Fix IOSource#readline for @pending_buffer (#215)

naitoh · web-flow · commit cf0fb9c9ca3d · 2024-10-19T15:27:25.000+09:00
## Why?
Fixed a problem that `@pending_buffer` is not processed when `IOError`
occurs in `@source.readline` although `@pending_buffer` exists when
reading XML file.
diff --git a/lib/rexml/parsers/baseparser.rb b/lib/rexml/parsers/baseparser.rb
@@ -167,6 +167,7 @@ def initialize( source )
         @entity_expansion_count = 0
         @entity_expansion_limit = Security.entity_expansion_limit
         @entity_expansion_text_limit = Security.entity_expansion_text_limit
+        @source.ensure_buffer
       end
 
       def add_listener( listener )
diff --git a/lib/rexml/source.rb b/lib/rexml/source.rb
@@ -295,14 +295,19 @@ def current_line
 
     private
     def readline(term = nil)
-      str = @source.readline(term || @line_break)
       if @pending_buffer
+        begin
+          str = @source.readline(term || @line_break)
+        rescue IOError
+        end
         if str.nil?
           str = @pending_buffer
         else
           str = @pending_buffer + str
         end
         @pending_buffer = nil
+      else
+        str = @source.readline(term || @line_break)
       end
       return nil if str.nil?
 
diff --git a/test/parse/test_text.rb b/test/parse/test_text.rb
@@ -4,6 +4,23 @@
 module REXMLTests
   class TestParseText < Test::Unit::TestCase
     class TestInvalid < self
+      def test_text_only
+        exception = assert_raise(REXML::ParseException) do
+          parser = REXML::Parsers::BaseParser.new('a')
+          while parser.has_next?
+            parser.pull
+          end
+        end
+
+        assert_equal(<<~DETAIL.chomp, exception.to_s)
+          Malformed XML: Content at the start of the document (got 'a')
+          Line: 1
+          Position: 1
+          Last 80 unconsumed characters:
+
+        DETAIL
+      end
+
       def test_before_root
         exception = assert_raise(REXML::ParseException) do
           parser = REXML::Parsers::BaseParser.new('b<a></a>')
