[quagga-dev 9649] Re: [PATCH RFC] OSPF vertices memory exhaustion

Joakim Tjernlund joakim.tjernlund at transmode.se
Mon Aug 6 19:05:22 BST 2012


Paul Jakma <paul at jakma.org> wrote on 2012/08/06 19:24:26:
>
> On Mon, 6 Aug 2012, Joakim Tjernlund wrote:
>
> > But it is OK for some to do "code churn" for no/little gain?:
>
> > lib: prefix.c nano-optimisation
> >
> >    * lib/prefix.c: (prefix_match) nano-optimisation, let it return early
> >      without copying pointers.
>
> Well, it's subjective but that one is very self-contained and easy to
> understand. It was done as part of a contract I had in early 2010 to
> improve the performance of ospfd. Though not stated in that commit
> message, that tweak was made because profiling data showing prefix_match
> was hot, and was profiled itself. Though, I don't remember though if I
> profiled that patch on its own, or with other patches - and I didn't
> document that.
>
> Further, I posted that patch here, you commented on it and I addressed
> your comment by changing the patch. You didn't NACK the patch.

This was just an example, I would be surprised if you did get that, to illustrate
that such comments are very subjective and means just about nothing. I
end with " ... and I rather not go there, please?" to let you know that
I didn't want to continue this "debate" and you reply with a silly
defence which is invalid(why should contract matter, questionable profiling,
I never said I was against such changes, you did and me NACKing your patch would have
little bearing).

>
> I'm not claiming to be perfect. Many of the nits I raise are things I've
> learned from my own changes. ;) I need as much help with my blind-spots as
> anyone else. :)

Clearly, the above comment would have been enough.

 Jocke




More information about the Quagga-dev mailing list