* transcode.c (rb_trans_conv): report last transcode_obuf_full.

(econv_max_output): new method Encoding::Converter#max_output.


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@18547 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

ChangeLog

@@ -1,3 +1,8 @@
+Wed Aug 13 07:41:03 2008  Tanaka Akira  <akr@fsij.org>
+
+	* transcode.c (rb_trans_conv): report last transcode_obuf_full.
+	  (econv_max_output): new method Encoding::Converter#max_output.
+
 Wed Aug 13 02:46:01 2008  NARUSE, Yui  <naruse@ruby-lang.org>
 
 	* error.c (rb_eEncCompatError): add Exception.

test/ruby/test_econv.rb

@@ -1,24 +1,53 @@
 require 'test/unit'
 
 class TestEncodingConverter < Test::Unit::TestCase
-  def assert_econv(ret_expected, src_expected, dst_expected, from, to, src, dst, flags=0)
+  def assert_econv(ret_expected, dst_expected, src_expected, to, from, src, opt={})
+    opt[:obuf_len] ||= 100
+    src = src.dup
     ec = Encoding::Converter.new(from, to)
-    ret = ec.primitive_convert(src, dst, flags)
-    assert_equal(ret_expected, ret)
-    assert_equal(src_expected, src)
-    assert_equal(dst_expected, dst)
+    dst = ''
+    while true
+      ret = ec.primitive_convert(src, dst2=" "*opt[:obuf_len], 0)
+      dst << dst2
+      #p [ret, dst, src]
+      break if ret != :obuf_full
+    end
+    assert_equal([ret_expected, dst_expected, src_expected], [ret, dst, src])
   end
 
   def test_eucjp_to_utf8
-    assert_econv(:finished, "", "", "EUC-JP", "UTF-8", "", "")
-    assert_econv(:ibuf_empty, "", "", "EUC-JP", "UTF-8", "", "", Encoding::Converter::PARTIAL_INPUT)
-    assert_econv(:finished, "", "", "EUC-JP", "UTF-8", "", " "*10)
-    assert_econv(:obuf_full, "", "", "EUC-JP", "UTF-8", "a", "")
+    assert_econv(:finished, "", "", "EUC-JP", "UTF-8", "")
+    assert_econv(:finished, "a", "", "EUC-JP", "UTF-8", "a")
+  end
+
+  def test_iso2022jp
+    assert_econv(:finished, "", "", "ISO-2022-JP", "Shift_JIS", "")
   end
 
   def test_invalid
-    assert_econv(:invalid_input, "", "", "EUC-JP", "UTF-8", "\x80", " "*10)
-    assert_econv(:invalid_input, "", "a", "EUC-JP", "UTF-8", "a\x80", " "*10)
-    assert_econv(:invalid_input, "\x80", "a", "EUC-JP", "UTF-8", "a\x80\x80", " "*10)
+    assert_econv(:invalid_input, "", "", "EUC-JP", "UTF-8", "\x80")
+    assert_econv(:invalid_input, "a", "", "EUC-JP", "UTF-8", "a\x80")
+    assert_econv(:invalid_input, "a", "\x80", "EUC-JP", "UTF-8", "a\x80\x80")
+    assert_econv(:invalid_input, "abc", "def", "EUC-JP", "UTF-8", "abc\xFFdef")
+    assert_econv(:invalid_input, "abc", "def", "EUC-JP", "Shift_JIS", "abc\xFFdef")
+    assert_econv(:invalid_input, "abc", "def", "EUC-JP", "Shift_JIS", "abc\xFFdef", :obuf_len=>1)
+    assert_econv(:invalid_input, "abc", "def", "Shift_JIS", "ISO-2022-JP", "abc\xFFdef")
+  end
+
+  def test_errors
+    ec = Encoding::Converter.new("UTF-16BE", "EUC-JP")
+    src = "\xFF\xFE\x00A\xDC\x00"
+    ret = ec.primitive_convert(src, dst=" "*10, 0)
+    assert_equal("", src)
+    assert_equal("", dst)
+    assert_equal(:undefined_conversion, ret)
+    ret = ec.primitive_convert(src, dst=" "*10, 0)
+    assert_equal("", src)
+    assert_equal("A", dst)
+    assert_equal(:invalid_input, ret)
+    ret = ec.primitive_convert(src, dst=" "*10, 0)
+    assert_equal("", src)
+    assert_equal("", dst)
+    assert_equal(:finished, ret)
   end
 end

transcode.c

@@ -749,7 +749,7 @@ rb_trans_conv(rb_trans_t *ts,
     int flags)
 {
     int i;
-    int start, err_index, no_error;
+    int start, err_index;
 
     unsigned char empty_buf;
     unsigned char *empty_ptr = &empty_buf;
@@ -764,20 +764,11 @@ rb_trans_conv(rb_trans_t *ts,
         output_stop = empty_ptr;
     }
 
-    no_error = 1;
     err_index = -1;
     for (i = ts->num_trans-1; 0 <= i; i--) {
-        if (ts->elems[i].last_result == transcode_invalid_input ||
-            ts->elems[i].last_result == transcode_undefined_conversion) {
-            if (no_error) {
-                /* last error */
-                no_error = 0;
-            }
-            else {
-                /* second last error */
-                err_index = i;
-                break;
-            }
+        if (ts->elems[i].last_result != transcode_ibuf_empty) {
+            err_index = i;
+            break;
         }
     }
 
@@ -786,12 +777,14 @@ rb_trans_conv(rb_trans_t *ts,
         err_index = trans_sweep(ts, input_ptr, input_stop, output_ptr, output_stop, flags, start);
     } while (err_index != -1 && err_index != ts->num_trans-1);
 
-    if (err_index == ts->num_trans-1)
-        return ts->elems[ts->num_trans-1].last_result;
-    else if (start == 0)
-        return ts->elems[ts->num_trans-1].last_result;
-    else
-        return ts->elems[start-1].last_result;
+    for (i = ts->num_trans-1; 0 <= i; i--) {
+        if (ts->elems[i].last_result != transcode_ibuf_empty) {
+            rb_trans_result_t res = ts->elems[i].last_result;
+            ts->elems[i].last_result = transcode_ibuf_empty;
+            return res;
+        }
+    }
+    return transcode_ibuf_empty;
 }
 
 static void
@@ -1303,6 +1296,16 @@ econv_primitive_convert(VALUE self, VALUE input, VALUE output, VALUE flags_v)
     }
 }
 
+static VALUE
+econv_max_output(VALUE self)
+{
+    rb_trans_t *ts = check_econv(self);
+    int n;
+    n = ts->elems[ts->num_trans-1].tc->transcoder->max_output;
+
+    return INT2FIX(n);
+}
+
 void
 Init_transcode(void)
 {
@@ -1323,5 +1326,6 @@ Init_transcode(void)
     rb_define_alloc_func(rb_cEncodingConverter, econv_s_allocate);
     rb_define_method(rb_cEncodingConverter, "initialize", econv_init, 2);
     rb_define_method(rb_cEncodingConverter, "primitive_convert", econv_primitive_convert, 3);
+    rb_define_method(rb_cEncodingConverter, "max_output", econv_max_output, 0);
     rb_define_const(rb_cEncodingConverter, "PARTIAL_INPUT", INT2FIX(PARTIAL_INPUT));
 }