Fix extraction of block element contents.

Don't confound normal characters with formatting markers. Handle escaping of
    otherwise headline starters with a comma.

tests/04-code.org

@@ -36,3 +36,16 @@ exit 0 # Comment
     This is another test
     with two lines too
     :end:
+
+* Escaped code
+    :PROPERTIES:
+    :ID:       04-code-escaped-code-id
+    :CREATED:  [2020-01-01 Wed 01:01]
+    :END:
+
+    #+BEGIN_SRC c :results drawer
+/* This code has to be escaped to
+ ,* avoid confusion with new headlines.
+ ,*/
+main(){}
+    #+END_SRC

tests/test_org.py

@@ -433,7 +433,7 @@ class TestSerde(unittest.TestCase):
             doc = load(f)
 
         snippets = list(doc.get_code_snippets())
-        self.assertEqual(len(snippets), 2)
+        self.assertEqual(len(snippets), 3)
         self.assertEqual(
             snippets[0].content,
             'echo "This is a test"\n'
@@ -456,6 +456,14 @@ class TestSerde(unittest.TestCase):
             snippets[1].result, "This is another test\n" + "with two lines too"
         )
 
+        self.assertEqual(
+            snippets[2].content,
+            '/* This code has to be escaped to\n'
+            + ' * avoid confusion with new headlines.\n'
+            + ' */\n'
+            + 'main(){}',
+        )
+
     def test_mimic_write_file_05(self):
         with open(os.path.join(DIR, "05-dates.org")) as f:
             orig = f.read()