luci-py: Making __file__ usage unicode safe.

client/utils/zip_package.py

 # Copyright 2013 The LUCI Authors. All rights reserved.
 # Use of this source code is governed under the Apache License, Version 2.0
 # that can be found in the LICENSE file.
 
 """Utilities to work with importable python zip packages."""
 
 import atexit
 import collections
 import cStringIO as StringIO
 import hashlib
@@ skipping 17 matching lines @@
   r'.*\.pyc$',
   r'.*\.pyo$',
 )
 
 
 # Temporary files extracted by extract_resource. Removed in atexit hook.
 _extracted_files = []
 _extracted_files_lock = threading.Lock()
 
 
+# Patch zipimport.zipimporter hook to accept unicode strings
+def zipimporter_unicode(archivepath):
+  if isinstance(archivepath, unicode):
+    archivepath = archivepath.encode(sys.getfilesystemencoding())
+  return zipimport.zipimporter(archivepath)
+
+
+for i, hook in enumerate(sys.path_hooks):
+  if hook is zipimport.zipimporter:
+    sys.path_hooks[i] = zipimporter_unicode
+
+
 class ZipPackageError(RuntimeError):
   """Failed to create a zip package."""
 
 
 class ZipPackage(object):
   """A set of files that can be zipped to file on disk or into memory buffer.
 
   Usage:
     package = ZipPackage(root)
     package.add_file('some_file.py', '__main__.py')
@@ skipping 183 matching lines @@
 def get_main_script_path():
   """If running from zip returns path to a zip file, else path to __main__.
 
   Basically returns path to a file passed to python for execution
   as in 'python <main_script>' considering a case of executable zip package.
 
   Returns path relative to a current directory of when process was started.
   """
   # If running from interactive console __file__ is not defined.
   main = sys.modules['__main__']
-  return get_module_zip_archive(main) or getattr(main, '__file__', None)
+  path = get_module_zip_archive(main)
+  if path:
+    return path
+
+  path = getattr(main, '__file__', None)
+  if path:
+    return path.decode(sys.getfilesystemencoding())
 
 
 def _write_temp_data(name, data, temp_dir):
   """Writes content-addressed file in `temp_dir` if relevant."""
   filename = '%s-%s' % (hashlib.sha1(data).hexdigest(), name)
   filepath = os.path.join(temp_dir, filename)
   if os.path.isfile(filepath):
     with open(filepath, 'rb') as f:
       if f.read() == data:
         # It already exists.
@@ skipping 20 matching lines @@
     package: is a python module object that represents a package.
     resource: should be a relative filename, using '/'' as the path separator.
     temp_dir: if set, it will extra the file in this directory with the filename
         being the hash of the content. Otherwise, it uses tempfile.mkstemp().
 
   Raises ValueError if no such resource.
   """
   # For regular non-zip packages just construct an absolute path.
   if not is_zipped_module(package):
     # Package's __file__ attribute is always an absolute path.
-    path = os.path.join(os.path.dirname(package.__file__),
+    ppath = package.__file__.decode(sys.getfilesystemencoding())
+    path = os.path.join(os.path.dirname(ppath),
         resource.replace('/', os.sep))
     if not os.path.exists(path):
       raise ValueError('No such resource in %s: %s' % (package, resource))
     return path
 
   # For zipped packages extract the resource into a temp file.
   data = pkgutil.get_data(package.__name__, resource)
   if data is None:
     raise ValueError('No such resource in zipped %s: %s' % (package, resource))
 
@@ skipping 43 matching lines @@
   h = hashlib.sha1()
   with zipfile.ZipFile(get_main_script_path(), 'r') as z:
     for name in sorted(z.namelist()):
       with z.open(name) as f:
         h.update(str(len(name)))
         h.update(name)
         content = f.read()
         h.update(str(len(content)))
         h.update(content)
   return h.hexdigest()