Merge browsing of binary files. Fixes bug 1052217. Close #386.

dxr/app.py

@@ -264,7 +264,8 @@ def _browse_folder(tree, path, config):
              f['name'],
              decode_es_datetime(f['modified']) if 'modified' in f else None,
              f.get('size'),
-             url_for('.browse', tree=tree, path=f['path'][0]))
+             url_for('.browse', tree=tree, path=f['path'][0]),
+             f.get('is_binary', [False])[0])
             for f in files_and_folders])
 
 

dxr/build.py

@@ -486,34 +486,30 @@ def docs():
         # indexing. We could interpose an external queueing system, but I'm
         # willing to potentially sacrifice a little speed here for the easy
         # management of self-throttling.
-        #
-        # Conditional until we figure out how to display arbitrary binary
-        # files:
-        if is_text or is_image(rel_path):
-            file_info = stat(path)
-            folder_name, file_name = split(rel_path)
-            # Hard-code the keys that are hard-coded in the browse()
-            # controller. Merge with the pluggable ones from needles:
-            doc = dict(# Some non-array fields:
-                       folder=folder_name,
-                       name=file_name,
-                       size=file_info.st_size,
-                       modified=datetime.fromtimestamp(file_info.st_mtime),
-                       is_folder=False,
-
-                       # And these, which all get mashed into arrays:
-                       **needles)
-            links = [{'order': order,
-                      'heading': heading,
-                      'items': [{'icon': icon,
-                                 'title': title,
-                                 'href': href}
-                                for icon, title, href in items]}
-                     for order, heading, items in
-                     chain.from_iterable(linkses)]
-            if links:
-                doc['links'] = links
-            yield es.index_op(doc, doc_type=FILE)
+        file_info = stat(path)
+        folder_name, file_name = split(rel_path)
+        # Hard-code the keys that are hard-coded in the browse()
+        # controller. Merge with the pluggable ones from needles:
+        doc = dict(# Some non-array fields:
+                    folder=folder_name,
+                    name=file_name,
+                    size=file_info.st_size,
+                    modified=datetime.fromtimestamp(file_info.st_mtime),
+                    is_folder=False,
+
+                    # And these, which all get mashed into arrays:
+                    **needles)
+        links = [{'order': order,
+                    'heading': heading,
+                    'items': [{'icon': icon,
+                                'title': title,
+                                'href': href}
+                            for icon, title, href in items]}
+                    for order, heading, items in
+                    chain.from_iterable(linkses)]
+        if links:
+            doc['links'] = links
+        yield es.index_op(doc, doc_type=FILE)
 
         # Index all the lines. If it's an empty file (no lines), don't bother
         # ES. It hates empty dicts.

dxr/config.py

@@ -207,7 +207,8 @@ def __init__(self, name, unvalidated_tree, sections, config):
             Optional('enabled_plugins', default=plugin_list('*')): Plugins,
             Optional('ignore_patterns',
                      default=['.hg', '.git', 'CVS', '.svn', '.bzr',
-                              '.deps', '.libs']): WhitespaceList,
+                              '.deps', '.libs', '.DS_Store', '.nfs*', '*~',
+                              '._*']): WhitespaceList,
             Optional('object_folder', default=None): AbsPath,
             'source_folder': AbsPath,
             Optional('source_encoding', default='utf-8'): basestring,

dxr/plugins/core.py

@@ -83,6 +83,10 @@
                 'type': 'binary',
                 'index': 'no'
             },
+            'is_binary': { # assumed False if not present
+                'type': 'boolean',
+                'index': 'no'
+            },
 
             # Sidebar nav links:
             'links': {
@@ -381,6 +385,9 @@ def needles(self):
             bytestring = (self.contents.encode('utf-8') if self.contains_text()
                           else self.contents)
             yield 'raw_data', b64encode(bytestring)
+        # binary, but not an image
+        elif not self.contains_text():
+            yield 'is_binary', True
 
     def needles_by_line(self):
         """Fill out line number and content for every line."""
@@ -389,8 +396,8 @@ def needles_by_line(self):
                    ('content', text)]
 
     def is_interesting(self):
-        """Core plugin puts text and image files in the search index."""
-        return self.contains_text() or is_image(self.path)
+        """Core plugin puts all files in the search index."""
+        return True
 
 
 # Match file name and line number: filename:n. Strip leading slashes because

dxr/static/css/dxr.css

@@ -148,6 +148,13 @@ td#line-numbers span {
     text-align: right;
     padding: 0 0.5rem;
 }
+tr.binary_row > td {
+    padding-top: .5em;
+    padding-bottom: .5em;
+}
+tr.binary_row {
+    opacity: 0.6;
+}
 .highlighted,
 .multihighlight {
     background: none repeat scroll 0 0 rgb(255, 255, 204);

dxr/static/templates/folder.html

@@ -32,12 +32,20 @@
       </tr>
     </thead>
     <tbody>
-      {% for icon, name, modified, size, relative_url in files_and_folders %}
-        <tr>
-          <td><a href="{{ relative_url }}" class="icon {{ icon }}">{{ name }}</a></td>
-          <td><a href="{{ relative_url }}">{{ '' if modified is none else modified.strftime("%Y %b %d %H:%m") }}</a></td>
-          <td><a href="{{ relative_url }}">{{ '' if size is none else size|filesizeformat }}</a></td>
-        </tr>
+      {% for icon, name, modified, size, relative_url, is_binary in files_and_folders %}
+        {% if is_binary %}
+          <tr class="binary_row">
+            <td><span class="icon {{ icon }}">{{ name }}</span></td>
+            <td>{{ '' if modified is none else modified.strftime("%Y %b %d %H:%m") }}</td>
+            <td>{{ '' if size is none else size|filesizeformat }}</td>
+          </tr>
+        {% else %}
+          <tr>
+            <td><a href="{{ relative_url }}" class="icon {{ icon }}">{{ name }}</a></td>
+            <td><a href="{{ relative_url }}">{{ '' if modified is none else modified.strftime("%Y %b %d %H:%m") }}</a></td>
+            <td><a href="{{ relative_url }}">{{ '' if size is none else size|filesizeformat }}</a></td>
+          </tr>
+        {% endif %}
       {% endfor %}
     </tbody>
   </table>

tests/test_binary_files/test_binary_files.py

@@ -38,6 +38,8 @@ def test_svg(self):
         ok_(response.status_code, 200)
 
     def test_some_bytes(self):
-        """We do not want some_bytes to show in the folder index."""
+        """We want some_bytes to show in the folder index, but without its own page."""
         response = self.client().get('/code/source/')
-        ok_('some_bytes' not in response.data)
+        ok_('some_bytes' in response.data)
+        response = self.client().get('/code/source/some_bytes')
+        ok_(response.status_code, 404)