[crossfire] Alchemy formulae file broken

[Preston Crow]

I was doing some testing, and I found several basic alchemy formulas 
never showed up in books (mercury, philosophical oil, and philosophical 
salt).  After a lot of digging, I found that this accounted for a chance 
of 30, and there are 30 chance-1 formulas near the end of the file that 
aren't getting added to the total correctly.  This is because in 
recipe.c, it adds the chance to the total_chance when it reads the 
ingredients, but on those 30 formulas, the chance is listed in the file 
after the ingredients.

While we could say the bug is in the formula file for putting the 
entries in the wrong order, it would seem like the right fix is to 
change recipe.c to not care about the order.  I think trying to set the 
total_chance field as we parse the file is wrong, and we should scan the 
lists at the end to set the value.