I believe the problem is with padding. One of these (don't remember which
one - Java or Crypto++) does not use the current revision, so they produce
encryption incompatible with each other.

Norbert, it would be interesting to see how AES-CFB fares - because it does
not need padding.

> -----Original Message-----
> From: Jeffrey Walton [mailto:[EMAIL PROTECTED] 
> Sent: Tuesday, November 14, 2006 13:25
> To: [email protected]
> Subject: Re: using AES with JAVA and C++
> 
> Hi Norbert,
> 
> > Crypto ++ Version 4.2, because I'm not able to get
> > 5.1 (or 5.2) running on our VS 6.0 project
> I would recommend getting this issue resolved first. Have you 
> applied the processor pack? Also, a static link may be easier 
> if you do not require FIPS certification.  Dor the basic "how 
> to Compile and use a Static Library", see 
> http://www.codeguru.com/article-preview.php/12799
> 
> > encrypt in Java  -> decrypt in Java    ok   (with CBC or 
> some other block
> > SNIP...
> > encrypt in C++ -> decrypt in Java    fail   (except ECB 
> with nopadding)
> Examples of your code (which compiles and fails) would be nice.
> 
> The mailing list has an example of Java/Crypto++ 
> interoperability, but I don't recall the details.
> 
> Jeff
> 
> On 11/14/06, Norbert Thek <[EMAIL PROTECTED]> wrote:
> > I know
> >
> > this isn't the problem
> >
> > encrypt in Java  -> decrypt in Java    ok   (with CBC or 
> some other block
> > cipher)
> > encrypt in C++ -> decrypt in C++     ok   (with CBC or some 
> other block
> > cipher)
> >
> >  encrypt in Java -> decrypt in C++    fail   (except ECB 
> with nopadding)
> > encrypt in C++ -> decrypt in Java    fail   (except ECB 
> with nopadding)
> >
> > The problem with ECB with nopadding is...
> >
> > that if encrypt for example a binary where most of the data is NULL 
> > the ecnrypted has lots of repeated data exampe Hex 00 00 00 
> 00 00 00 
> > 00 00 00 ...00 00 00 00 00 00 00 becomes to Hex 01 EA EA 2B 
> 01 EA EA 
> > 2B ...  01 EA EA 2B 01 EA EA 2B (no real data, only a example)
> >
> > I forgot to mention I use Crypto ++ Version 4.2, because 
> I'm not able 
> > to get
> > 5.1 (or 5.2) running on our
> > VS 6.0 project  ( i get a lot of linker errors, but working 
> with 4.2 
> > is ok
> > !?!?!??!)
> >
> > If somebody can give me some example
> > How JAVA and Crypto CPP can work together  (with AES, if 
> possible) I 
> > would be very glad :-)
> >
> > regards
> > Norbert
> >
> >
> > 2006/11/14, Jeffrey Walton < [EMAIL PROTECTED]>:
> > > Hi Norbert,
> > >
> > > The FAQ has a few examples of Block Cipher use with AES as a
> > demonstration.
> > >
> > > Jeff
> > >
> > >
> > http://www.eskimo.com/~weidai/cgi-bin/fom-serve/cache/79.html
> > >
> > > On 11/14/06, Norbert Thek < [EMAIL PROTECTED]> wrote:
> > > > Hello
> > > >
> > > > This question was asked several days ago, but I'm not 
> able to find 
> > > > the
> > mails
> > > > anymore and the mailist archive is also not working.
> > > > ( http://www.mail-archive.com/[email protected]
> > isn't
> > > > working)
> > > >
> > > > I tried to use the build in JAVA Cipher but the only 
> way to get it
> > working
> > > > was by using
> > > > "Cipher.getInstance ("AES/ECB/NoPadding")"
> > > >
> > > > But this is not very clever  (I want to crypt a 
> licensfile which 
> > > > have
> > about
> > > > 3 kb or little less)
> > > >
> > > > Can somebody tell, which padding I can use to let java and C++ 
> > > > work together?
> > > >
> > > > It would be optimal if someone can send some democode!
> > > >
> > > >
> > > > regards
> > > > Norbert
> > >
> > >
> >
> >
> 

Reply via email to