[Mono-bugs] [Bug 80439][Nor] Changed - symmetric decryption fails

[bugzilla-daemon at bugzilla.ximian.com]

Please do not reply to this email- if you want to comment on the bug, go to the
URL shown below and enter your comments there.

Changed by sebastien at ximian.com.

http://bugzilla.ximian.com/show_bug.cgi?id=80439

--- shadow/80439	2007-01-04 05:46:12.000000000 -0500
+++ shadow/80439.tmp.25675	2007-01-04 10:38:02.000000000 -0500
@@ -1,13 +1,13 @@
 Bug#: 80439
 Product: Mono: Class Libraries
 Version: 1.2
 OS: unknown
 OS Details: 
-Status: RESOLVED   
-Resolution: FIXED
+Status: REOPENED   
+Resolution: 
 Severity: Unknown
 Priority: Normal
 Component: CORLIB
 AssignedTo: mono-bugs at ximian.com                            
 ReportedBy: lauren.bedoule at gmail.com               
 QAContact: mono-bugs at ximian.com
@@ -68,6 +68,26 @@
 
 ------- Additional Comments From lauren.bedoule at gmail.com  2007-01-04 05:46 -------
 By reducing the input count, it seems to me you won't buffer the right
 block for next transformation for the special case when you need to
 KeepLastBlock (not ECB, padding on, decryption). Does your fix work
 work with two successives transformations for that case?
+
+------- Additional Comments From sebastien at ximian.com  2007-01-04 10:38 -------
+Well it looks like it's more complex than that (again ;-). The only
+time you can call TransformBlock with the "wrong/smaller"(*) value is
+at the end of the decryption process (i.e. there cannot be subsequent
+calls to TransformBlock, well unless you have a working test case ?).
+
+(*) I already consider the previous fix "a bug" we introduce to be
+compatible with MS implementation. IMO the "right" behaviour is to
+thrown an exception. All decryption should (1) use TransformBlock and
+finish with (2) a single call to TransformFinalBlock. In fact because
+each transform could be different code (and because this case is
+undocumented) it has a bug "real-life" chance of not working for
+non-Fx provided symmetric algorithms.
+
+I appreciate your feedback :) but please use the mailing-lists for
+discussions about untested/potential issues or (better) open new (or
+re-open existing) bugzilla entries for bugs (with sample code showing
+the problem). Adding comments to closed issues has a high probability
+to get "lost" (as it won't show up in pre-defined queries).