Re: variants for all games

[Hans Ronne <hronne at pp dot sbbs dot se>]

>> However, Stan didn't like the idea, so he removed the code about a
>> year later. He thought we should fix the real problem (the game files)
>> instead of using this ugly hack. We never got around to doing that,
>> though.
>
>Well, I don't see any record of this in the mailing list archives so I
>don't know what his reasoning was.  I don't think we need to bring
>back the add_more_variants code; I think it is as simple as changing
>"int force_all_variants = FALSE;" to TRUE in read.c.  If the .g file
>mentions the variant anyway, the code seems like it can deal with it.
>
>It might be better if the .g file can set force_all_variants (with the
>default to TRUE, probably), but I don't really see what the problem
>with having it always TRUE would be.

Yes, I forgot about force_all_variants. This is a variable that Stan added
when he removed add_more_variants. I tried setting it to TRUE. It works in
the standard game, but crashes several other games. Not sure why, but I
suspect that some of the variants that are forced by it are not well
defined in all games.

The solution is probably to find the minimal set of variants that can (and
should) be enabled in all games, and change the code in read.c accordingly.
I would suggest that world-seen, see-all, sequential and a new variant
ais-may-resign (which would set g_ai_may_resign) should be enabled
everywhere. In contrast, people, supply and economy should be enabled only
in games that support them. And world-size only in games without fixed maps
(base modules). Not sure about real-time. It's currently enabled only in
the panzer game.

Hans

Hans Ronne

hronne@pp.sbbs.se