[Solved] 0.77.17 - Crucible Mixing Percentages

[Annother 0]

Hotfix #: 77.17Suggested Name: Super Annoying Crucible Percentage Anti-Alloy Bug 

Suggested Category: Severe/AnnoyingDescription: Recently I was working with my crucible on a server and got the percentages just right for blacksteel.  When I placed a ceramic mold into the bottom slot a small amount of blacksteel poured into it, then stopped.  As this happened the pecentages on all of the metals in the crucible went up (totaling over 100%).  I tried to get the percentages back to the correct values by adding metal into them, but i had to make black bronze to do so, and the same thing happened with a different crucible (one that I had placed after I initally found the bug).  I relogged to see if that had any effect on the percentages and now they all total 0%.

 

I have included some screenshots for your viewing pleasure. 

Have you deleted your config files and are still able to reproduce this bug?: No, though I am unsure of how this effects a serverside bug 

EDIT:  I deleted my config files and tried this again on singleplayer making bismuth bronze in a crucible.  The percentages of metals change as they are being poured into a mold from the crucible.  It seemed like the bismuth and zinc percentages were going down and the copper percentage went up (if that means anything).

 

 

 

 

 

 

EDIT 2- electriceditaloo: After a bit more use with the cruible on hotfix 17 ive noticed that the percentages do fluctuate a bit as the metal is being poured and, after a few ingots worth of bronze were poured I got this little glitch with the metal percentages.

 

[Tsuarok 47]

Just to chime in, I'm also experiencing this bug. As you drain the unknown alloy, the percentages of some components drop below 0 into negative values. It happens with every alloy I've made, including, black, blue, and red steels, rose gold, sterling silver, and brass.

[aliceingame 114]

reporting now

[Annother 0]

Fixed as of Hotfix 17, thank you for fixing your "derp math".

Edited 25 Sep 2013 by Annother

[JohnTzimisces 0]

I've just encountered this in hotfix 17. After adding bismuth to correct the ratios, it happened again. I then added a tetrahedrite nugget (having originally started with 10, just wanted to clear some inventory space) and I've gotten metal percentages that total over 100%.

First encounter with the bug:

After adding a copper nugget to correct the screwed up percentages:

[Annother 0]

ok i take it back, your math is still derpy

 

 

able to have over a hundred percent total?

[wyrmofvt 14]

It may be due to rounding errors. I don't know how Bioxx chose to represent the amount of each metal in the mixture, but if he used integers to represent the amount of each metal, then I can easily see how this could happen in some cases.

 

Let's invent a new unit, centi-ingots (one-hundreth of an ingot, unit* cIg) to represent the amount of metal present in a mixture, and we have some bismuth bronze (100 cIg) cooking in a crucible with a mixture of 60% Cu, 14% Bi, and 26% Zn. That means that there is 60 cIg of copper, 14 cIg of bismuth, and 26 cIg of zinc in the mixture. If you pour 30 cIg of bismuth bronze into a mold, then the amount of copper, bismuth and zinc you actually poured into the mold are not even numbers — it would be 18 cIg Cu, 4.2 cIg Bi and 7.8 cIg Zn. This would leave a mixture of 42 cIg Cu, 9.8 cIg Bi and 18.2 cIg Zn, but because we are storing these as integers, they get rounded to 42 cIg Cu, 10 cIg Bi and 18 cIg Zn. Note that amount of bismuth gets rounded up, so there's more bismuth in the mixture than before, so its percentage is higher.

 

Anyway, the math needed to keep this from acting screwy isn't hard. You just need to regard the mixture as a vector of all the components in the mixture, with its 'length' defined as the sum of the elements of the vector (taxicab norm). Adding to the mixture is a straightforward vector addition. Drawing from the mixture is the scaling of the length of this mixture by the amount of metal removed from the mixture. This should produce the correct behavior.

 

But this is just me mouthing off. I may be completely off base.

 

* SI review of this unit is pending at this time. 

[Kittychanley 1,002]

Exactly that. It's rounding errors.. I just couldn't think of a simple way to explain it like you did.

[dunkleosteus 2,425]

The values for the metals are stored as floats, but the method that returns the total amount of metal in the crucible returns an int, which would just cut off the digits after the decimal. The problem in the picture can only occur when there is a very small amount of metal in the crucible. If there is 1.8 units of 1 metal and 1.3 of the other, the total metal calculated would be 2, giving 90% metal A and 65% metal B.