Fix a bug in mega distill tower

[oier-namespace-std]

Problem: In game the incompleted mega distill tower is detected as completed.

Possible reason: structure check forgot to ensure that the top layer is completed.

PS: I do not have a Java debug environment. :<

[oier-namespace-std]

Bug analysis:

In file bartworks/common/tileentities/multis/mega/GT_TileEntity_MegaDistillTower.java, 293-294

        while (this.mHeight < 12 && this.checkPiece(STRUCTURE_PIECE_LAYER, 7, this.mHeight * 5, 0)
                && !this.mTopLayerFound) {

If the top layer structure check fails (the top layer has air), but checkPiece() found at least one machine casing (which will call onTopLayerFound() (line 181) and set this.mTopLayerFound to True), then we will get out of the loop and enter final check (305-307):

        return this.mCasing >= 75 * this.mHeight + 10 && this.mHeight >= 2
                && this.mTopLayerFound
                && this.mMaintenanceHatches.size() == 1;

And these checks will eventually pass, resulting in a weird 'completed structure' above.

So maybe rewritting the logic on line 293-294 (report incomplete structure instantly after checkPiece() fails) is better ig?

[AbdielKavash]

Thanks for the explanation! I have tested it in the full pack, and it fixes the problem.

Can I ask for one more thing? Explicitly check whether the top "layer" contains an output hatch. The regular DT will not form if any layer doesn't have an output hatch, but the MDT allows the top layer without an output.

[Glease]

this is not how it should be done.

1. HINT piece should never be used in checkMachine()
2. checking the same space twice for the convenience of coding is slightly too wasteful in a mega machine

this is really just a duplication of GTNewHorizons/GT5-Unofficial#865 and the correct fix is do what that PR was doing.
I'll prepare another PR in the mean time to do what I suggested.

[oier-namespace-std]

this is not how it should be done.

1. HINT piece should never be used in checkMachine()
2. checking the same space twice for the convenience of coding is slightly too wasteful in a mega machine

this is really just a duplication of GTNewHorizons/GT5-Unofficial#865 and the correct fix is do what that PR was doing. I'll prepare another PR in the mean time to do what I suggested.

After the bug analysis I realized that the piece check is actually done but the old code didn't deal with the returning results correctly. So I totally agree with changes in the next PR.

Thanks for review!