Villagers not breeding, recognizing houses or farming

Villagers still don't breed in 18w02a.

Villagers still don't breed in 18w03b. Wonder why this isn't assigned to anyone? Sort of a serious bug for anyone who wants to make a base full of villagers, or someone who just wants to repopulate a zombified village.

I have confirmed from my own experiments that villagers do not detect doors.

In villages.dat in the nether, end, and overworld, not a single village is created. Villagers do not track towards doors, do not breed, and do not produce iron golems.

Possibly related to MC-122308?

Testing with 18w03b, villages.dat is being called to save village data when doing save-all, but village door data is non-existent, or a single incorrect entry is included for a village with multiple doors. So, houses are either not being added, perhaps due to how door direction is determined, or possibly something is removing houses from the list. Either way, should be pretty straightforward to debug why villages.dat gets no door entries if someone looks into this issue.

With the changes in the codebase I would assume that the variable/ID for doors has changed and that the villagers have not been redirected to the new var.

Just tested 18W05A, villagers still don't breed. Crafting erotic magazines and handing them to villagers also doesn't help them to breed. This new version is still broke.

As of this morning 18w05a is still not saving info in the data/villages.dat file. Consequently villages are not recognising doors and not breeding. I'm surprised this bug is still Unassigned.

Tried it again in 18w07c, breeding still doesn't work. Also, /locate says that villages are at y=0? Still a bug...

In 1.12.2 they sometimes breed but with some door configurations that should work according to the minecraft wiki, they don't work.

I was able to open Minecraft 18w08b with the ASM Village Marker mod, which draws the village radius and detected doors to the screen. In a void world, I made a line of doors and spawned a ton of villagers. The villagers are detecting the doors, BUT the villages instantly despawn. See this video to see what I'm talking about: https://www.youtube.com/watch?v=OAsust8I-m0
(Confirmed for 18w08b)

Tried getting villagers to breed in 18w09a, villagers still act like no doors are around and they're still not interested in breeding. I imagine Mojang is probably knuckling down on the 'look and feel' part of 1.13 now and they'll hit the 'debug' portion of the release later, but hopefully this village/villager bug gets fixed before 1.13 is released.

Agreed, cannot stand villagers not breeding.

I can confirm this for 10w08d, using the ASM village marker mod (https://www.youtube.com/watch?v=p8IcoPeeIOg).

It's worth pointing out that when the villages are created (then immediately destroyed) their centres are pretty far away from where the doors are. It's possible that for some reason the villagers detect the doors, save the wrong coordinates, then the game checks if there is a door there and fails, deleting the village.

OK, I've just done a bit of testing. I made a strip of 12 "house" doors and found where the village's centre was. I made a platform of doors in that area, and to my surprise the village stayed after "detecting" a few doors in that area, but every few seconds its centre would tick to the left or right. These doors also wouldn't've counted as houses as there were no roof blocks. When I put down a bit of farmland and seeds and two villagers, they entered love mode and began farming!

And to add to all of this, after looking at the NBT of the villages.dat file, there were 12 door entries (which is correct) BUT they all had the same X/Y/Z coords. When I broke all the doors on the platform EXCEPT for the door at that position, the village still broke, and the centre moved when I placed all the doors back. I minimized the platform to the smallest I could with the village intact, but after a few seconds (presumably an autosave) the village broke again, then later reappeared. The platform with the doors is 16 blocks east, 16 blocks north, and 3 blocks up from the original doors with 5 strips doors on either side of the centre strip (after some adjustments to my experiment).

It's possible that for some reason the villagers detect the doors, save the wrong coordinates, then the game checks if there is a door there and fails, deleting the village.

That's exactly what it's doing. The loop being used to check for a door within an area around the villager uses a mutable BlockPos instead of immutable, so while the door's position is detected correctly, the final coords will be in +16X/Z and +4Y from the villager's location (since that's where the loop ends). See the description of the report for this info; the fix is to simply call toImmutable() when storing the BlockPos.

OK, interesting. This seems like a pretty simple fix.

Still an issue in 18w14a.

Seems like it might be fixed in 18w15a. I've seen villagers take shelter in the houses and also seen them breed in a simple pen surrounded by 'houses'. I haven't tested all the mechanics in detail though.

Fixed in 18w15a. I loaded my village testing world in 18w15a with the ASM Village Marker Mod (it has a mod builder to work with any version) and the village doors are recognized in the right places.

The doors are also saved properly after closing; I checked the villages.dat file.

I noted 18w15a is also affected but later it is removed.

Can someone confirms if the villagers can breed in 18w15a?

The previous testing showed that it was saved properly (see the above comment); 18w15a was added because MC-128611 was reported and it's currently being checked whether this is actually still an issue or not.

This issue as reported, has been confirmed as being resolved for 18w19b.

Summary:

1. Villagers are confirmed as breeding with a typical infinite breeding test using 4 doors placed 5 blocks below satisfied villagers.

2. A farming radius test around a single door made the expected pattern of crops 31 blocks to each side of the center block, however an anomaly was observed in the south direction possibly due to villager AI behavior (see image).

3. A village merge test between two villages placed with 63 blocks in between was successful, the first village farms were created 31 blocks to either side, after the second village only the blocks at or between the doors were planted and farmed. The farms from the opposite side of the first door was no longer harvested as expected. (see image)

4. A villager iron golem test was successful with 10 villagers and 21+ doors, in the 16x16 area, and retested with 20 villager and 21+ doors which spawned 2 golems over time but stopped there as expected.

For the intent of the bug reported, this issue seems to be confirmed as resolved, however, this new anomaly observation of the farming ai in the south direction needs further investigation.