[Mono-bugs] [Bug 609376] HashAlgorithm subclasses function differently than on MS .NET

[bugzilla_noreply at novell.com]

http://bugzilla.novell.com/show_bug.cgi?id=609376

http://bugzilla.novell.com/show_bug.cgi?id=609376#c4


--- Comment #4 from Juho Vähä-Herttua <juhovh at iki.fi> 2010-05-27 18:33:19 UTC ---
(In reply to comment #3)
> why ? because it's a bug ;-) and that is likely undocumented (or was back when
> we implemented this part of Mono)

It's indeed true that the whole System.Security.Cryptography is lacking
documentation from Microsoft's part.

> Seriously I gave you a workaround but I did not close the bug report (instead I
> assigned it to myself). That means it will get fixed in a future release of
> Mono.

Ok, I missed the assignation part and got the impression that the bug is not
relevant unless otherwise proven. I'm not in a hurry with a fix, because it
doesn't affect correct code, just wish that it wouldn't be forgotten. Thank
you.

> Now, like any issue, it can cause compatibility problems (but it won't be
> random ones). In fact you noticed something that no one reported in more than 7
> years :-) OTOH that's likely because most code is ported from Windows to Linux
> (and not Linux to Windows) where Initialize would need to be called.

With seemingly random I meant some code paths using TransformBlock and some
code paths using ComputeHash, making it fail on the first hash after the first
TransformFinalBlock. But of course it's easily debugged, especially with the
quite clear error message given in exception.

Personally I always port my code from OSX/Linux to Windows, but might be better
off the other way around in the .NET world. :P

-- 
Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.