Re: [gnugo-devel] GNU Go 3.4 or 3.5 in Debian sarge?

[Arend Bayer]

Gunnar wrote:
> Dan wrote:
> > I suppose we should tune, then release. If we could get
> > persistent caching for the semeai code we could increase
> > the semeai nodes which might make the engine stronger.
> 
> I think we have to live with 3.6 not being significantly stronger than
> 3.4. Not that a little tuning would hurt, of course.

I think waiting for tuning only makes sense if someone starts working on
that very soon. Otherwise I think that due to slow development lately we
already have the usual stabilization period behind us. Providing Debian
sarge with an officially stable version seems like another good excuse to
do it now.

I implemented persistent semeai caching tonight, will post a patch
tomorrow. Other than that, I would suggest to integrate the outstanding
patches, release 3.5.10, ask as many gnugo bots as possible to upgrade
and report crashes if they happen, and then release this as 3.6 soonish,
i.e. around 15-20th of August?

Arend