From 9a0b263b7fc9d43e6acd16729ebfc2120ae501d6 Mon Sep 17 00:00:00 2001
From: Lars Wirzenius <liw@liw.fi>
Date: Wed, 22 Dec 2010 09:04:38 +0000
Subject: Add --remove option.

---
 debian/changelog |  6 ++++++
 dupfiles         | 10 ++++++++++
 dupfiles.1       |  6 +++++-
 3 files changed, 21 insertions(+), 1 deletion(-)

diff --git a/debian/changelog b/debian/changelog
index df2a680..76a87d2 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,9 @@
+dupfiles (5.1) squeeze; urgency=low
+
+  * Add --remove option.
+
+ -- Lars Wirzenius <liw@liw.fi>  Wed, 22 Dec 2010 09:04:17 +0000
+
 dupfiles (5) squeeze; urgency=low
 
   * Use python-ttystatus for progress reporting.
diff --git a/dupfiles b/dupfiles
index 8c06c3d..3d9e8db 100755
--- a/dupfiles
+++ b/dupfiles
@@ -197,6 +197,12 @@ def make_hardlinks(duplicates):
         os.link(canonical, pathname)
 
 
+def remove_all_but_one(duplicates):
+    keep = duplicates.pop()
+    for pathname in duplicates:
+        os.remove(pathname)
+
+
 def report(duplicates):
     sys.stdout.write('\n'.join(duplicates))
     sys.stdout.write('\n\n')
@@ -208,6 +214,8 @@ def main():
                       help='hardlink duplicate files to each other')
     parser.add_option('--progress', action='store_true',
                       help='report progress')
+    parser.add_option('--remove', action='store_true',
+                      help='remove all but one copy of identical files')
 
     opts, args = parser.parse_args()
 
@@ -218,6 +226,8 @@ def main():
     for duplicates in dupfinder.duplicates():
         if opts.make_hardlinks:
             make_hardlinks(duplicates)
+        elif opts.remove:
+            remove_all_but_one(duplicates)
         else:
             report(duplicates)
 
diff --git a/dupfiles.1 b/dupfiles.1
index 1bcf159..ee66d9b 100644
--- a/dupfiles.1
+++ b/dupfiles.1
@@ -20,6 +20,7 @@ dupfiles \- find identical files and optionally hardlink them
 .B dupfiles
 .RB [ --make-hardlinks ]
 .RB [ --progress ]
+.RB [ --remove ]
 .RI [ file ]...
 .SH DESCRIPTION
 .B dupfiles
@@ -43,7 +44,10 @@ Report progress while program is running.
 .BR --make-hardlinks
 Instead of reporting files that have been found,
 make sure all names point at the same content.
-FIXME.
+.TP
+.BR --remove
+Remove all but one copy of identical files.
+The copy that gets kept is chosen arbitrarily.
 .SH EXAMPLE
 To find all duplicate packages in Debian and Ubuntu:
 .IP
-- 
cgit v1.2.1