The bug

For some plants when the block supporting it is instantly (e.g. using 0-ticks) replaced with the same block type the plant will grow.

The expected behavior is that the the plant should break, or that nothing special should happen.

As it currently stands this behavior allows for mass duplication of these items on the order of tens of thousands of items per hour. Far, far faster than any conventional farming method.

This bug may relate to MC-165023 where crop growth is randomly updated via falling sand on boats.

Affected plants

• Carrots, Potatoes, Wheat, and Beetroot
• Bamboo: https://www.youtube.com/watch?v=FROPCjl2DQM
• Cactus: https://www.youtube.com/watch?v=NkV9iXii9_4
• Chorus flower: https://www.youtube.com/watch?v=Ar0Zve-AzeU
• Kelp: https://www.youtube.com/watch?v=Ctc6anMVPW4
• Sugarcane: https://www.youtube.com/watch?v=qadaI38zRZc
• Twisting vines: https://www.youtube.com/watch?v=ojiaUGdQmDk
• Weeping vines: https://www.youtube.com/watch?v=VWbc8GNnkns

How to reproduce

1. Set the value for the gamerule randomTickSpeed to 0

   /gamerule randomTickSpeed 0
   

2. Build a contraption like the one shown in Chorus flower update setup (17w06a).png and activate it

Code analysis

Based on 1.13.1 decompiled using MCPConfig

The affected blocks check in their Block.updatePostPlacement(IBlockState, EnumFacing, IBlockState, IWorld, BlockPos, BlockPos) implementation (which is called when neighboring blocks change) whether they should break because the supporting block is missing. In this case a tick is scheduled. The problem is that when this tick is executed, the support block exists again and therefore instead of being broken, the plant grows (both actions are part of the update method Block#tick).