So induktio mentioned to me that they noticed a bug where the faction defined societal priority / opposition effects are completely disabled. While working on the read_faction() function I've come across where the actual issue is in the code. I vaguely remember seeing posts or reading about this issue awhile ago. I don't recall who or where since it's all a jumble in my brain at this point. Maybe kyrub or Yitzi addressed this already via patching?
Just to outline the issue, inside Player structure there are 6 values. The read_faction function initializes both the categories and effects as -1 and model to 0. Inside the faction editor, category and model priority are determined by social ideology. Opposition is determined by anti-ideology (not sure where I came up with term opposition, I may rename code variable names). The bugged value is referenced as emphasis or socPriorityEffect below. It looks like you can't directly set anti-emphasis via faction editor nor is it directly referenced at all in the code. I haven't dug into indirect references other than read_faction().
An example from gaians.txt:
Economics, Green, PLANET
Economics, Free Market, nil
struct Player {
...
int socPriorityCategory; // Gaians will be Economics
int socOppositionCategory; // " "
int socPriorityModel; // Gaians will be Green
int socOppositionModel; // Gaians will be Free Market
int socPriorityEffect; // Gaians should be PLANET, vanilla code sets to -1 or disabled
int socOppositionEffect; // unused/padding/ignored
};
void __cdecl read_faction(Player *player, int type) {
...
for (int i = 0; i < 2; i++) { // first priority, then opposition
*(&player->socPriorityCategory + i) = -1;
*(&player->socPriorityModel + i) = 0;
*(&player->socPriorityEffect + i) = -1;
<parse and set societal categories; all correct>
<parse and set societal models; all correct>
LPSTR socEffect = text_item();
for (int j = 0; j < MaxSocialEffectNum; j++) {
if (!_strcmpi(SocialEffect[j].set1, socEffect)) {
*(&player->socPriorityEffect + i) = -1;
break;
}
}
}
So the issue is, rather than setting it to j it sets it to -1. I couldn't find anything stating this mechanic was suppose to be disabled versus something like the freefac value under #FACILITIES. It is actually referenced in two functions: communicate() and social_ai(). After skimming direct references, communicate() doesn't seem that interesting but there is a large section of code that gets skipped in social_ai() if socPriorityEffect is -1. So perhaps fixing this will improve the AI in some ways or at least make them act more as they were intended.
Comments, thoughts?
New SMAC quizzes available.
