Re: [cp-patches] GNU Crypto and Jessie merge

[Casey Marshall]

On Jan 6, 2006, at 5:41 PM, Raif S. Naffah wrote:

> On Saturday 07 January 2006 11:32, Casey Marshall wrote:
> > You have to do that if you want to present a unified factory
> > interface to (e.g.) all PRNG implementations. But we don't have to do
> > that, here, because there already is a unified factory for PRNGs:
> > java.security.SecureRandom, and that's the one that makes sense for
> > Classpath. The weak provider queries the weak PRNG factory, and the
> > strong provider queries the strong factory.
>
> i was under the impression that it would be desirable to have that
> behind the JCE facade so that internally to Classpath, all
> crypto-related classes can get what they need without using the JCE
> mechanism.

Yes, it would be nice to have Classpath's crypto internals (based on  
GNU Crypto and Jessie) had a decent API for maintainers to use, but  
it's something I'm willing to sacrifice to more cleanly split weak/ 
strong crypto packages. I'm saying that since this API is internal,  
it is OK if it's a little rough in places.

HOWEVER, if we write all Classpath-internal classes to use the JCE  
whenever possible, the code is a lot more portable, and won't depend  
on any specific crypto implementation.

> > I'm talking about GNU-Crypto-in-Classpath, though, so why does a
> > Factory make sense in that context? It seems to me like it's little
> > more than an additional level of indirection (security-getInstance ->
> > wrapper -> factory-getInstance -> class, instead of security-
> > getInstance -> wrapper -> class).
>
> is not that what the current JCE Adapters do, or are you suggesting  
> that
> this will change?

I'm saying that the `Factory.getInstance' call is another method  
call, which isn't strictly necessary. That is, when I need a cipher  
in an adapter (or anywhere else), if I do

  new CipherClass();

instead of

  Factory.getInstance ("Cipher");

I'm saving a method call. Nothing more, since the factory's  
`getInstance' is itself just calling the constructor (it also calls  
the `selfTest' method for the algorithm, but I'm less convinced of  
the utility of that).

This is headed to bikeshed-territory, and this particular issue isn't  
that big a concern for me.

Thanks.