Bryan Brown
- bbinme
- bbinme
- Europe/Stockholm
- Yes
- No
MC-30997
A simple repeater T-Flip Flop no longer works in 13w02a. I tested the t-flip flop (see attached image) in snapshot and
ittheeoutput was always off. I also tested in the 1.4.7 prerelease, and it worked, ie the output toggled between on and off properly.A simple repeater T-Flip Flop no longer works in 13w02a. I tested the t-flip flop (see attached image) in snapshot and the output was always off. I also tested it in the 1.4.7 prerelease, and it worked, ie the output toggled between on and off properly.
The LC parameter which shows
inwhen you have debug enabled,shows theindex of thetop most block in the section of the chunk that has anonair blockin it. This parameter is important in the calculations for entity spawning.When you remove a block that was the last non-air block in that section of the chunk. The LC param should be recalculated, it is not. It is recalculated everytime you place a block.
The LC parameter is only recalculated when you restart the world, on single player this is an annoyance, but not that bad. On multiplayer this is a really annoying because the server needs to be restarted. On LAN worlds I am assuming (didn't test) that the host would need to restart.
The value should really be recalculated everytime a block is removed. The calc is there when placing a block. Why not on remove ?
Also I have seen the LC parameter change inconsistently due to the player jumping, or moving back and forth between adjacent chunks. This behaviour is inconsistent as far as I can tell.
The LC parameter which shows when you have debug enabled, is an index of the upper y coordinate of the chunk section which has the highest non-air block for that chunk. This parameter is important in the calculations for entity spawning. When you place a block above that section or remove every non-air block out of that section, the section should be unloaded or loaded as applicable and the LC parameter should correctly reflect this change.
What happens when you remove a section:
When you remove a section of a chunk, e.g. by removing all non-air blocks. The section stays loaded and the server still uses these empty sections for calculations like mobspawning. This does not change if you reload the world! However, if you reload the world, the client's LC parameter will reflect that it has been recalculated (even though the section has not been unloaded on the server).What should happen:
When a section is completely empty it should not be loaded, the LC parameter should reflect that it is not loaded. And any calculations for mob spawning etc, should use the correct LC value. The client should always show the correct LC value.How to fix:
There is a method in the Chunk class (abw.class) which checks for the topmost filled section in a chunk (int h()). Don't just check that the section is not null, but also that it is not empty.This video details the problem and the fix
http://youtu.be/LIDNyB-Ys8wThis bug has existed for a long time, let's get it fixed !
LC parameter is not recalculated when you remove a block It only recalculates on restart of the worldIncorrect LC value
Incorrect LC valueLC value displayed in client may not reflect the LC value calculated by server
Pre 1.6.3 it was possible to have a Desert or Jungle Temple that was in the proper biome, and also partway in a swamp biome.
These structures were previously valid locations for witches to spawn. The code didn't distinguish between a temple, or a witch hut.
In 1.6.3 the Desert and Jungle Temples that are partway in swamp biome no longer spawn witches.
Is this a correction of a defect (witches should not have been able to spawn in temples) ? Or a new bug, that the code (in 1.6.3) that saves the structure data, is not correctly handling temples that span a swamp biome ?
Attached screen shots in 1.6.3, and then in 1.6.2
I noticed in the latest snapshot that when a dispenser with bonemeal in it is triggered the bonemeal is consumed even if their is not a crop/tree/flower to grow in front of the dispenser.
This is a change from the behavior that has been in place for quite some time.What did I do:
I had a dispenser with a bonemeal in it, and the dispenser is pointing at block, which has a fully grown wheat plantinit. I click a button to activate the dispenser.What happened:
The bonemeal was consumed, even though it was not used to grow the wheat (it was already fully grown). The green particle effect is displayed.What I expected to happen:
The dispenser should have 'clicked' but since the bonemeal was not actually used to grow a crop, the bonemeal should not have been consumed.I noticed in the latest snapshot that when a dispenser with bonemeal in it is triggered, the bonemeal is consumed even if their is not a crop/tree/flower to grow in front of the dispenser.
This is a change from the behavior that has been in place for quite some time.What did I do:
I had a dispenser with a bonemeal in it, and the dispenser is pointing at a block, which has a fully grown wheat plant on top of it. I click a button to activate the dispenser.What happened:
The bonemeal was consumed, even though it was not used to grow the wheat (it was already fully grown). The green particle effect is displayed.What I expected to happen:
The dispenser should have 'clicked' but since the bonemeal was not actually used to grow a crop, the bonemeal should not have been consumed.
Works correctly in 1.4.7 (full release). Fails on snapshot 13w01a.
I can confirm the the LC parameter is inconsistent. I logged in to my SMP server today and the LC parameter of a chunk that used to be 47 is not back to 79 again. Even though there has been nothing built in that chunk at all. Very weird behavior.
I can also confirm this issue. I had 16 pens of sheep. I wouldn't breed the pens right next to each other at the same time (to keep the sheep from breeding with different color sheep), when I breeded two collections of sheep they could be pretty far apart, the next sheep born would occasionally be the wrong color. I've attached a screenshot of the result after it happened a couple of times.
Well in my defense I logged this bug a long long long time ago, and my observations were based on playing the game. Now that I see your explanation and video and look at the code, I have corrected the "errors" in my bug description.
The description asks a legitimate question, did 1.6.3 fix a bug in 1.6.2. Or is there a bug in 1.6.3
@Th3F4114n0n3 witches have always been able to spawn in any structure in a swamp biome. It was not a bug in 1.6.2. If it is a bug it goes right back to the scary Halloween update when witches were introduced.
Previously to 1.6.2 a witch is a valid entity to spawn on a block which exists in a structure, and in a swamp biome, providing it meets the other spawning criteria (bounding box, light level etc). So if a portion of a desert or jungle temple (which share the same parent class which contains the code) is in a swamp biome, then it is possible for a witch to spawn there.
Was it a bug to spawn a witch in a temple ? Possibly. Why not let Mojang answer that.
The situation has not changed in 1.6.4
Attached crash report
Resolved in 14w34d