Fixed HTML-to-native conversion of unsupported HTML elements

The conversion of HTML elements that have no native representation was flawed because they were processed like elements with a native representation.
cdlm · May 5, 2011 · 3317b5e · 3317b5e
1 parent b598a80
commit 3317b5e
Show file tree

Hide file tree

Showing 7 changed files with 52 additions and 47 deletions.
diff --git a/doc/news/release_0_X_X.page b/doc/news/release_0_X_X.page
@@ -14,7 +14,7 @@ modified_at: 2011-02-21 18:01:00 +02:00
 
   - Added support for correctly parsing more HTML5 elements (requested by Bernt Carstenschulz)
 
-* 8 bug fixes:
+* 9 bug fixes:
 
   - Table line `` |a|`b` `` was parsed as `` |a`b` `` (patch by Masahiro Kitajima)
   - Table line `` |`a` `` lead to error condition (patch by Masahiro Kitajima)
@@ -27,6 +27,7 @@ modified_at: 2011-02-21 18:01:00 +02:00
   - Fixed problem with footnotes in LaTeX tables -- now using `longtable` instead of `tabular`
     environment (reported by Michael Franzl)
   - The `style` attribute is now used for outputting table cell alignments in HTML
+  - Fixed HTML-to-native conversion of unsupported HTML elements
 
 * 1 documentation fix:
 

diff --git a/lib/kramdown/parser/html.rb b/lib/kramdown/parser/html.rb
@@ -217,28 +217,32 @@ def process(el, do_conversion = true, preserve_text = false, parent = nil)
           else return
           end
 
-          type = el.value
-          remove_text_children(el) if REMOVE_TEXT_CHILDREN.include?(type)
-
           mname = "convert_#{el.value}"
           if do_conversion && self.class.method_defined?(mname)
             send(mname, el)
-          elsif do_conversion && SIMPLE_ELEMENTS.include?(type)
-            set_basics(el, type.intern)
-            process_children(el, do_conversion, preserve_text)
           else
-            process_html_element(el, do_conversion, preserve_text)
-          end
+            type = el.value
+            remove_text_children(el) if do_conversion && REMOVE_TEXT_CHILDREN.include?(type)
 
-          strip_whitespace(el) if STRIP_WHITESPACE.include?(type)
-          remove_whitespace_children(el) if REMOVE_WHITESPACE_CHILDREN.include?(type)
-          wrap_text_children(el) if WRAP_TEXT_CHILDREN.include?(type)
+            if do_conversion && SIMPLE_ELEMENTS.include?(type)
+              set_basics(el, type.intern)
+              process_children(el, do_conversion, preserve_text)
+            else
+              process_html_element(el, do_conversion, preserve_text)
+            end
+
+            if do_conversion
+              strip_whitespace(el) if STRIP_WHITESPACE.include?(type)
+              remove_whitespace_children(el) if REMOVE_WHITESPACE_CHILDREN.include?(type)
+              wrap_text_children(el) if WRAP_TEXT_CHILDREN.include?(type)
+            end
+          end
         end
 
         def process_children(el, do_conversion = true, preserve_text = false)
           el.children.map! do |c|
             if c.type == :text
-              process_text(c.value, preserve_text)
+              process_text(c.value, preserve_text || !do_conversion)
             else
               process(c, do_conversion, preserve_text, el)
               c
@@ -279,7 +283,7 @@ def process_text(raw, preserve = false)
 
         def process_html_element(el, do_conversion = true, preserve_text = false)
           el.options.replace(:category => HTML_SPAN_ELEMENTS.include?(el.value) ? :span : :block,
-                             :content_model => HTML_CONTENT_MODEL[el.value])
+                             :content_model => (do_conversion ? HTML_CONTENT_MODEL[el.value] : :raw))
           process_children(el, do_conversion, preserve_text)
         end
 
@@ -413,6 +417,7 @@ def convert_table(el)
             process_html_element(el, false)
             return
           end
+          remove_text_children(el)
           process_children(el)
           set_basics(el, :table)
 

diff --git a/test/testcases/block/09_html/content_model/tables.html b/test/testcases/block/09_html/content_model/tables.html
@@ -1,14 +1,14 @@
 <table class="examples">
-  <tr>
-    <th><em>Usage</em></th>
-    <th>
+<tr>
+  <th><em>Usage</em></th>
+  <th>
 Output
   </th>
-  </tr>
-  <tr>
-    <td>Some <em>data</em></td>
-    <td>
+</tr>
+<tr>
+  <td>Some <em>data</em></td>
+  <td>
       <h1 id="some-more">Some more</h1>
     </td>
-  </tr>
+</tr>
 </table>
diff --git a/test/testcases/block/09_html/content_model/tables.text b/test/testcases/block/09_html/content_model/tables.text
@@ -7,7 +7,7 @@ Output
 </tr>
 <tr>
   <td markdown="span">Some *data*</td>
-  <td>
+  <td markdown="1">
 # Some more
   </td>
 </tr>

diff --git a/test/testcases/block/09_html/html_to_native/emphasis.html b/test/testcases/block/09_html/html_to_native/emphasis.html
@@ -2,4 +2,5 @@
 
 <p>This is <strong>strong<em>italic</em>, yes!</strong>.</p>
 
-<p>This is <em> not</em> converted, as <em>is </em> this.</p>
+<p>This is <em> not</em> converted, as <em>is
+</em> this.</p>
diff --git a/test/testcases/block/09_html/html_to_native/table_normal.html b/test/testcases/block/09_html/html_to_native/table_normal.html
@@ -1,14 +1,12 @@
 <table class="examples">
-  <tr>
-    <th>Usage</th>
-    <th>Other</th>
-  </tr>
-  <tr>
-    <td>
-Some *data*
-    </td>
-    <td>
-      <p>Some more</p>
-    </td>
-  </tr>
+<tr>
+  <th>Usage</th>
+  <th>Other</th>
+</tr>
+<tr>
+  <td>Some *data*</td>
+  <td>
+    <p>Some more</p>
+  </td>
+</tr>
 </table>
diff --git a/test/testcases/block/09_html/html_to_native/table_simple.html b/test/testcases/block/09_html/html_to_native/table_simple.html
@@ -33,16 +33,16 @@
 </table>
 
 <table class="examples">
-  <tr>
-    <th>Usage</th>
-    <th>Output</th>
-  </tr>
-  <tr>
-    <td>
-Some *data*
-    </td>
-    <td>
+<tr>
+  <th>Usage</th>
+  <th>
+Output
+  </th>
+</tr>
+<tr>
+  <td>Some *data*</td>
+  <td>
 Some more
-    </td>
-  </tr>
+  </td>
+</tr>
 </table>