Fix checksum calculation in CAB data records.

[MS-CAB] is pretty unclearly worded in this area. Turns out that if
the sequence of bytes being checksummed is not a multiple of 4 bytes,
you are supposed to _first_ reverse the order of the trailing (n % 4)
bytes, and _then_ append zero bytes to pad to a multiple of 4, before
you do the main checksum operation of XORing together all the 32-bit
words of the input.

1 files changed, 5 insertions, 1 deletions

diff --git a/makecab.py b/makecab.py
index befa94a..77c1e4f 100755
--- a/makecab.py
+++ b/makecab.py
@@ -34,7 +34,11 @@ def packtime(h,m,s):
     return ((h << 11) | (m << 5) | (s >> 1))
 
 def checksum(data):
-    data += "\0" * (3 & -len(data)) # pad to multiple of 4 bytes
+    data_full_words = data[:len(data) & ~3]
+    data_last_word = data[len(data_full_words):]
+    data_last_word = "".join(reversed(data_last_word))
+    data_last_word += "\0" * (3 & -len(data)) # pad to multiple of 4 bytes
+    data = data_full_words + data_last_word
     toret = 0
     for offset in xrange(0, len(data), 4):
         toret ^= struct.unpack_from("<L", data, offset)[0]