* string.c (rb_str_comparable): need not to check asciicompat here.

* encoding.c (rb_enc_check): ditto.

* string.c (rb_enc_str_coderange): tuned a bit; no broken check.

* encoding.c (rb_enc_check): new encoding comparison criteria.

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@13547 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

3 files changed, 49 insertions, 35 deletions

diff --git a/ChangeLog b/ChangeLog
index 2bb7667358..07c76e8abb 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -65,6 +65,16 @@ Fri Sep 28 12:49:05 2007  Koichi Sasada  <ko1@atdot.net>
 
 	* benchmark/driver.rb: fix file selection algorithm.
 
+Fri Sep 28 02:05:42 2007  Yukihiro Matsumoto  <matz@ruby-lang.org>
+
+	* string.c (rb_str_comparable): need not to check asciicompat here.
+
+	* encoding.c (rb_enc_check): ditto.
+
+	* string.c (rb_enc_str_coderange): tuned a bit; no broken check.
+
+	* encoding.c (rb_enc_check): new encoding comparison criteria.
+
 Thu Sep 27 17:36:28 2007  NAKAMURA Usaku  <usa@ruby-lang.org>
 
 	* win32/REAMDE.win32: follow recent changes.
diff --git a/encoding.c b/encoding.c
index bb9e77b5bd..4df874292f 100644
--- a/encoding.c
+++ b/encoding.c
@@ -196,25 +196,27 @@ rb_enc_check(VALUE str1, VALUE str2)
 	return rb_enc_from_index(idx1);
     }
 
-    if (idx1 == 0) {
-	enc = rb_enc_from_index(idx2);
-	if (rb_enc_asciicompat(enc)) {
-	    return enc;
-	}
+    if (BUILTIN_TYPE(str1) != T_STRING) {
+	VALUE tmp = str1;
+	str1 = str2;
+	str2 = tmp;
     }
-    else if (idx2 == 0) {
-	enc = rb_enc_from_index(idx1);
-	if (rb_enc_asciicompat(enc)) {
-	    return enc;
+    if (BUILTIN_TYPE(str1) == T_STRING) {
+	int cr1, cr2;
+
+	cr1 = rb_enc_str_coderange(str1);
+	if (BUILTIN_TYPE(str2) == T_STRING) {
+	    cr2 = rb_enc_str_coderange(str2);
+	    if (cr1 != cr2) {
+		/* may need to handle ENC_CODERANGE_BROKEN */
+		if (cr1 == ENC_CODERANGE_SINGLE) return rb_enc_from_index(idx2);
+		if (cr2 == ENC_CODERANGE_SINGLE) return rb_enc_from_index(idx1);
+	    }
+	    if (cr1 == ENC_CODERANGE_SINGLE) return ONIG_ENCODING_ASCII;
 	}
-    }
-    if (BUILTIN_TYPE(str1) == T_STRING &&
-	BUILTIN_TYPE(str2) == T_STRING &&
-	rb_enc_asciicompat(rb_enc_from_index(idx1)) &&
-	rb_enc_asciicompat(rb_enc_from_index(idx2)) &&
-	rb_enc_str_coderange(str1) == ENC_CODERANGE_SINGLE &&
-	rb_enc_str_coderange(str2) == ENC_CODERANGE_SINGLE) {
-	return ONIG_ENCODING_ASCII;
+	if (cr1 == ENC_CODERANGE_SINGLE &&
+	    rb_enc_asciicompat(enc = rb_enc_from_index(idx2)))
+	    return enc;
     }
     rb_raise(rb_eArgError, "character encodings differ");
 }
diff --git a/string.c b/string.c
index 3f09196093..ab48a66bb6 100644
--- a/string.c
+++ b/string.c
@@ -99,26 +99,32 @@ VALUE rb_fs;
 int
 rb_enc_str_coderange(VALUE str)
 {
-    long i;
     int cr = ENC_CODERANGE(str);
 
     if (cr == ENC_CODERANGE_UNKNOWN) {
-	cr = ENC_CODERANGE_SINGLE;
-	for (i = 0; i < RSTRING_LEN(str); ++i) {
-	    const char *p = &RSTRING_PTR(str)[i];
-	    int c = (unsigned char)*p;
-
-	    if (!ISASCII(c)) {
-		c = rb_enc_codepoint(p, RSTRING_END(str), rb_enc_get(str));
-		if (c == -1) {
-		    cr = ENC_CODERANGE_BROKEN;
-		}
-		else {
+	rb_encoding *enc = rb_enc_get(str);
+
+	if (!rb_enc_asciicompat(enc)) {
+	    cr = ENC_CODERANGE_MULTI;
+	    ENC_CODERANGE_SET(str, cr);
+	    return cr;
+	}
+	else {
+	    const char *p = RSTRING_PTR(str);
+	    const char *e = p + RSTRING_LEN(str);
+
+	    cr = ENC_CODERANGE_SINGLE;
+	    while (p < e) {
+		int c = (unsigned char)*p;
+
+		if (c > 0x80) {
 		    cr = ENC_CODERANGE_MULTI;
+		    break;
 		}
+		p++;
 	    }
+	    ENC_CODERANGE_SET(str, cr);
 	}
-	ENC_CODERANGE_SET(str, cr);
     }
     return cr;
 }
@@ -1169,8 +1175,7 @@ rb_str_hash(VALUE str)
     if (e && is_ascii_string(str)) {
 	e = 0;
     }
-    return hash((const void *)RSTRING_PTR(str), RSTRING_LEN(str),
-		e);
+    return hash((const void *)RSTRING_PTR(str), RSTRING_LEN(str), e);
 }
 
 /*
@@ -1196,8 +1201,6 @@ rb_str_comparable(VALUE str1, VALUE str2)
     int idx2 = rb_enc_get_index(str2);
 
     if (idx1 == idx2) return Qtrue;
-    if (!rb_enc_asciicompat(rb_enc_from_index(idx1))) return Qfalse;
-    if (!rb_enc_asciicompat(rb_enc_from_index(idx2))) return Qfalse;
     if (!is_ascii_string(str1)) return Qfalse;
     if (!is_ascii_string(str2)) return Qfalse;
     return Qtrue;
@@ -1263,7 +1266,6 @@ rb_str_eql(VALUE str1, VALUE str2)
 	return Qfalse;
 
     if (!rb_str_comparable(str1, str2)) return Qfalse;
-
     if (memcmp(RSTRING_PTR(str1), RSTRING_PTR(str2),
 	       lesser(RSTRING_LEN(str1), RSTRING_LEN(str2))) == 0)
 	return Qtrue;