| Time |
Nick |
Message |
| 00:57 |
|
Noisytoot joined #luanti-dev |
| 02:25 |
MTDiscord |
<herowl> I agree that banning all AI written code shouldn't be done, simply because it's unenforceable. |
| 02:27 |
MTDiscord |
<herowl> In all seriousness, I wouldn't really trust a new contributor more than I trust AI. All code has to be closely reviewed. |
| 02:29 |
MTDiscord |
<herowl> AI code may be shadowbanned, i.e. coredevs may care less about such PRs without disclosing it and deprioritize them without disclosing AI code as the reason, but that's about all we can do. |
| 02:29 |
MTDiscord |
<herowl> (by new Contributor above I meant pre-AI) |
| 02:34 |
MTDiscord |
<herowl> And it's not like those "legalize weed, because people will be smoking it anyway". |
| 02:34 |
MTDiscord |
<herowl> Weed can be weeded out. Can be found on people. Government can arrest people making it available and destroy their farms. Etc. |
| 02:35 |
MTDiscord |
<herowl> Some people will be afraid to possess an illegal item on them, because that can easily be proven and they can be arrested. |
| 02:36 |
MTDiscord |
<herowl> But here, we have no way to prove something is AI and people know it. |
| 02:37 |
MTDiscord |
<herowl> We have no way to limit people's usage of AI |
| 04:00 |
|
MTDiscord joined #luanti-dev |
| 04:17 |
MTDiscord |
<et086> if the project makes it clear that it does not want LLM code in it, people are going to submit less LLM code, it may not be perfect but it will be okay |
| 04:18 |
MTDiscord |
<et086> *not perfect as in a tiny amount of people are still going to submit code made by an LLM, and i think that is somewhat fine |
| 04:18 |
MTDiscord |
<et086> it is a lot better than to have strange hidden rules |
| 04:20 |
MTDiscord |
<et086> *i think it is a lot better to have that than hidden rules |
| 04:30 |
|
turtleman joined #luanti-dev |
| 14:34 |
Krock |
Is there any documentation on the mapgen order (emerge thread + MMVManip)? |
| 14:56 |
Krock |
sfan5: I think it would be a good first step to use the 2nd patch from https://github.com/luanti-org/luanti/issues/17063#issuecomment-4148744891 to at least ensure *some* order |
| 15:03 |
Krock |
or rather... there should be a general status indicator for mapblocks since not all will be generated |
| 15:11 |
Krock |
> we can confirm that the block is indeed empty (ignore) at activation time: |
| 15:11 |
Krock |
> block (-2,0,0) contained |
| 15:12 |
Krock |
get(it).name returns "ignore" for ignore nodes, thus the mapblock must consist of unknown nodes |
| 15:38 |
Krock |
ServerMap::emergeBlock() might actually create an air-only mapblock, causing side-effects in mapgen (it's the same "pool" of mapblocks) |
| 16:08 |
sfan5 |
Krock: I planned to write a comment in there, still |
| 16:08 |
sfan5 |
the patch is for testing only and would cause issues in real use |
| 16:09 |
sfan5 |
(imagine an object becoming deactivated in a non-generated block and the engine refusing to ever activate it again) |
| 16:38 |
Krock |
the object would have to wait until the mapblock is ready |
| 17:47 |
[MatrxMT] |
<Zughy> Somehow related? #16834 |
| 17:47 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/16834 -- Spawning an entity in a block that is loading might fail |
| 17:55 |
sfan5 |
I don't think so |
| 22:33 |
|
panwolfram joined #luanti-dev |
