| Time |
Nick |
Message |
| 01:33 |
MTDiscord |
<exe_virus> maybe 2 bugs, or just one bug here with this mapgen multithreaded thing. Looks like the same mapchunk is called for mapgen twice - maybe we're literally double generating sometimes |
| 01:39 |
MTDiscord |
<exe_virus> oh and if you breakpoint while the first of the group and second of the group are generating, our thread pool get's another 2 lined up to tackle it. So literally 4 of the same mapgen chunk generation occuring right now.... So need to figure out why... even if intentional.... seems unintentional |
| 01:43 |
MTDiscord |
<exe_virus> (oh and for the record, making biome generation go +- y direction did have a large effect on fixing the issue, but doesn't resolve the core problem - instead it nearly masks it completely, but not quite. |
| 01:59 |
MTDiscord |
<exe_virus> Darn it, my brain is going to mush trying to understand our thread pool and emerge logic for he first time. I'll just keep focusing on masking the "issue". Getting biome generation to go +-1 Y did get most of the issue. The remaining one(s) are less repeatable with that fix in though so it'll take longer to hunt/mask it |
| 02:14 |
|
aliasalreadytake joined #luanti-dev |
| 02:41 |
MTDiscord |
<exe_virus> hurm, this is all due to overgenerated terrain. what if I just set that back to contentIgnore..... |
| 02:52 |
MTDiscord |
<exe_virus> oooh, that might be the solution!! just set that overgenerated +1 -1 Y layers to CONTENT_IGNORE again. works for mapgenV7, let's try some others |
| 03:10 |
|
SpaceManiac joined #luanti-dev |
| 03:10 |
MTDiscord |
<nathan4220776> Excellent. |
| 03:20 |
|
SFENCE_arch joined #luanti-dev |
| 03:22 |
MTDiscord |
<exe_virus> Woohoo, okay branch and PR are up: https://github.com/luanti-org/luanti/pull/16224 |
| 03:30 |
|
nekobit joined #luanti-dev |
| 03:58 |
|
nekobit joined #luanti-dev |
| 04:00 |
|
MTDiscord joined #luanti-dev |
| 04:36 |
|
SpaceManiac joined #luanti-dev |
| 04:41 |
|
Warr1024 joined #luanti-dev |
| 04:42 |
|
SpaceManiac joined #luanti-dev |
| 04:47 |
|
fluxionary joined #luanti-dev |
| 04:52 |
|
SpaceManiac joined #luanti-dev |
| 04:53 |
|
SpaceManiac joined #luanti-dev |
| 04:54 |
|
v-rob joined #luanti-dev |
| 05:09 |
|
pattmax joined #luanti-dev |
| 05:13 |
|
SpaceManiac joined #luanti-dev |
| 05:15 |
|
SpaceManiac joined #luanti-dev |
| 06:03 |
|
v-rob joined #luanti-dev |
| 07:06 |
|
MTDiscord joined #luanti-dev |
| 09:54 |
sfan5 |
<exe_virus> maybe 2 bugs, or just one bug here with this mapgen multithreaded thing. Looks like the same mapchunk is called for mapgen twice - maybe we're literally double generating sometimes |
| 09:54 |
sfan5 |
same issue as here https://irc.luanti.org/luanti/2025-05-28? |
| 10:52 |
MTDiscord |
<exe_virus> Yes, aware that may be the case. Did some research, from what I understand, when we overgenerate chunks, they overlap by an entire mapblock +-1 node in the Y direction, correct? That is all my issue was from testing. That said, I did get 4 hits in the same node spot instead of 2 when using breakpoints to slow the system down really badly, so there's definitely a chance of double generation somewhere in there. |
| 10:53 |
MTDiscord |
<exe_virus> But, it wasn't relevant to the y slice issue, it would be a separate bugfix |
| 11:15 |
|
wsor4035 joined #luanti-dev |
| 11:41 |
|
fluxionary joined #luanti-dev |
| 13:15 |
|
matteobin joined #luanti-dev |
| 13:24 |
|
SFENCE_arch joined #luanti-dev |
| 15:14 |
|
Juri joined #luanti-dev |
| 15:15 |
|
Juri joined #luanti-dev |
| 15:15 |
|
Niklp joined #luanti-dev |
| 15:59 |
|
matteobin left #luanti-dev |
| 16:56 |
|
v-rob joined #luanti-dev |
| 17:31 |
|
SFENCE joined #luanti-dev |
| 17:41 |
|
SFENCE joined #luanti-dev |
| 18:00 |
|
v-rob joined #luanti-dev |
| 19:48 |
|
SFENCE joined #luanti-dev |
| 21:29 |
|
Sharpman joined #luanti-dev |
| 22:18 |
|
v-rob joined #luanti-dev |
| 22:33 |
|
panwolfram joined #luanti-dev |
| 23:05 |
|
Eragon joined #luanti-dev |
| 23:16 |
|
fluxionary joined #luanti-dev |
| 23:39 |
|
Guest72 joined #luanti-dev |
| 23:44 |
Guest72 |
Feature request:option in Touchscreen settings to show Aux1 button when "Virtual joystick triggers Aux1 button" enabled. Why:sometimes i need to use Aux1 button without moving but i don't have Aux1 button because virtual_joystick_triggers_aux1 enabled |
