From 58398ff7e4b47ec39a9106e193c2ea54eda4f723 Mon Sep 17 00:00:00 2001
From: Lars Wirzenius <liw@xander>
Date: Fri, 26 Nov 2010 18:29:48 +0200
Subject: Optionall run program under profiling.

---
 dupfiles | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/dupfiles b/dupfiles
index b1de7fc..1956c1d 100755
--- a/dupfiles
+++ b/dupfiles
@@ -228,4 +228,9 @@ def main():
 
 
 if __name__ == '__main__':
-    main()
+    profname = os.environ.get('DUPFILES_PROFILE')
+    if profname:
+        import cProfile
+        cProfile.run('main()', profname)
+    else:
+        main()
-- 
cgit v1.2.1


From 279879c20a82614f3c323be77e1d4af601bdb70f Mon Sep 17 00:00:00 2001
From: Lars Wirzenius <liw@xander>
Date: Sun, 28 Nov 2010 00:14:37 +0200
Subject: Use os.read for speed.

---
 dupfiles | 20 +++++++++++---------
 1 file changed, 11 insertions(+), 9 deletions(-)

diff --git a/dupfiles b/dupfiles
index 1956c1d..1c2550d 100755
--- a/dupfiles
+++ b/dupfiles
@@ -58,30 +58,31 @@ class File(object):
 
     def read(self, num_bytes):
         if self.pathname in filepool:
-            f = filepool[self.pathname]
+            fd = filepool[self.pathname]
             filestats.hit_count += 1
         else:
             try:
-                f = file(self.pathname)
-            except IOError, e:
+                fd = os.open(self.pathname, os.O_RDONLY)
+            except OSError, e:
                 if e.errno != errno.EMFILE:
                     raise
                 victim = random.choice(filepool.keys())
-                filepool[victim].close()
+                os.close(filepool[victim])
                 del filepool[victim]
                 filestats.close_count += 1
-                f = file(self.pathname)
-            f.seek(self.offset)
-            filepool[self.pathname] = f
+                fd = os.open(self.pathname, os.O_RDONLY)
+            os.lseek(fd, self.offset, os.SEEK_SET)
+            filepool[self.pathname] = fd
             filestats.open_count += 1
-        data = f.read(num_bytes)
+        data = os.read(fd, num_bytes)
         self.offset += len(data)
         return data
 
     def close(self):
         if self.pathname in filepool:
-            filepool[self.pathname].close()
+            os.close(filepool[self.pathname])
             del filepool[self.pathname]
+            filestats.close_count += 1
 
 
 class DuplicateFileFinder(object):
@@ -220,6 +221,7 @@ def main():
     for dirname in sorted(args):
         dupfinder.collect(dirname)
 
+
     for duplicates in dupfinder.duplicates():
         if opts.make_hardlinks:
             make_hardlinks(duplicates)
-- 
cgit v1.2.1


From ee06c931432ed23019e34baf66627ae3269ed998 Mon Sep 17 00:00:00 2001
From: Lars Wirzenius <liw@xander>
Date: Sun, 28 Nov 2010 10:16:20 +0200
Subject: Read only 4 KiB, for speed.

---
 dupfiles | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/dupfiles b/dupfiles
index 1c2550d..7c8af79 100755
--- a/dupfiles
+++ b/dupfiles
@@ -185,7 +185,7 @@ class DuplicateFileFinder(object):
         
         done = []
         not_done = []
-        chunk_size = 64 * 1024
+        chunk_size = 4 * 1024
         for pathname, f in group:
             data = f.read(chunk_size)
             if not data:
-- 
cgit v1.2.1