chore: refactor nullchecks

[Dorus]

This change needs some investigation re #109 (review)

The old code did effectively this:

Recipe[] allProduction;
if (goods == null)
{
  allProduction = [];
}
else
{
  if (variants == null)
  {
    allProduction = goods.production;
  }
  else
  {
    allProduction = variants.SelectMany(x => x.production).Distinct().ToArray();
  }
}

The new code is doing this:

Recipe[] allProduction;
if (variants != null)
{
    allProduction = variants.SelectMany(x => x.production).Distinct().ToArray();
}
else if (goods != null)
{
   allProduction = goods.production;
}
else
{
  allProduction = [];
}

goods	variants	old	new
n	n	[]	[]
n	y	[]	variants
y	n	goods	goods
y	y	variants	variants

Need to investigate what it means for goods or variants to be null. I've been running some scenario's in YaFC but couldn't find one where goods was null and variants wasn't. Some help would be appreciated.

[Dorus]

What i did find so far:

The only place where variants is not null is when we're called from ProductionTableView.cs:424:

var ingredient = recipe.recipe.ingredients[i];
var link = recipe.links.ingredients[i];
var goods = recipe.links.ingredientGoods[i];
grid.Next();
view.BuildGoodsIcon(gui, goods, link, (float)(ingredient.amount * recipe.recipesPerSecond), ProductDropdownType.Ingredient, recipe, recipe.linkRoot, ingredient.variants);

So goods = recipe.links.ingredientGoods[i] and variants = recipe.recipe.ingredients[i].variants.
recipe is in both cases a RecipeRow (ProductionTableContent.cs:173)

So goods will end up as Goods and variants will end up as Goods[]. After goods.production and variants.SelectMany(...) both are Recipe[].

[Dorus]

One thing i noticed while investigating this:

Before when you disable a recipe, the item row is emptied. Old projects still have this state stored:

If i enable them now and disable again, i get the recipes visible:

I'm not sure if this was intended, i tried to go to older versions along the master branch, but couldn't find any yet without this problem until i ran into the ones with the libf problem.