Kenneth R. Robinette wrote:
> Or put the real world way, the only way we desire one to have a cert within
> NSS is to get that cert from one of our vendors that pay us a lot of money.
> 
> Ken

What does that mean?

I think you are confusing NSS with the builtin token.  NSS is a set of 
libraries that manage multiple tokens, including the softoken and 
builtin token.  The softoken is a writeable token with databases for 
storing certs and keys.  The builtin token is a read-only list of 
default trusted roots.  If you want "a cert within NSS", naturally, you 
would import it into the softoken, since it is writeable.

If you want a root that is trusted by default for any application that 
uses NSS, that is a much larger statement.

-Ian


> 
> "Ian McGreer" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED].;
> 
>>Kenneth R. Robinette wrote:
>>
>>>Yea, but that function also calls CERT_DecodeDERCertificate, along with
>>>
> a
> 
>>>whole bunch of other functions.  I thought the problem was with
>>>CERT_DecodeDERCertificate, as was pointed out in a previous reply as
>>>follows:
>>>
>>
>>It is perfectly okay for an exported NSS function to call another NSS
>>function internally.  In fact, that is exactly the behavior we desire.
>>
>>The reason we don't want applications to call CERT_DecodeDERCertificate
>>is that we want to reach a point where having a CERTCertificate implies
>>you actually have a cert in the NSS system, either on a token or in the
>>temporary cert store.  CERT_DecodeDERCertificate returns a
>>CERTCertificate with no state information, just a decoding.  If one was
>>to pass this CERTCertificate to a function needing state information,
>>bad things would happen.
>>
>>PK11_ImportDERCertForKey uses CERT_DecodeDERCertificate to obtain the
>>decoding, then imports the cert and fills in all the necessary state
>>information (token, handle, etc.).  It then returns a valid
>>CERTCertificate, and you're on your way.
>>
>>-Ian
>>
>>
> 
> 



Reply via email to