* encoding.c (rb_enc_mbclen): return minlen instead of 1 when

  a character is not found properly.

* string.c (rb_enc_strlen): round up string length with fixed
  multibyte encoding such as UTF-32.
  (rb_enc_strlen_cr): ditto.
  (rb_str_substr): fix substring with fixed multibyte encoding.
  (rb_str_justify): check number of characters.



git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@15573 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

4 files changed, 48 insertions, 9 deletions

diff --git a/ChangeLog b/ChangeLog
index 2d19f72903..82b6fd960a 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,14 @@
+Fri Feb 22 15:47:36 2008  Tanaka Akira  <akr@fsij.org>
+
+	* encoding.c (rb_enc_mbclen): return minlen instead of 1 when
+	  a character is not found properly.
+
+	* string.c (rb_enc_strlen): round up string length with fixed
+	  multibyte encoding such as UTF-32.
+	  (rb_enc_strlen_cr): ditto.
+	  (rb_str_substr): fix substring with fixed multibyte encoding.
+	  (rb_str_justify): check number of characters.
+
 Fri Feb 22 12:11:12 2008  NARUSE, Yui  <naruse@ruby-lang.org>
 
 	* string.c (rb_str_inspect): string of ascii incompatible encoding
diff --git a/encoding.c b/encoding.c
index 002c0c04fb..813b4b5062 100644
--- a/encoding.c
+++ b/encoding.c
@@ -738,8 +738,10 @@ rb_enc_mbclen(const char *p, const char *e, rb_encoding *enc)
     int n = ONIGENC_PRECISE_MBC_ENC_LEN(enc, (UChar*)p, (UChar*)e);
     if (MBCLEN_CHARFOUND_P(n) && MBCLEN_CHARFOUND_LEN(n) <= e-p)
         return MBCLEN_CHARFOUND_LEN(n);
-    else
-        return 1;
+    else {
+        int min = rb_enc_mbminlen(enc);
+        return min <= e-p ? min : e-p;
+    }
 }
 
 int
diff --git a/string.c b/string.c
index 19e8072462..90ed51b108 100644
--- a/string.c
+++ b/string.c
@@ -618,7 +618,7 @@ rb_enc_strlen(const char *p, const char *e, rb_encoding *enc)
     const char *q;
 
     if (rb_enc_mbmaxlen(enc) == rb_enc_mbminlen(enc)) {
-        return (e - p) / rb_enc_mbminlen(enc);
+        return (e - p + rb_enc_mbminlen(enc) - 1) / rb_enc_mbminlen(enc);
     }
     else if (rb_enc_asciicompat(enc)) {
         c = 0;
@@ -651,7 +651,7 @@ rb_enc_strlen_cr(const char *p, const char *e, rb_encoding *enc, int *cr)
 
     *cr = 0;
     if (rb_enc_mbmaxlen(enc) == rb_enc_mbminlen(enc)) {
-	return (e - p) / rb_enc_mbminlen(enc);
+	return (e - p + rb_enc_mbminlen(enc) - 1) / rb_enc_mbminlen(enc);
     }
     else if (rb_enc_asciicompat(enc)) {
 	c = 0;
@@ -1223,10 +1223,9 @@ rb_str_substr(VALUE str, long beg, long len)
 	len = 0;
     }
     else if (rb_enc_mbmaxlen(enc) == rb_enc_mbminlen(enc)) {
-	long rest = (e - p) / rb_enc_mbmaxlen(enc);
-	if (len > rest)
-	    len = rest;
-	else
+        if (len * rb_enc_mbmaxlen(enc) > e - p)
+            len = e - p;
+        else
 	    len *= rb_enc_mbmaxlen(enc);
     }
     else {
@@ -5777,7 +5776,7 @@ rb_str_justify(int argc, VALUE *argv, VALUE str, char jflag)
 	flen = RSTRING_LEN(pad);
 	fclen = str_strlen(pad, enc);
 	singlebyte = single_byte_optimizable(pad);
-	if (flen == 0) {
+	if (flen == 0 || fclen == 0) {
 	    rb_raise(rb_eArgError, "zero width padding");
 	}
     }
diff --git a/test/ruby/test_utf32.rb b/test/ruby/test_utf32.rb
new file mode 100644
index 0000000000..f81524f29e
--- /dev/null
+++ b/test/ruby/test_utf32.rb
@@ -0,0 +1,27 @@
+require 'test/unit'
+
+class TestUTF32 < Test::Unit::TestCase
+  def encdump(str)
+    d = str.dump
+    if /\.force_encoding\("[A-Za-z0-9.:_+-]*"\)\z/ =~ d
+      d
+    else
+      "#{d}.force_encoding(#{str.encoding.name.dump})"
+    end
+  end
+
+  def assert_str_equal(expected, actual, message=nil)
+    full_message = build_message(message, <<EOT)
+#{encdump expected} expected but not equal to
+#{encdump actual}.
+EOT
+    assert_block(full_message) { expected == actual }
+  end
+
+  def test_substr
+    assert_str_equal(
+      "abcdefgh".force_encoding("utf-32be"),
+      "abcdefgh".force_encoding("utf-32be")[0,3])
+  end
+end
+