I had the problem, that when I wanted to deconstruct something that is bigger than the range of the construction area of a constructor, the deconstruction bugged out and didn't continue. I think the problem is that when an entity is at the border of the construction area, but just out of reach so that robots won't get it, the constructor will not move as it still thinks there is something in its range.
I took some pictures of the problem: https://imgur.com/a/cSPgCyS
The first picture shows the remainder of a deconstruction job. The constructor picked up everything in its range but there is some still remaining north of it. But he stood there for several minutes without moving. When I cancelled the deconstruction of the two rail segments that reach into its range, he starts to move upwards and deconstructs the rest as seen in the second picture. I had similar problems quite often and at first I thought the range was just too small, as I was still using mk1 personal roboports. But upgrading to mk2 didn't solve the problem.
I changed line 795 in \script\constructron.lua to this: (added the substraction)
area = chunk_util.get_area_from_position(constructron.position, constructron.logistic_network.cells[1].construction_radius - 4)
which seems to solve the problem for me. At first I put "-1" but still had some issues with bigger entities reaching into the range and generating the same bug so I just increased it to 4 and haven't had problems since. I am not really familiar with programming Factorio mods nor LUA so I don't know if this will cause some other issues as I also didn't look too far into the code, but rather just tried around.
Moreover I had the problem, that constructors left before all bots returned back into the inventory of the constructor. Which left them flying slowly over the map until they caught the constructor somewhere again. I couldn't find a solution to this, it is not really a critical bug but rather annoying and sad to see momma spidertron leave her little kids behind D: Would it be possible to increase the wait time or something like that to counteract this problem?
