Re: [Monotone-devel] speed of "mtn ls branches"

[William Uther]

On 17/01/2008, at 8:20 PM, Nathaniel Smith wrote:

> On Thu, Jan 17, 2008 at 07:40:09PM +1100, William Uther wrote:
> > I'm not sure if that should be committed or not.  It reduces 'time ls
> > branches --ignore-suspend-certs' on my monotone db from 7s down to
> > less than 0.1s.  But it means ignoring the validity of certs in this
> > case.
>
> I am doubtful that anyone cares if 'ls branches' might include branch
> names that only exist in invalid certs -- at the moment, there isn't
> even any sane way to *have* invalid certs.  I don't care, myself, and
> am perfectly happy to leave it broken until a user actually notices.
>
> If one really does want to get 'ls branches' exactly correct in that
> case, the right way isn't to do expensive head-calculation anyway.
> All we need to know is that there is some valid branch cert out there
> somewhere, and we can tell that by just pulling a cert out of the db
> and checking it, which is very cheap.

Fair enough.  I'll commit the patch.  I just wanted people to be clear
on what was going on.

Will        :-}

P.S.  Yeah - I agree that we didn't need a full head calculation, but
it isn't as simple as a single cert check either... if the check fails
then you need to keep checking until you find no more of those certs,
or until you find one that doesn't fail.