akr	2008-09-06 14:47:30 +0900 (Sat, 06 Sep 2008)

  New Revision: 19179

  http://svn.ruby-lang.org/cgi-bin/viewvc.cgi?view=rev&revision=19179

  Log:
    * include/ruby/encoding.h (ECONV_HTML_TEXT_ENCODER): new constant.
      (ECONV_HTML_ATTR_ENCODER): ditto.
    
    * transcode.c (rb_econv_open): check ECONV_HTML_TEXT_ENCODER and
      ECONV_HTML_ATTR_ENCODER.
      (Init_transcode): Encoding::Converter::HTML_TEXT_ENCODER and
      Encoding::Converter::HTML_ATTR_ENCODER defined.

  Modified files:
    trunk/ChangeLog
    trunk/include/ruby/encoding.h
    trunk/test/ruby/test_econv.rb
    trunk/transcode.c

Index: include/ruby/encoding.h
===================================================================
--- include/ruby/encoding.h	(revision 19178)
+++ include/ruby/encoding.h	(revision 19179)
@@ -263,6 +263,8 @@
 /* usable only if source encoding is ascii compatible */
 #define ECONV_CRLF_NEWLINE_ENCODER              0x0200
 #define ECONV_CR_NEWLINE_ENCODER                0x0400
+#define ECONV_HTML_TEXT_ENCODER                 0x0800
+#define ECONV_HTML_ATTR_ENCODER                 0x1000
 
 /* end of flags for rb_econv_open */
 
Index: ChangeLog
===================================================================
--- ChangeLog	(revision 19178)
+++ ChangeLog	(revision 19179)
@@ -1,3 +1,13 @@
+Sat Sep  6 14:46:12 2008  Tanaka Akira  <akr@f...>
+
+	* include/ruby/encoding.h (ECONV_HTML_TEXT_ENCODER): new constant.
+	  (ECONV_HTML_ATTR_ENCODER): ditto.
+
+	* transcode.c (rb_econv_open): check ECONV_HTML_TEXT_ENCODER and
+	  ECONV_HTML_ATTR_ENCODER.
+	  (Init_transcode): Encoding::Converter::HTML_TEXT_ENCODER and
+	  Encoding::Converter::HTML_ATTR_ENCODER defined.
+
 Sat Sep  6 14:15:25 2008  Tanaka Akira  <akr@f...>
 
 	* transcode.c (struct trans_open_t): defined to pass num_additional.
Index: test/ruby/test_econv.rb
===================================================================
--- test/ruby/test_econv.rb	(revision 19178)
+++ test/ruby/test_econv.rb	(revision 19179)
@@ -754,4 +754,18 @@
     assert_equal('"&amp;&lt;&gt;&quot;', ec.convert("&<>\""))
     assert_equal('"', ec.finish)
   end
+
+  def test_html_escape_with_charref
+    ec = Encoding::Converter.new("utf-8", "euc-jp", Encoding::Converter::HTML_TEXT_ENCODER|Encoding::Converter::UNDEF_HEX_CHARREF)
+    assert_equal('&lt;&#x2665;&gt;&amp;"&#x2661;"', ec.convert("<\u2665>&\"\u2661\""))
+    assert_equal('', ec.finish)
+
+    ec = Encoding::Converter.new("utf-8", "euc-jp", Encoding::Converter::HTML_ATTR_ENCODER|Encoding::Converter::UNDEF_HEX_CHARREF)
+    assert_equal('"&lt;&#x2665;&gt;&amp;&quot;&#x2661;&quot;', ec.convert("<\u2665>&\"\u2661\""))
+    assert_equal('"', ec.finish)
+
+    ec = Encoding::Converter.new("utf-8", "iso-2022-jp", Encoding::Converter::HTML_TEXT_ENCODER)
+    assert_equal("&amp;\e$B$&\e(B&amp;".force_encoding("iso-2022-jp"), ec.convert("&\u3046&"))
+    assert_equal('', ec.finish)
+  end
 end
Index: transcode.c
===================================================================
--- transcode.c	(revision 19178)
+++ transcode.c	(revision 19179)
@@ -895,6 +895,10 @@
         (ecflags & ECONV_UNIVERSAL_NEWLINE_DECODER))
         return NULL;
 
+    if ((ecflags & ECONV_HTML_TEXT_ENCODER) &&
+        (ecflags & ECONV_HTML_ATTR_ENCODER))
+        return NULL;
+
     num_encoders = 0;
     if (ecflags & ECONV_CRLF_NEWLINE_ENCODER)
         if (!(encoders[num_encoders++] = get_transcoder_entry("", "crlf_newline")))
@@ -902,6 +906,12 @@
     if (ecflags & ECONV_CR_NEWLINE_ENCODER)
         if (!(encoders[num_encoders++] = get_transcoder_entry("", "cr_newline")))
             return NULL;
+    if (ecflags & ECONV_HTML_TEXT_ENCODER)
+        if (!(encoders[num_encoders++] = get_transcoder_entry("", "html-text-escaped")))
+            return NULL;
+    if (ecflags & ECONV_HTML_ATTR_ENCODER)
+        if (!(encoders[num_encoders++] = get_transcoder_entry("", "html-attr-escaped")))
+            return NULL;
 
     num_decoders = 0;
     if (ecflags & ECONV_UNIVERSAL_NEWLINE_DECODER)
@@ -3510,6 +3520,8 @@
     rb_define_const(rb_cEncodingConverter, "UNIVERSAL_NEWLINE_DECODER", INT2FIX(ECONV_UNIVERSAL_NEWLINE_DECODER));
     rb_define_const(rb_cEncodingConverter, "CRLF_NEWLINE_ENCODER", INT2FIX(ECONV_CRLF_NEWLINE_ENCODER));
     rb_define_const(rb_cEncodingConverter, "CR_NEWLINE_ENCODER", INT2FIX(ECONV_CR_NEWLINE_ENCODER));
+    rb_define_const(rb_cEncodingConverter, "HTML_TEXT_ENCODER", INT2FIX(ECONV_HTML_TEXT_ENCODER));
+    rb_define_const(rb_cEncodingConverter, "HTML_ATTR_ENCODER", INT2FIX(ECONV_HTML_ATTR_ENCODER));
 
     rb_define_method(rb_eConversionUndefined, "source_encoding_name", ecerr_source_encoding_name, 0);
     rb_define_method(rb_eConversionUndefined, "destination_encoding_name", ecerr_destination_encoding_name, 0);

--
ML: ruby-changes@q...
Info: http://www.atdot.net/~ko1/quickml/