Add option to collapse symlinks in isolate.py

client/isolated_format.py

 # Copyright 2014 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.
 
 """Understands .isolated files and can do local operations on them."""
 
 import hashlib
 import json
 import logging
 import os
@@ skipping 291 matching lines @@
           expand_directory_and_symlink(
               indir, relfile, blacklist, follow_symlinks))
     except MappingError as e:
       if not ignore_broken_items:
         raise
       logging.info('warning: %s', e)
   return outfiles
 
 
 @tools.profile
-def file_to_metadata(filepath, prevdict, read_only, algo):
+def file_to_metadata(filepath, prevdict, read_only, algo, collapse_symlinks):
   """Processes an input file, a dependency, and return meta data about it.
 
   Behaviors:
   - Retrieves the file mode, file size, file timestamp, file link
     destination if it is a file link and calcultate the SHA-1 of the file's
     content if the path points to a file and not a symlink.
 
   Arguments:
     filepath: File to act on.
     prevdict: the previous dictionary. It is used to retrieve the cached sha-1
               to skip recalculating the hash. Optional.
     read_only: If 1 or 2, the file mode is manipulated. In practice, only save
                one of 4 modes: 0755 (rwx), 0644 (rw), 0555 (rx), 0444 (r). On
                windows, mode is not set since all files are 'executable' by
                default.
     algo:      Hashing algorithm used.
+    collapse_symlinks: True if symlinked files should be treated like they were
+                       the normal underlying file.
 
   Returns:
     The necessary dict to create a entry in the 'files' section of an .isolated
     file.
   """
   # TODO(maruel): None is not a valid value.
   assert read_only in (None, 0, 1, 2), read_only
   out = {}
   # Always check the file stat and check if it is a link. The timestamp is used
   # to know if the file's content/symlink destination should be looked into.
   # E.g. only reuse from prevdict if the timestamp hasn't changed.
   # There is the risk of the file's timestamp being reset to its last value
   # manually while its content changed. We don't protect against that use case.
   try:
-    filestats = os.lstat(filepath)
+    if collapse_symlinks:
+      # os.stat follows symbolic links
+      filestats = os.stat(filepath)
+    else:
+      # os.lstat does not follow symbolic links, and thus preserves them.
+      filestats = os.lstat(filepath)
   except OSError:
     # The file is not present.
     raise MappingError('%s is missing' % filepath)
   is_link = stat.S_ISLNK(filestats.st_mode)
 
   if sys.platform != 'win32':
     # Ignore file mode on Windows since it's not really useful there.
     filemode = stat.S_IMODE(filestats.st_mode)
     # Remove write access for group and all access to 'others'.
     filemode &= ~(stat.S_IWGRP | stat.S_IRWXO)
@@ skipping 213 matching lines @@
     data['files'] = dict(
         (k.replace(wrong_path_sep, os.path.sep), v)
         for k, v in data['files'].iteritems())
     for v in data['files'].itervalues():
       if 'l' in v:
         v['l'] = v['l'].replace(wrong_path_sep, os.path.sep)
   if 'relative_cwd' in data:
     data['relative_cwd'] = data['relative_cwd'].replace(
         wrong_path_sep, os.path.sep)
   return data