str)
     VALUE str;
 {
-    int base = 8;
-
-    if (RSTRING(str)->len > 2 && RSTRING(str)->ptr[0] == '0') {
-	switch (RSTRING(str)->ptr[1]) {
-	  case 'x':
-	  case 'X':
-	    base = 16;
-	    break;
-	  case 'b':
-	  case 'B':
-	    base = 2;
-	    break;
-	}
-    }
-    return rb_str2inum(str, base);
+    return rb_str_to_inum(str, -8, Qfalse);
 }
 
 static VALUE
diff --git a/time.c b/time.c
index c4c3da9bba..32b9f1118c 100644
--- a/time.c
+++ b/time.c
@@ -218,7 +218,7 @@ obj2long(obj)
     VALUE obj;
 {
     if (TYPE(obj) == T_STRING) {
-	obj = rb_str2inum(obj, 10);
+	obj = rb_str_to_inum(obj, 10, Qfalse);
     }
 
     return NUM2LONG(obj);