In the tick when a marker Armorstand is summoned it seems to have the Hitbox of a normal Armorstand. One tick later the hitbox is corrected.
This can be observed with two repeat commandblocks on top of each other:
lower block:
/summon armor_stand ~ ~ ~ {NoGravity:1b,Marker:1b}
upper block:
/execute @e[dy=0] ~ ~ ~ /say hi
If you power the upper commandblock first and then the second, everything is as expected and nothing is printed in chat, because the upper command is triggered first in the tick.
If you first power the lower block and then the upper, the chat will be spammed with hi as the Armorstand is summoned and immediately tested for. The correct behaviour would be, that the Armorstand is still not found.
You can also change the upper repeating commandblock to a chain commandblock, wich is always active to observe the bug.
A useful command to get rid of all the Armorstands
/kill @e[type=armor_stand,r=5]
Here is a OneCommand that creates the setup
summon falling_block ~ ~1 ~ {Block:redstone_block,Time:1,Passengers:[{id:falling_block,Block:activator_rail,Time:1},{id:commandblock_minecart,Command:fill ~1 ~-2 ~1 ~3 ~-1 ~1 air},{id:commandblock_minecart,Command:setblock ~2 ~-2 ~1 repeating_command_block 1 replace {Command:summon armor_stand ~ ~ ~ {NoGravity:1b,Marker:1b}}},{id:commandblock_minecart,Command:setblock ~2 ~-1 ~1 repeating_command_block 1 replace {Command:execute @e[dy=0] ~ ~ ~ /say hi}},{id:commandblock_minecart,Command:setblock ~ ~2 ~ command_block 0 replace {Command:fill ~ ~-4 ~ ~ ~ ~ air}},{id:commandblock_minecart,Command:setblock ~ ~1 ~ redstone_block},{id:commandblock_minecart,Command:kill @e[type=commandblock_minecart,r=0]}]}
Linked issues
Attachments
Comments 2
This was fixed in 16w44a and is still fixed in 1.11. It is difficult to pinpoint the exact fix version, because 16w43a behaves super weird and MC-88181 can not reliably be reproduced.
Same issue as MC-88181.