Re: Improving RLS qual pushdown

Dean,

* Dean Rasheed (dean(dot)a(dot)rasheed(at)gmail(dot)com) wrote:
> Thanks for looking at this.

Thanks for working on it. :)

> On 28 February 2015 at 04:25, Stephen Frost <sfrost(at)snowman(dot)net> wrote:
> > * Dean Rasheed (dean(dot)a(dot)rasheed(at)gmail(dot)com) wrote:
> >> --- 1318,1347 ----
> >> }
> >>
> >> /*****************************************************************************
> >> ! * Check clauses for non-leakproof functions that might leak Vars
> >> *****************************************************************************/
> >
> > Might leak data in Vars (or somethhing along those lines...)
> >
>
> Perhaps just "Check clauses for Vars passed to non-leakproof
> functions". The more detailed explanation below is probably then
> sufficient to cover why that's significant.

Agreed, that looks fine to me.

> That sounds OK (except s/sensetive/sensitive).

I'm horrible with spelling. :) Thanks for catching it.

> I'm not sure that's really much clearer, because of the 2
> double-negatives in the final if-clause, and it's less efficient in
> the case where the input function is leaky, when it's not nessecary to
> check the output function (which involves 2 cache lookups).

Yeah, it is a bit less performant.

> Shrug. Not sure it makes much difference either way.

Alright.

> > The rest looked good to me.
> >
>
> Thanks. Do you want me to post an update, or are you going to hack on it?

Either works for me, though I'm happy to handle the modifications to
this if it means you have time to look at the other patches.. :)

Thanks!

Stephen