1 files changed, 50 insertions, 0 deletions
diff --git a/ftok_experiment.py b/ftok_experiment.py
new file mode 100644
index 0000000..95fcd21
--- /dev/null
+++ b/ftok_experiment.py
@@ -0,0 +1,50 @@
+"""This is an experiment to see how often ftok() returns duplicate keys for
+different filenames. 
+"""
+
+
+import sys
+import os
+import sysv_ipc
+
+if len(sys.argv) == 2:
+    start_path = sys.argv[1]
+else:
+    msg = "Start path? [Default = your home directory] "
+    start_path = raw_input(msg)
+    if not start_path:
+        start_path = "~"
+
+# Expand paths that start with a tilde and then absolutize.
+start_path = os.path.expanduser(start_path)
+start_path = os.path.abspath(start_path)
+
+# For every filename in the tree, generate a key and associate the filename
+# with that key via a dictionary.
+d = { }
+nfilenames = 0
+for path, dirnames, filenames in os.walk(start_path):
+    for filename in filenames:
+        # Fully qualify the path
+        filename = os.path.join(path, filename)
+
+        nfilenames += 1
+
+        #print "Processing %s..." % filename
+
+        key = sysv_ipc.ftok(filename, 42, True)
+
+        if key not in d:
+            d[key] = [ ]
+
+        d[key].append(filename)
+
+# Print statistics, including files with non-unique keys.
+ndups = 0
+for key in d:
+    if len(d[key]) > 1:
+        ndups += len(d[key])
+        print key, d[key]
+
+print "Out of {0} unique filenames, I found {1} duplicate keys.".format(nfilenames, ndups)
+