Only investigate regular files, ignoring e.g. symlinks.

Also, sort filenames, it's nicer for people.

1 files changed, 9 insertions, 6 deletions

diff --git a/dupfiles b/dupfiles
index 1aa9dc0..3822eaa 100755
--- a/dupfiles
+++ b/dupfiles
@@ -20,6 +20,7 @@
 import hashlib
 import optparse
 import os
+import stat
 import sys
 import time
 
@@ -56,14 +57,16 @@ class DuplicateFileFinder(object):
         for dirname, subdirs, filenames in os.walk(root):
             self.progress.write(dirname)
             subdirs.sort()
+            filenames.sort()
             pathnames = [os.path.join(dirname, f) for f in filenames]
             for pathname in pathnames:
-                stat = os.lstat(pathname)
-                t = (stat.st_dev, stat.st_ino, pathname)
-                if stat.st_size in self.by_size:
-                    self.by_size[stat.st_size].append(t)
-                else:
-                    self.by_size[stat.st_size] = [t]
+                st = os.lstat(pathname)
+                if stat.S_ISREG(st.st_mode):
+                    t = (st.st_dev, st.st_ino, pathname)
+                    if st.st_size in self.by_size:
+                        self.by_size[st.st_size].append(t)
+                    else:
+                        self.by_size[st.st_size] = [t]
         self.progress.finished()
 
     def duplicates(self):