Re: [gnutls-dev] Work in progress: GnuTLS 2.2 release notes on API changes

[Simon Josefsson]

Nikos Mavrogiannopoulos <address@hidden> writes:

> On Thursday 15 November 2007, Simon Josefsson wrote:
>
>> I think it would be better if we had a function like:
>>
>>   int gnutls_set_priority (gnutls_session_t session,
>>                            const char *priority);
>
> I just remembered that there was a reason this priority function was kept 
> simple from the begging (integers only). This function is called per session, 
> thus having a parsing routing like this would add some overhead... This could 
> be insignificant compared to RSA/DH etc, but still in a busy server it might 
> become significant.

Ah, I understand.

> What I had thought then was to make this parsing routine output the result
> in a gnutls_priority_st structure and then associate this struction with 
> every 
> session. If found that solution complex then...

How about implementing the simple gnutls_set_priority function now, and
if it turns out that it is actually a performance bottle-neck for some
applications, we can add a gnutls_parse_priority and a new
gnutls_set_preparsed_priority function to handle that.  I think for 90 %
of the applications, the inefficiency doesn't matter.  Premature
optimization is the root of all evil etc...

/Simon