* string.c (rb_str_index): check if substring is broken.

* string.c (rb_str_rindex): ditto.

* string.c (rb_str_succ): should carry over.

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@14268 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

ChangeLog

@@ -1,3 +1,11 @@
+Mon Dec 17 17:50:30 2007  Yukihiro Matsumoto  <matz@ruby-lang.org>
+
+	* string.c (rb_str_index): check if substring is broken.
+
+	* string.c (rb_str_rindex): ditto.
+
+	* string.c (rb_str_succ): should carry over.
+
 Mon Dec 17 17:47:26 2007  Nobuyoshi Nakada  <nobu@ruby-lang.org>
 
 	* common.mk (encs): new target to compile external encodings.

bootstraptest/test_knownbug.rb

@@ -219,12 +219,12 @@ assert_equal 'false', %q{
   "\xa1\xa2\xa3\xa4".force_encoding("euc-jp").include?("\xa3".force_encoding("euc-jp"))
 }
 
-assert_equal 'nil', %q{
-  "\xa1\xa2\xa3\xa4".force_encoding("euc-jp").index("\xa3".force_encoding("euc-jp"))
+assert_equal 'ok', %q{
+  "\xa1\xa2\xa3\xa4".force_encoding("euc-jp").index("\xa3".force_encoding("euc-jp")) or :ok
 }
 
-assert_equal 'nil', %q{
-  "\xa1\xa2\xa3\xa4".force_encoding("euc-jp").rindex("\xa3".force_encoding("euc-jp"))
+assert_equal 'ok', %q{
+  "\xa1\xa2\xa3\xa4".force_encoding("euc-jp").rindex("\xa3".force_encoding("euc-jp")) or :ok
 }
 
 assert_equal 'false', %q{

encoding.c

@@ -553,7 +553,8 @@ rb_enc_precise_mbclen(const char *p, const char *e, rb_encoding *enc)
     return n;
 }
 
-int rb_enc_ascget(const char *p, const char *e, int *len, rb_encoding *enc)
+int
+rb_enc_ascget(const char *p, const char *e, int *len, rb_encoding *enc)
 {
     int c, l;
     if (e <= p)

string.c

@@ -94,6 +94,7 @@ VALUE rb_cSymbol;
 
 #define is_ascii_string(str) (rb_enc_str_coderange(str) == ENC_CODERANGE_7BIT)
 #define IS_7BIT(str) (ENC_CODERANGE(str) == ENC_CODERANGE_7BIT)
+#define is_broken_string(str) (rb_enc_str_coderange(str) == ENC_CODERANGE_BROKEN)
 
 VALUE rb_fs;
 
@@ -1443,6 +1444,9 @@ rb_str_index(VALUE str, VALUE sub, long offset)
     rb_encoding *enc;
 
     enc = rb_enc_check(str, sub);
+    if (is_broken_string(sub)) {
+	return -1;
+    }
     len = str_strlen(str, enc);
     slen = str_strlen(sub, enc);
     if (offset < 0) {
@@ -1553,6 +1557,9 @@ rb_str_rindex(VALUE str, VALUE sub, long pos)
     int asc = IS_7BIT(str);
 
     enc = rb_enc_check(str, sub);
+    if (is_broken_string(sub)) {
+	return -1;
+    }
     len = str_strlen(str, enc);
     slen = str_strlen(sub, enc);
     /* substring longer than string */
@@ -1863,7 +1870,7 @@ rb_str_succ(VALUE orig)
     if (c == -1) {		/* str contains no alnum */
 	c = '\001';
 	s = e;
-	while ((s = rb_enc_prev_char(sbeg, e, enc)) != 0) {
+	while ((s = rb_enc_prev_char(sbeg, s, enc)) != 0) {
 	    if (cc == 0) cc = rb_enc_codepoint(s, e, enc);
 	    cc += 1;
 	    l = rb_enc_mbcput(cc, carry, enc);