merge revision(s) 57302,57303,57304: [Backport #13119]

	string.c: block for scrub with ASCII-incompatible

	* string.c (rb_enc_str_scrub): honor the given block with
	  ASCII-incompatible encoding.  [ruby-core:79039] [Bug #13120]
	string.c: yield invalid part

	* string.c (rb_enc_str_scrub): yield the invalid part only with
	  ASCII-incompatible.  [ruby-core:79039] [Bug #13120]
	string.c: replacement and block

	* string.c (rb_enc_str_scrub): only one of replacement and block
	  is allowed.  [ruby-core:79038] [Bug #13119]

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_3@58175 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

3 files changed, 22 insertions, 7 deletions

diff --git a/string.c b/string.c
index c399f3eafd..f7480cd43b 100644
--- a/string.c
+++ b/string.c
@@ -8739,9 +8739,15 @@ rb_enc_str_scrub(rb_encoding *enc, VALUE str, VALUE repl)
     int encidx;
     VALUE buf = Qnil;
     const char *rep;
-    long replen;
+    long replen = -1;
     int tainted = 0;
 
+    if (rb_block_given_p()) {
+	if (!NIL_P(repl))
+	    rb_raise(rb_eArgError, "both of block and replacement given");
+	replen = 0;
+    }
+
     if (ENC_CODERANGE_CLEAN_P(cr))
 	return Qnil;
 
@@ -8765,9 +8771,8 @@ rb_enc_str_scrub(rb_encoding *enc, VALUE str, VALUE repl)
 	const char *e = RSTRING_END(str);
 	const char *p1 = p;
 	int rep7bit_p;
-	if (rb_block_given_p()) {
+	if (!replen) {
 	    rep = NULL;
-	    replen = 0;
 	    rep7bit_p = FALSE;
 	}
 	else if (!NIL_P(repl)) {
@@ -8878,7 +8883,10 @@ rb_enc_str_scrub(rb_encoding *enc, VALUE str, VALUE repl)
 	const char *e = RSTRING_END(str);
 	const char *p1 = p;
 	long mbminlen = rb_enc_mbminlen(enc);
-	if (!NIL_P(repl)) {
+	if (!replen) {
+	    rep = NULL;
+	}
+	else if (!NIL_P(repl)) {
 	    rep = RSTRING_PTR(repl);
 	    replen = RSTRING_LEN(repl);
 	}
@@ -8929,7 +8937,7 @@ rb_enc_str_scrub(rb_encoding *enc, VALUE str, VALUE repl)
 		    rb_str_buf_cat(buf, rep, replen);
 		}
 		else {
-		    repl = rb_yield(rb_enc_str_new(p, e-p, enc));
+		    repl = rb_yield(rb_enc_str_new(p, clen, enc));
 		    repl = str_compat_and_valid(repl, enc);
 		    tainted |= OBJ_TAINTED_RAW(repl);
 		    rb_str_buf_cat(buf, RSTRING_PTR(repl), RSTRING_LEN(repl));
diff --git a/test/ruby/test_m17n.rb b/test/ruby/test_m17n.rb
index f0a6b4c95b..641b8286b7 100644
--- a/test/ruby/test_m17n.rb
+++ b/test/ruby/test_m17n.rb
@@ -1623,8 +1623,9 @@ class TestM17N < Test::Unit::TestCase
     assert_raise(ArgumentError){ u("\xE3\x81\x82\xE3\x81\x82\xE3\x81").scrub{u("\x81")} }
     assert_equal(e("\xA4\xA2\xA2\xAE"), e("\xA4\xA2\xA4").scrub{e("\xA2\xAE")})
 
-    assert_equal(u("\x81"), u("a\x81").scrub {|c| break c})
+    assert_equal(u("\x81"), u("a\x81c").scrub {|c| break c})
     assert_raise(ArgumentError) {u("a\x81").scrub {|c| c}}
+    assert_raise(ArgumentError) {u("a").scrub("?") {|c| c}}
   end
 
   def test_scrub_widechar
@@ -1640,6 +1641,12 @@ class TestM17N < Test::Unit::TestCase
     assert_equal("\uFFFD".encode("UTF-32LE"),
                  "\xff".force_encoding(Encoding::UTF_32LE).
                  scrub)
+    c = nil
+    assert_equal("?\u3042".encode(Encoding::UTF_16LE),
+                 "\x00\xD8\x42\x30".force_encoding(Encoding::UTF_16LE).
+                   scrub {|e| c = e; "?".encode(Encoding::UTF_16LE)})
+    assert_equal("\x00\xD8".force_encoding(Encoding::UTF_16LE), c)
+    assert_raise(ArgumentError) {"\uFFFD\u3042".encode("UTF-16BE").scrub("") {}}
   end
 
   def test_scrub_dummy_encoding
diff --git a/version.h b/version.h
index 7a740e98f9..9eb5d5b06d 100644
--- a/version.h
+++ b/version.h
@@ -1,6 +1,6 @@
 #define RUBY_VERSION "2.3.3"
 #define RUBY_RELEASE_DATE "2017-03-28"
-#define RUBY_PATCHLEVEL 284
+#define RUBY_PATCHLEVEL 285
 
 #define RUBY_RELEASE_YEAR 2017
 #define RUBY_RELEASE_MONTH 3