---- James Cameron <quozl at us.netrek.org> wrote: 
> I've tried this, and it hasn't worked.  Which is surprising, given that
> the code to point the ship at destination coordinates worked fine.  In
> my latest patch I've commented it out until I understand it.

That's strange.  I'll have a look too.  Two possibilities that spring to mind are that, in order of liklihood:

a) I don't understand p_dir, and I'm not incrementing it enough to escape the nint function (although + 1 seems to be enough, is it rounded somewhere else to the nearest five or whatever?)

b) the point-the-ship-at-the-destination code doesn't work either, it just appears to, because some other code is actually updating ship direction, and happens to be  correctly pointing the ship as well.

> I've also added a third layout which I've made the default.  

Sweet, looks nice.

> The current code in my repository has been through several runs of
> tests/inl-draft ... it is rather fun selecting robots, and when the
> draft ends they all take it out on the captains.  ;-)

I'm going to have to go through a few runs tonight, sounds like fun :P

> I've not yet finished thinking about how to handle observers in the same
> draft sequence.

I've been thinking about this as well.  Is it possible to generate robot ships, assign them an observers number, and stick them in the pool?  Then we assign player slots (using setship) according to pick_sequence, randomly, or whatever scheme we devise.

The most important thing is to not break from graphical selection, which is great.  So they need some kind of placeholder on screen.