cope with utf-8 and very long filenames (bug #649817)

apps/files/helpers.py

@@ -6,6 +6,7 @@ import stat
 
 from django.conf import settings
 from django.utils.datastructures import SortedDict
+from django.utils.encoding import smart_unicode
 
 import commonware.log
 from jingo import register
@@ -100,6 +101,22 @@ class FileViewer:
             return {}
         return self._get_files()
 
+    def truncate(self, filename, pre_length=15,
+                 post_length=10, ellipsis=u'..'):
+        """
+        Truncates a filename so that
+           somelongfilename.htm
+        becomes:
+           some...htm
+        as it truncates around the extension.
+        """
+        root, ext = os.path.splitext(filename)
+        if len(root) > pre_length:
+            root = root[:pre_length] + ellipsis
+        if len(ext) > post_length:
+            ext = ext[:post_length] + ellipsis
+        return root + ext
+
     @memoize(prefix='file-viewer')
     def _get_files(self):
         all_files, res = [], SortedDict()
@@ -108,7 +125,8 @@ class FileViewer:
                 all_files.append([os.path.join(root, filename), filename])
 
         for path, filename in sorted(all_files):
-            short = path[len(self.dest) + 1:]
+            filename = smart_unicode(filename)
+            short = smart_unicode(path[len(self.dest) + 1:])
             mime, encoding = mimetypes.guess_type(filename)
             binary = self.is_binary(mime, filename)
             # TODO: handling for st.a.m.o will be going in here
@@ -116,6 +134,7 @@ class FileViewer:
             directory = os.path.isdir(path)
             res[short] = {'full': path,
                           'filename': filename,
+                          'truncated': self.truncate(filename),
                           'short': short,
                           'directory': directory,
                           'url': url,

apps/files/templates/files/viewer.html

@@ -22,7 +22,8 @@
                     style="padding-left: {{ value['depth'] }}em"
                     class="{% if value['depth'] %}hidden{% endif %}">
                     <a class="{{ file_viewer_class(value, selected) }}"
-                        href="{{ value['url'] }}">{{ value['filename'] }}</a>
+                        href="{{ value['url'] }}"
+                        title="{{ value['filename'] }}">{{ value['truncated'] }}</a>
                 </li>
             {% endfor %}
         {% endif %}

apps/files/tests/test_helpers.py

@@ -1,3 +1,4 @@
+# -*- coding: utf-8 -*-
 import os
 import mimetypes
 import tempfile
@@ -60,6 +61,16 @@ class TestFileHelper(test_utils.TestCase):
             m, encoding = mimetypes.guess_type(f)
             assert binary(m, f), '%s should be binary' % f
 
+    def test_truncate(self):
+        truncate = self.viewer.truncate
+        for x, y in (['foo.rdf', 'foo.rdf'],
+                     ['somelongfilename.rdf', 'somelongfilenam...rdf'],
+                     [u'unicode삮.txt', u'unicode\uc0ae.txt'],
+                     [u'unicodesomelong삮.txt', u'unicodesomelong...txt'],
+                     ['somelongfilename.somelongextension',
+                      'somelongfilenam...somelonge..'],):
+            eq_(truncate(x), y)
+
     def test_get_files_not_extracted(self):
         assert not self.viewer.get_files()
 

apps/files/tests/test_views.py

@@ -1,12 +1,13 @@
 import json
+import os
 import shutil
 import tempfile
 
 from django.conf import settings
 from django.core.cache import cache
+from django.utils.encoding import iri_to_uri
 
 from nose.tools import eq_
-import os
 from pyquery import PyQuery as pq
 import test_utils
 
@@ -17,6 +18,10 @@ from files.models import File
 from users.models import UserProfile
 
 
+dictionary = 'apps/files/fixtures/files/dictionary-test.xpi'
+unicode_filenames = 'apps/files/fixtures/files/unicode-filenames.xpi'
+
+
 class FilesBase:
 
     def login_as_editor(self):
@@ -42,7 +47,6 @@ class FilesBase:
         settings.TMP_PATH = tempfile.mkdtemp()
         settings.ADDONS_PATH = tempfile.mkdtemp()
 
-        dictionary = 'apps/files/fixtures/files/dictionary-test.xpi'
         for file_obj in [self.file, self.file_two]:
             src = os.path.join(settings.ROOT, dictionary)
             try:
@@ -240,6 +244,12 @@ class TestFileViewer(FilesBase, test_utils.TestCase):
         res = self.client.get(self.file_url('doesnotexist.js'))
         eq_(res.status_code, 404)
 
+    def test_unicode(self):
+        self.file_viewer.src = unicode_filenames
+        self.file_viewer.extract()
+        res = self.client.get(self.file_url(iri_to_uri(u'\u1109\u1161\u11a9')))
+        eq_(res.status_code, 200)
+
 
 class TestDiffViewer(FilesBase, test_utils.TestCase):
     fixtures = ['base/addon_3615', 'base/users']