| Time |
Nick |
Message |
| 00:13 |
|
SFENCE joined #minetest-dev |
| 00:19 |
MTDiscord |
<exe_virus> Another merge I'd like to get into 5.10 earlier than later is: 聽https://github.com/minetest/minetest/pull/14643 聽It'd be nice to say, we made Minetest faster |
| 00:49 |
rubenwardy |
do we have any consensus on which PR is the best approach yet? |
| 01:07 |
|
SFENCE joined #minetest-dev |
| 01:12 |
luatic |
i don't think we have some good data yet |
| 01:16 |
|
SFENCE joined #minetest-dev |
| 01:52 |
|
SFENCE joined #minetest-dev |
| 02:11 |
|
SFENCE joined #minetest-dev |
| 02:59 |
MTDiscord |
<exe_virus> No, we don't have that data, luatics implementation is much harder to get right than mine, and his was having issues during testing. Speaking of which, luatic is that work on your near term (2-mo) roadmap? |
| 03:03 |
MTDiscord |
<exe_virus> Alternatively, we have extremely similar methods of integration. It should be very easy to swap out mine when a better one by luatic is done. In the meantime, it's a good performance boost for many entity servers. 聽 After this one the next biggie for entities is tackling collision handling. Those two alone account for like 70% of current time for larger SAO counts when you have entities fighting each other, which happens a lot on |
| 03:03 |
MTDiscord |
Minetest servers haha. |
| 03:11 |
MTDiscord |
<exe_virus> As for my idea for making that more performant... Google/highway looks like a promising cross platform SIMD lib that I could try to use to perform those extremely repetitive calculations in parallel. 聽 If it works, I get it would be really helpful (SIMD) also for ABMs. |
| 03:11 |
MTDiscord |
<exe_virus> I bet* it would be really... |
| 04:00 |
|
MTDiscord joined #minetest-dev |
| 04:47 |
MTDiscord |
<bastrabun> Aplogies, I entirely forgot about the test. If you want, we can pick that up again |
| 06:59 |
|
sofar_ joined #minetest-dev |
| 07:39 |
MTDiscord |
<nrz0> @sfan5 is the merge window opened for the 5.10 ? |
| 07:51 |
MTDiscord |
<nrz0> note, i'm not sure what input now we use in master (after 5.9 release) but we have a weird thing on the / on azerty/FR keyboards. The / from the numpad open the console whereas the / over the : on the keyboard does nothing |
| 08:49 |
[MTMatrix] |
<Zughy> Yes, that's SDL2 |
| 08:52 |
MTDiscord |
<nrz0> when you all problems we have, you see that the keyboards API on OS are just plenty of dozens of years of wtf, and every dev have to care about it |
| 09:27 |
[MTMatrix] |
<Zughy> Core devs: it'd be nice if we could bring PRs under a hundred in the next few days, so to keep things a bit in order |
| 09:28 |
[MTMatrix] |
<Zughy> It shouldn't be difficult, about 20 of them already have at least one approval |
| 09:41 |
MTDiscord |
<nrz0> mergijng #14823, one less |
| 09:41 |
ShadowBot |
https://github.com/minetest/minetest/issues/14823 -- dev: add shell.nix by nerzhul |
| 09:41 |
MTDiscord |
<nrz0> and #14793 will be mergeable when sfan5 or SmallJoker have sufficient time to review the last touch 馃檪 |
| 09:41 |
ShadowBot |
https://github.com/minetest/minetest/issues/14793 -- refacto: rework the GUI element handler function by nerzhul |
| 10:19 |
MTDiscord |
<herowl> Idk what @andrey2470t is doing with the Ambient Light PR (#14343), but IIRC the new vertex format had a bunch of problems, and I found a way to handle it with the old format (just needs some more work from me), so if Andrey doesn't want to continue it, I could pick it up. |
| 10:19 |
ShadowBot |
