Provide some progress reporting for the collection phase.

1 files changed, 33 insertions, 4 deletions

diff --git a/dupfiles b/dupfiles
index 950c170..039ecdd 100755
--- a/dupfiles
+++ b/dupfiles
@@ -21,15 +21,40 @@ import hashlib
 import optparse
 import os
 import sys
+import time
+
+
+class ProgressReporter(object):
+
+    def __init__(self, do_report):
+        self.written = ''
+        self.when = 0
+        self.do_report = do_report
+        
+    def write(self, msg):
+        if self.do_report and time.time() - self.when >= 1:
+            sys.stdout.flush()
+            sys.stderr.write('\b \b' * len(self.written))
+            msg = msg[:79] # FIXME: use real screen width
+            sys.stderr.write(msg)
+            sys.stderr.flush()
+            self.written = msg
+            self.when = time.time()
+
+    def finished(self):
+        self.write('')
 
 
 class DuplicateFileFinder(object):
 
-    def __init__(self):
+    def __init__(self, progress):
         self.by_size = dict()
+        self.progress = progress
         
     def collect(self, root):
         for dirname, subdirs, filenames in os.walk(root):
+            self.progress.write(dirname)
+            subdirs.sort()
             pathnames = [os.path.join(dirname, f) for f in filenames]
             for pathname in pathnames:
                 stat = os.stat(pathname)
@@ -40,6 +65,7 @@ class DuplicateFileFinder(object):
                 else:
                     self.by_size[stat.st_size] = (stat.st_dev, stat.st_ino,
                                                   set([pathname]))
+        self.progress.finished()
 
     def duplicates(self):
         for dev, ino, pathnames in self.by_size.itervalues():
@@ -66,18 +92,21 @@ def make_hardlinks(duplicates):
 
 def report(duplicates):
     sys.stdout.write('\n'.join(duplicates))
-    sys.stdout.write('\n')
+    sys.stdout.write('\n\n')
 
 
 def main():
     parser = optparse.OptionParser()
     parser.add_option('--make-hardlinks', action='store_true',
                       help='hardlink duplicate files to each other')
+    parser.add_option('--progress', action='store_true',
+                      help='report progress')
 
     opts, args = parser.parse_args()
 
-    dupfinder = DuplicateFileFinder()
-    for dirname in args:
+    progress = ProgressReporter(opts.progress)
+    dupfinder = DuplicateFileFinder(progress)
+    for dirname in sorted(args):
         dupfinder.collect(dirname)
     for duplicates in dupfinder.duplicates():
         if opts.make_hardlinks: