To be clear when I say "pops off" I mean that where the comparator is supposed to be attached within its block space it is not attached and a floating comparator ITEM is left floating in its place [which will despawn after the regular time].
I spend a lot of time on my realm doing redstone stuff and since the 1.13 change I have not been able to use my WorldEdit mod so I have been using the structure blocks for copy, paste, and rotation a lot so I can combine large sections of redstone machines so that they work together and this is a constant persistent issue, especially when I have to replace comparators in tucked in spaces you just can't get to, where I have at times resorted to using chat commands to place them because it is actually faster than a partial breakdown and rebuild.
I had not noticed any real pattern to how it pops off comparators as it will pop off some while others are fine. It can properly place a whole row of them and pop off other specific ones or sometimes pop off a whole row and leave other specific ones. It is common as it happens all the time and it is only sometimes that it properly places all of them. Keep in mind that I typically have a large number of comparators in my redstone contraptions [like classic item sorters to even more complex cart type sorters], so this may be an indicator as to why it happens to me so much since I do so much redstone. It does not seem to do this with repeaters or other types of redstone [that I have noticed], it seems to be just the comparators. I am not sure how to get structure files from my realm since that is where all my work is and there is no file browser, otherwise I would have attached some of them.
I originally came here to see if the bug is being worked on as this has been happening for a while now and I expected to see a report but I am guessing not too many people are using the structure blocks, especially in this way since I just found this one: https://bugs.mojang.com/browse/MC-134637 which shows it as resolved but that is apparently on windows 10 which is apparently somehow different on that OS ['better together'?], and this problem is in reference to the java edition on windows 7 and I have the latest update 1.13.1 and this is still continually happening. After seeing this other bug report I cannot be certain if it only happens on slabs and never on solids as I had not noticed, but I can assure you in win 7 java edition 1.13.1 it is NOT RESOLVED [as of 2018-10-01].
Linked issues
Comments 3
Thank you for addressing this. I seem to remember the original poster stated they had windows 10, so I thought it was for win10, especially since it was set as resolved and I was on a different OS and it was clearly still happening. Again I am not so sure that it only happens on slabs as I did come across one since I originally posted [literally just 1 at one time] that popped off that was NOT on a slab, so I feel it puts into question the assertion of it happening ONLY on slabs. I tried to reproduce the same result from that one incident [repeating the load op in diff locations] and it did not repeat, so I cannot be sure of why it happened that one time or how it may or may not be related. This also should create concern that it may also have a relation to world position, orientation, and such where you may get different results [if even slightly] depending on position, rotation, chunk boundaries, or other position and orientation related effects [e.g. the order in which block updates happen].
I at least hope that this can help resolve this issue and help make the code that much more robust. Again, thank you for addressing this issue.
Ok, so after reading the new changes made to the other issue referenced, I had a thought and suspected something and went to go check. The original solid quartz block that the comparitor popped off of that one time looks like a solid quartz block, however when I use the 'pick block' op on it I actually get a quartz slab instead of a solid block despite its appearance of being a solid quartz block. I think it is safe to assume I built it originally as a slab and added another slab to make it a solid block at some point after. So it may still be related to the slab, but I feel this makes the appearance of a solid block misleading in any context if in fact it remembers it as a slab but then has to still treat it as a solid,... This sounds to me like a bug just waiting to happen. This ambiguity of a solid remembering it is a slab may come to bite in the end.
Duplicate of MC-134637 (which is in fact not for Windows 10 Edition, but for Java Edition)