dupekill: Skip files that can't be accessed

Also included checks for dead symlinks and network links that won't
resolve.

Added -a, which will display all processed files. The -v flag will
only output dupes, errors, and notices.

2 files changed, 36 insertions, 4 deletions

diff --git a/TODO b/TODO
new file mode 100644
index 0000000..81cd368
--- /dev/null
+++ b/TODO
@@ -0,0 +1,3 @@
+dupekill
+--------
+* Find a way to indicate which file a dupe is a copy of.
diff --git a/dupekill b/dupekill
index 79d5c04..19d9123 100755
--- a/dupekill
+++ b/dupekill
@@ -44,7 +44,7 @@ def need_to_add(filepath, datalist):
 
     return found != 1
 
-def dupekill(dry_run=False, recursive=False, verbose=False, path=os.getcwd()):
+def dupekill(dry_run=False, recursive=False, verbose=False, all_files=False, path=os.getcwd()):
 
     if not os.path.isdir(path):
         print("Error: Unable to fetch directory to work with.")
@@ -72,8 +72,36 @@ def dupekill(dry_run=False, recursive=False, verbose=False, path=os.getcwd()):
         for item in file:
             # Set up a few variables we'll be needing.
             filepath = os.path.join(root, item)
-            filesize = os.stat(filepath).st_size
+
+            # Check for information. If we can't fetch any, there's likely
+            # a dead symlink or something else wrong.
+            try:
+                filesize = os.stat(filepath).st_size
+            # This occurs when the data points to something that can't be found or
+            # resolved.
+            except IOError:
+                if verbose:
+                    print("NOT FOUND:", filepath)
+                raise
+                break
+
+            # This occurs mostly with symlinks.
+            except OSError:
+                if verbose:
+                    print("DEAD LINK:", filepath)
+                raise
+                break
+
             deleted = False # We need this flag to determine state before adding to the list
+            if not os.access(filepath, os.R_OK):
+                if verbose:
+                    print("NOTICE: Cannot read from", filepath)
+
+                break
+
+            if not os.access(filepath, os.W_OK):
+                if verbose:
+                    print("NOTICE: Cannot write to", filepath)
 
             # Funny, processed_files will always equal the index of the file
             # in our list. :D We might not need it, though
@@ -140,7 +168,7 @@ def dupekill(dry_run=False, recursive=False, verbose=False, path=os.getcwd()):
             if need_to_add(filepath, hashList) and not deleted:
                 hashList.append([filepath, filesize])
 
-                if verbose:
+                if all_files:
                     print("FILE:", filepath)
 
     print()
@@ -158,11 +186,12 @@ if __name__ == '__main__':
         parser.add_option("-d", "--dry", dest='dry_run', action='store_true', default=False, help="displays a list of files dupekill will delete if you run it again without this flag")
         parser.add_option("-r", "--recursive", dest='recursive', action='store_true', default=False, help="Recurses into all directories below the starting point")
         parser.add_option("-v", "--verbose", dest='verbose', action='store_true', default=False, help="Provide more detailed output")
+        parser.add_option("-a", "--all-files", dest='all_files', action='store_true', default=False, help="Show all processed files, not just dupes and errors")
         (options, args) = parser.parse_args()
         if args and os.path.isdir(args[0]):
             path = os.path.abspath(args[0])
         else:
             path = os.getcwd()
-        dupekill(options.dry_run, options.recursive, options.verbose, path)
+        dupekill(options.dry_run, options.recursive, options.verbose, options.all_files, path)
     except KeyboardInterrupt:
         print("Aborted")