Re: [boost] Crypto Proposal

Subject: Re: [boost] Crypto Proposal
From: Sohail Somani (sohail_at_[hidden])
Date: 2010-04-11 21:07:40

On 11/04/2010 7:25 PM, Rob Riggs wrote:
> On 04/11/2010 10:53 AM, Chad Seibert wrote:
>> Here's an example of boostified code, from one of the examples on the
>> site. It's not too different from the original code (it's just
>> commented out); this is because I don't want it to be a chore to port
>> from Botan to Boost.Botan. Also, it is intuitively designed and I see
>> no reason to change it. Of course, if it needs redesigning, we can
>> discuss that afterwards
> ...
>>        //std::auto_ptr<PKCS8_PrivateKey>
>> privkey(PKCS8::load_key(arg_ca_key, rng, arg_passphrase));
>>        std::auto_ptr<boost::botan::pkcs8::private_key>
>> privkey(pkcs8::load_key(arg_ca_key, rng, arg_passphrase));
> Hi Chad,
>
> Would it make sense to have privkey() return the private key by value?
> Is it non-copyable?

I think the reason is because a PKCS#8 key can be RSA/DSA or whatever else.

I'd like to be able to hold keys by value but I'm not sure whether that 
is useful. I can't really see a need to copy keys around, but then again 
who knows?

A Boostified Botan should consider this problem as Botan requires the 
use of lots of yucky pointers in places because of the polymorphism.

-- 
Sohail Somani
http://uint32t.blogspot.com