From c3988f73004cbf327ea14372d5edda7404b68087 Mon Sep 17 00:00:00 2001
From: Adrian Sampson <adrian@radbox.org>
Date: Fri, 6 Aug 2010 11:17:57 -0700
Subject: [PATCH] safely interpret integers in packed values

---
 NEWS                   |  1 +
 beets/mediafile.py     | 78 +++++++++++++++++++++++++-----------------
 test/test_mediafile.py | 26 ++++++++++++++
 3 files changed, 74 insertions(+), 31 deletions(-)

diff --git a/NEWS b/NEWS
index 4871d5c8c..1a166a1b0 100644
--- a/NEWS
+++ b/NEWS
@@ -45,6 +45,7 @@
   Windows users can now just type "beet" at the prompt to run beets.
 * Fixed an occasional bug where Mutagen would complain that a tag was
   already present.
+* Fixed a bug with reading invalid integers from ID3 tags.
 
 1.0b3
 -----
diff --git a/beets/mediafile.py b/beets/mediafile.py
index f2ba9f131..dd5fef428 100644
--- a/beets/mediafile.py
+++ b/beets/mediafile.py
@@ -62,6 +62,50 @@ TYPES = {
 }
 
 
+# Utility.
+
+def _safe_cast(out_type, val):
+    """Tries to covert val to out_type but will never raise an
+    exception. If the value can't be converted, then a sensible
+    default value is returned. out_type should be bool, int, or
+    unicode; otherwise, the value is just passed through.
+    """
+    if out_type == int:
+        if val is None:
+            return 0
+        elif isinstance(val, int) or isinstance(val, float):
+            # Just a number.
+            return int(val)
+        else:
+            # Process any other type as a string.
+            if not isinstance(val, basestring):
+                val = unicode(val)
+            # Get a number from the front of the string.
+            val = re.match('[0-9]*', val.strip()).group(0)
+            if not val:
+                return 0
+            else:
+                return int(val)
+
+    elif out_type == bool:
+        if val is None:
+            return False
+        else:
+            try:
+                # Should work for strings, bools, ints:
+                return bool(int(val)) 
+            except ValueError:
+                return False
+
+    elif out_type == unicode:
+        if val is None:
+            return u''
+        else:
+            return unicode(val)
+
+    else:
+        return val
+
 # Flags for encoding field behavior.
 
 class Enumeration(object):
@@ -146,9 +190,9 @@ class Packed(object):
             out = None
     
         if out is None or out == self.none_val or out == '':
-            return self.out_type(self.none_val)
+            return _safe_cast(self.out_type, self.none_val)
         else:
-            return self.out_type(out)
+            return _safe_cast(self.out_type, out)
     
     def __setitem__(self, index, value):
 
@@ -336,35 +380,7 @@ class MediaField(object):
         if style.packing:
             out = Packed(out, style.packing)[style.pack_pos]
         
-        # return the appropriate type
-        if self.out_type == int:
-            if out is None:
-                return 0
-            elif isinstance(out, int) or isinstance(out, float):
-                # Just a number.
-                return int(out)
-            else:
-                # Process any other type as a string.
-                if not isinstance(out, basestring):
-                    out = unicode(out)
-                # Get a number from the front of the string.
-                out = re.match('[0-9]*', out.strip()).group(0)
-                if not out:
-                    return 0
-                else:
-                    return int(out)
-        elif self.out_type == bool:
-            if out is None:
-                return False
-            else:
-                return bool(int(out)) # should work for strings, bools, ints
-        elif self.out_type == unicode:
-            if out is None:
-                return u''
-            else:
-                return unicode(out)
-        else:
-            return out
+        return _safe_cast(self.out_type, out)
     
     def __set__(self, obj, val):
         """Set the value of this metadata field.
diff --git a/test/test_mediafile.py b/test/test_mediafile.py
index 84a491b83..114ed177c 100644
--- a/test/test_mediafile.py
+++ b/test/test_mediafile.py
@@ -61,6 +61,32 @@ class EdgeTest(unittest.TestCase):
         self.assertEqual(f.disc, 4)
         self.assertEqual(f.disctotal, 5)
 
+_sc = beets.mediafile._safe_cast
+class InvalidValueToleranceTest(unittest.TestCase):
+    def test_packed_integer_with_extra_chars(self):
+        pack = beets.mediafile.Packed("06a", beets.mediafile.packing.SLASHED)
+        self.assertEqual(pack[0], 6)
+
+    def test_packed_integer_invalid(self):
+        pack = beets.mediafile.Packed("blah", beets.mediafile.packing.SLASHED)
+        self.assertEqual(pack[0], 0)
+
+    def test_packed_index_out_of_range(self):
+        pack = beets.mediafile.Packed("06", beets.mediafile.packing.SLASHED)
+        self.assertEqual(pack[1], 0)
+
+    def test_safe_cast_string_to_int(self):
+        self.assertEqual(_sc(int, 'something'), 0)
+
+    def test_safe_cast_int_string_to_int(self):
+        self.assertEqual(_sc(int, '20'), 20)
+
+    def test_safe_cast_string_to_bool(self):
+        self.assertEqual(_sc(bool, 'whatever'), False)
+
+    def test_safe_cast_intstring_to_bool(self):
+        self.assertEqual(_sc(bool, '5'), True)
+
 class SafetyTest(unittest.TestCase):
     def _exccheck(self, fn, exc):
         fn = os.path.join('rsrc', fn)