The clean script can now also delete directories.

BUG=
TEST=

Review URL: https://webrtc-codereview.appspot.com/399004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@1763 4adac7df-926f-26a2-2b94-8c16560cd09d

tools/continuous_build/clean_old_files.py

@@ -9,6 +9,16 @@
 
 __author__ = "ivinnichenko@webrtc.org (Illya Vinnichenko)"
 
+"""This script will prune sufficiently old files and empty directories.
+
+   The algorithm is to look into the provided directory and delete any files
+   that is older than x days, recursively. Then all empty directories will be
+   deleted (we can't look at timestamps there since the act of deleting a file
+   will refresh the directory's timestamp).
+
+   Note: This script has only been tested on Linux.
+"""
+
 from optparse import OptionParser
 import os
 import sys
@@ -23,7 +33,8 @@ WHITELIST = ["buildbot.tac", "master.cfg", "public_html", "changes.pck",
 def is_whitelisted(path):
   """Check if file is whitelisted.
 
-    path: file path.
+    Args:
+      path: file path.
   """
   for entry in WHITELIST:
     if entry in path:
@@ -31,38 +42,79 @@ def is_whitelisted(path):
   return False
 
 
-def remove_old_filenames(path, num_days, verbose):
-  """Remove old files.
+def delete_directory(directory):
+  try:
+    os.rmdir(directory)
+    return True
+  except OSError as exception:
+    # The directory probably contains newer files.
+    print "Could not remove directory %s: reason %s." % (directory, exception)
+  return False
 
-    path: base directory for removal.
-    num_days: days limit for removal.
-    verbose: print every cmd?
+
+def delete_file(file):
+  try:
+    os.remove(file)
+  except OSError as exception:
+    print "Unexpectedly failed to remove file %s: reason %s." % (file,
+                                                                 exception)
+
+
+def log_removal(file_or_directory, time_stamp, verbose):
+  if verbose:
+    str_stamp = time.strftime("%a, %d %b %Y %H:%M:%S +0000",
+                              time.gmtime(time_stamp))
+    print "Removing [%s], stamped on %s" % (file_or_directory, str_stamp)
+
+
+def remove_old_files_and_directories(path, num_days, verbose, skip_dirs):
+  """Removes all files under path that are older than num_days days.
+     The algorithm also tried to delete all directories, except for those who
+     contain files that are sufficiently new.
+
+     Implementation note: it doesn't make sense to look at timestamps for
+     directories since their timestamps are updated when a file is deleted.
+
+    Args:
+      path: The starting point.
+      num_days: days limit for removal.
+      verbose: print every cmd?
   """
-  print "Cleaning up everything in %s older than %s days" % (path, num_days)
   current_time = time.time()
-  limit = 60*60*24*num_days
-  for root, unused_dirs, files in os.walk(path):
-    for filename in files:
-      current_file = os.path.join(root, filename)
+  limit = 60 * 60 * 24 * num_days
+
+  # Walk bottom-up so directories are deleted in the right order.
+  for root, directories, files in os.walk(path, topdown=False):
+    for file in files:
+      current_file = os.path.join(root, file)
+      time_stamp = os.stat(current_file).st_mtime
+
       if is_whitelisted(current_file):
         continue
-      time_stamp = os.stat(current_file).st_mtime
+
       if (current_time - time_stamp) > limit:
-        str_stamp = time.strftime("%a, %d %b %Y %H:%M:%S +0000",
-                    time.gmtime(time_stamp))
-        if verbose:
-          print "Removing [%s], stamped on %s" % (current_file, str_stamp)
-        os.remove(current_file)
+        delete_file(current_file)
+        log_removal(current_file, time_stamp, verbose)
+
+    if not skip_dirs:
+      for directory in directories:
+        current_directory = os.path.join(root, directory)
+        time_stamp = os.stat(current_directory).st_mtime
+        if delete_directory(current_directory):
+          log_removal(current_directory, time_stamp, verbose)
 
 
 def main():
-  usage = "usage: %prog [options] arg"
+  usage = "usage: %prog -p <base path> -n <number of days> [-q] [-d]"
   parser = OptionParser(usage)
   parser.add_option("-p", "--path", dest="cleanup_path", help="base directory")
   parser.add_option("-n", "--num_days", dest="num_days", help="number of days")
   parser.add_option("-q", "--quiet",
                     action="store_false", dest="verbose", default=True,
                     help="don't print status messages to stdout")
+  parser.add_option("-d", "--delete-dirs-too",
+                    action="store_false", dest="skip_dirs", default=True,
+                    help="number of days")
 
   options, args = parser.parse_args()
   if not options.cleanup_path:
@@ -71,8 +123,12 @@ def main():
   if not options.num_days:
     print "You must specify number of days old"
     sys.exit(2)
-  remove_old_filenames(options.cleanup_path, int(options.num_days),
-                       options.verbose)
+
+  if options.verbose:
+    print "Cleaning up everything in %s older than %s days" % (
+        options.cleanup_path, options.num_days)
+  remove_old_files_and_directories(options.cleanup_path, int(options.num_days),
+                                   options.verbose, options.skip_dirs)
 
 if __name__ == "__main__":
   main()