Elrond <[EMAIL PROTECTED]> writes:
> On Wed, Mar 22, 2006 at 04:55:03PM +0100, Simon Josefsson wrote:
> [...]
>> int
>> shishi_key_from_name (Shishi * handle,
>> int32_t type,
>> const char *name,
>> const char *password, size_t passwordlen,
>> const char *parameter, Shishi_key ** outkey)
>>
>> Which even does this step correctly. Shisa now use this API.
>
> I checked shisa now.
>
> We might have run a bit in the wrong direction...
>
> When going the shishi_key_from_name() way in shisa, salt
> isn't set correctly and doesn't go into the newly created
> dbkey.
Oops, agreed.
> I don't know, what the right way out of this is.
> Maybe Shishi_key should know the salt, maybe
> shishi_key_from_name should output the salt. I don't know,
> really.
It seems better to separate this, I've created:
int
shishi_derive_default_salt (Shishi * handle,
const char *name,
char **salt)
Which shishi_key_from_name and shisa.c now use.
As a result, shisa.c doesn't use shishi_key_from_name any more, but
that API may be useful anyway, so I'll keep it.
What do you think?
Thanks,
Simon
_______________________________________________
Help-shishi mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/help-shishi
