[Lou][2.0.33] Bots fail to build land mines on Gleba

[MEOWMI]

Construction bots will occasionally fail to build land mines (on Gleba in this case).

They seem very much stuck permanently, until manually canceled (and can seemingly cause further construction orders to be stuck as a result too). In my experience, this usually happens in locations close to edges or with tight space (I've seen it frequently fail around the Watercane doodad frequently and also in the green swamps).

Edit: To add, this issue happens maybe 1/100 mines, placed by hand or blueprint.

Log and test save are attached.

[Lou]

in most stuck mine ghosts the issue seems to be missing landfill(ghosts). I discovered that in some cases landfill ghost can be removed without removing the landmine ghosts, which might be the cause of this.

Do you remember what method you used for building the mines (manual physical, remote, blueprint?), and were you using tile deconstruction planner or undo/redo? Also, where you using just forced mode (Shift) or super forced mode (Ctrl + Shift)?

[MEOWMI]

in most stuck mine ghosts the issue seems to be missing landfill(ghosts). I discovered that in some cases landfill ghost can be removed without removing the landmine ghosts, which might be the cause of this.

Do you remember what method you used for building the mines (manual physical, remote, blueprint?), and were you using tile deconstruction planner or undo/redo? Also, where you using just forced mode (Shift) or super forced mode (Ctrl + Shift)?

In the past, on a few occasions I definitely did remove landfill ghosts (using a deconstruction blueprint for landfill only), but I do not think that was true in all cases.

I had difficulties replicating this issue today and wasn't actually able to replicate it with mines. (Nevertheless, in the past month or two it occurred over multiple sessions, dozens of times.)

However, my testing today has some other results. Regularly, something (usually a roboport) would fail to build, because something didn't get deconstructed (typical culprit being Watercane.) This might even maybe a different issue technically, but I'll add a continuation of the save file here demonstrating it, seeing as it's a followup. Let me know if you want me to make it a separate report or if it's a known issue.

I was especially careful with the reproduction steps for this issue today:

• Be in remote view
• Using shift+click (normal force build), paste blueprints.
• Observe that some of them never get built, because some things did not get deconstructed.

Remote view can be pure map view, zoomed out to where no sprites are drawn, or remote view zoomed in close enough for the sprites to show up. Being in editor mode was not tested.

For blueprints, a commonly used entity was roboport but can happen with other things like mines too. The blueprints can be for multiple entities but also just for single entities too. Super force build was not tested.

Seemingly, common causes of failure are Watercane blocking landfill in turn blocking roboport ghosts, but there were also cases of cliffs or trees doing the blocking. I also noticed more unusual things like landfill ghosts not getting built because stromatolite did not get ordered to be deconstructed (with no roboport ghost ever involved because it overlapped with other existing player entities thus causing force build to only add some empty landfill).

Attached is a save demonstrating this, with pins on the map to the south to help locating the cases. It is a continuation of the earlier save uploaded.

So it seems like there might be at least 2 slightly different things going on. One being that removing landfill ghosts sometimes doesn't cancel relevant build ghosts, one other being that certain doodads don't always get ordered to be deconstructed. And If I can replicate it at any point with mines, a possible 3rd issue with mines near the edges of unbuildable tiles unrelated to landfill.

[Lou]

Thank you for the report and for the cooperation do far.

The issue of removal of tile ghosts not removing marginal supported entity ghost has been just fixed for 2.0.34.

There used to be a bug regarding autofilled tiles and plants but

• it has been fixed in 2.0.31
• I did not find any suspiscous plants near the stuck ghosts in the first save

There is a known issue with water cane and placing tile ghost either manually or explicitely in a blueprint does not mark the cane for deconstruction.

I tried to reproduce your issue based on your steps, however ran into a bit different behaviour than you descirbed:

• Happened only when blueprint had (landfill) tiles explicitly included (not autofilled)
• Normal build (not force build)
• regardless of remote view

which is consistent with the known issue, but not with what you described. However, other than the differences in reproduction, from what I've seen in the second save, I still suspect it is the same issue.

If you are sure with your specific steps, could you maybe send a save with the blueprint and place where you observed the behaviour?
(If you are willing to look into it further, I would recommend placing some water canes on the margins of tiles in editor)

[Lou]

The water cane issue was fixed for 2.0.34. Marking this as resolved for next release. If you will still run into it after 2.0.34, we can reopen this.

[MEOWMI]

Thanks for fixing!

If you are sure with your specific steps, could you maybe send a save with the blueprint and place where you observed the behaviour?
(If you are willing to look into it further, I would recommend placing some water canes on the margins of tiles in editor)

I simply copy pasted already placed structures. I didn't change them in blueprint editor at all, and they seem to generate without any explicit landfill. I'll attach the blueprints for completeness, but realistically they shouldn't contain any surprises, they are just saved versions of ctrl+C copies.
Roboports:

0eNq1lE2OwyAMhe/iNYtAAvm5ymhUJa1bIRFIgYwmirh7SbLoYrog0rDEwu99era8wqBmnKzUHroV5NVoB93XCk4+dK+2mu5HhA6sGcxkrAcCz7lX0i+xqPCB+tbbBQIBqW/4Cx0N3wRQe+klHlr7Y7noeRzQxg/kg+ZkXGwwenOMIqxkBKIDrcNnPwIWnzM6f7lL5dG6rdHhddM4XN84IZA/ECwFoi7zQpQJELSpDogmE0SVlATdIXguCH4iCUEzQYiUJJjIC1GnQIj62ImCZ6JozmxmITJRtCcG8l8U8XhJj2NUeF9GAj+xbffkgrVV23LOm6ooyhBemLfMWg==

Copy blueprint
Mines:

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

Copy blueprint

I did some final testing with the roboports, and I did notice one new thing: If I paste things near non-generated chunks (particularly, trying to make sure that blueprints paste at least partly *outside* of generated chunks), then the issues in my last report apply. On the contrary, if I specifically avoid pasting in non-generated chunks, though, then things work just fine. This seems to cause the issues observed in the last post with remarkable regularity (dozens of occurrences in ~1000 pasted entities outside generated chunks, and with 0 occurrences among 1000+ when pasted within already generated chunks).

Aside from this new "out of bounds" thing, honestly what you have uncovered so far sounds close enough. I can just open an issue later on if issues persist after all the fixes, that's probably the most productive for everyone.