[Freeciv-Dev] Re: loading rulesets patch

["Per I. Mathisen" <Per.Inge.Mathisen@xxxxxxxxxxx>]

On Sat, 2 Mar 2002, Mike Kaufman wrote:
> > -Alfred, Bertil and Leonardo join the game-
> > Alfred: Hey peeps, whatsup?
> > Johnny: Kewl, we have enough guys to play, let's rock.
> > Johnny: /start
> > -game starts-
> > Alfred: WTF is this??
> > Bertil: Ehhhhh...
> > Leonardo: no shit, this is a crappy game, let's go play Quake.
>
> Ok, sure, Johnny and Bobby are real assholes in this case as they didn't
> inform their competitors. (or give their competitors time to inform
> themselves)

Well look at it from Johnny's perspective:

players_iterate(pplayer) {
   patiently_explain_a_modpack_is_loaded_and_what_it_does(johnny,pplayer);
   if (game.nplayers > MAX_AMOUNT_PATIENCE) {
      tired_and_quitting(johnny);
   }
}

> Changing to a different ruleset is just a much bigger change
> (in most cases). So yes, people should be informed about it (perhaps
> even before they join into pregame (should one be able to 'finger' a
> port and get a 'show' listing back?, shounds slightly dangerous)) but it is
> not so fundamentally different that one should be _only_ be able to
> change it at the console (before the server even starts).

I would definitely want to know about it before joining pregame, as I
argued earlier. It is quite possibly a fundamentally different game, and I
think the restriction that you can only pick rulesetdir when you launch
the server to be quite managable.

> About the mallocs, etc not getting freed, I'm not sure why this is
> a problem. doing a /set tolkien /set default /set ancients in pregame is
> not going to load the rulsets three times. we're not going to reload, (and
> least it shouldn't at present (maybe when we can dynamically edit
> rulesets...), so this seems a strawman.

The entire point of this patch is to load rulesets _before_ pregame, so
that we can use ruleset information _in_ pregame. So if we are to be able
to change rulesets in pregame, we must reload. It is not a strawman, it is
a real problem.

> Conclusion: this feature has got to go back in. If you want to write a
> warning into stdinhand for "help rulesetdir" about possible nastiness,
> fine. fix the ruleset code later.

If you insist...

Yours,
Per

"What we anticipate seldom occurs: but what we least expect generally
happens." -- Benjamin Disraeli