Add --min-size and --max-size options

2 files changed, 24 insertions, 3 deletions

diff --git a/NEWS b/NEWS
index 73d2b07..ef717ec 100644
--- a/NEWS
+++ b/NEWS
@@ -1,6 +1,11 @@
 NEWS for dupfiles
 =================
 
+Version 5.5, released UNRELEASED
+--------------------------------
+
+* Add `--min-size` and `--max-size` options.
+
 Version 5.4, released 2012-02-09
 --------------------------------
 
diff --git a/dupfiles b/dupfiles
index a4877d5..87862e8 100755
--- a/dupfiles
+++ b/dupfiles
@@ -79,9 +79,15 @@ class File(object):
 
 class DuplicateFileFinder(object):
 
-    def __init__(self, progress):
+    def __init__(self, progress, min_size, max_size):
         self.by_size = dict()
         self.progress = progress
+        self.min_size = min_size
+        self.max_size = max_size
+
+    def size_ok(self, size):
+        return (self.min_size <= size <= self.max_size or
+                (size >= self.min_size and self.max_size == -1))
 
     def collect(self, root):
         ts = ttystatus.TerminalStatus()
@@ -98,7 +104,7 @@ class DuplicateFileFinder(object):
                 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:
+                    if self.size_ok(st.st_size) and st.st_size in self.by_size:
                         self.by_size[st.st_size].append(t)
                     else:
                         self.by_size[st.st_size] = [t]
@@ -206,9 +212,19 @@ class Dupfiles(cliapp.Application):
         self.settings.boolean(['no-act', 'dry-run', 'pretend'],
                               'do not really remove, just report what would '
                                 'be removed')
+        self.settings.bytesize(['min-size'], 
+                               'compare files at least SIZE in size',
+                               metavar='SIZE')
+        self.settings.bytesize(['max-size'], 
+                               'compare files at most SIZE in size '
+                                '(-1 for infinity)',
+                               metavar='SIZE',
+                               default=-1)
 
     def process_args(self, args):
-        dupfinder = DuplicateFileFinder(self.settings['progress'])
+        dupfinder = DuplicateFileFinder(self.settings['progress'],
+                                        self.settings['min-size'],
+                                        self.settings['max-size'])
         for dirname in sorted(args):
             dupfinder.collect(dirname)