https://github.com/minetest/minetest/issues/14343 -- Ambient light and server control for it by Andrey2470T |
| 10:20 |
MTDiscord |
<herowl> Also the reviews on #14543 found only nitpicks IIRC, so there's that (eligible for full review from ruben, although I'm obviously open to do more refactoring). |
| 10:20 |
ShadowBot |
https://github.com/minetest/minetest/issues/14543 -- Add gameid aliases by nauta-turbidus |
| 10:21 |
[MTMatrix] |
<Zughy> Closed a few stale PRs, back to 113 |
| 10:41 |
MTDiscord |
<exe_virus> Bastrabun: no not needed right now. You already used my PR and found success with it, and Luatic/Lars branch isn't ready yet for testing with your setup. Once it passes my stress-test setups, then a real world server is a good test to have a go at for measuring performance and stability impacts. 聽 I'll reach out if/when we get to that point |
| 10:56 |
MTDiscord |
<andrey2470t> No, I didn't abandon the PR. I'm just currently working on the texture atlasing and optimizing the clientmap rendering which I've already mostly written and testing now. |
| 11:17 |
sfan5 |
rubenwardy: the artifacts are double-zipped and the website is not updated yet |
| 11:19 |
sfan5 |
security advisories published |
| 11:32 |
rubenwardy |
web#308 |
| 11:32 |
ShadowBot |
https://github.com/minetest/minetest.github.io/issues/308 -- Update downloads for 5.9.0 by rubenwardy |
| 11:32 |
rubenwardy |
why does github actions double zip artifacts |
| 11:34 |
MTDiscord |
<rollerozxa> great question |
| 11:35 |
rubenwardy |
updated |
| 12:47 |
|
Desour joined #minetest-dev |
| 13:09 |
sfan5 |
merging #14825, #14806, #14942, #14918, #14897, maybe #14878, #14869 in 15m |
| 13:09 |
ShadowBot |
https://github.com/minetest/minetest/issues/14825 -- Rework object attachment handling to fix bugs by sfan5 |
| 13:09 |
ShadowBot |
https://github.com/minetest/minetest/issues/14806 -- Fix warning about getVertexTypeDescription reaching its end by Emojigit |
| 13:09 |
ShadowBot |
https://github.com/minetest/minetest/issues/14942 -- Show IME candidate list in Windows by y5nw |
| 13:09 |
ShadowBot |
https://github.com/minetest/minetest/issues/14918 -- TouchScreenGUI: Replace buttonbars with grid menu by grorp |
| 13:09 |
ShadowBot |
https://github.com/minetest/minetest/issues/14897 -- Render clouds as flat when thickness is zero by sfan5 |
| 13:09 |
ShadowBot |
https://github.com/minetest/minetest/issues/14878 -- Sanitize formspec fields server-side by sfan5 |
| 13:09 |
ShadowBot |
https://github.com/minetest/minetest/issues/14869 -- Clamp hotbar selection to slots that exist by Emojigit |
| 13:09 |
MTDiscord |
<luatic> sfan5: luk3yx has voiced a concern on #14878, have you looked at that yet? |
| 13:09 |
ShadowBot |
https://github.com/minetest/minetest/issues/14878 -- Sanitize formspec fields server-side by sfan5 |
| 13:14 |
sfan5 |
I am in the process of replying |
| 13:14 |
sfan5 |
but let's just postpone merging it |
| 14:29 |
|
SFENCE joined #minetest-dev |
| 14:31 |
|
Desour joined #minetest-dev |
| 15:12 |
MTDiscord |
<luatic> first draft for the blog post, please give feedback since it's 34 掳C here and i'm melting: https://github.com/minetest/blog/pull/152 |
| 15:18 |
MTDiscord |
<athozus> LGTM (not a dev) |
| 15:31 |
pgimeno |
@ Discord users, we IRC users are considerate with you by using @ for invocations, please be considerate with us by using # followed by the bug number instead of just giving a link, thanks |
| 15:33 |
Desour |
+1 |
| 15:35 |
MTDiscord |
<warr1024> https://github.com/minetest/blog/pull/152 and #152 are very different things |
| 15:35 |
ShadowBot |
https://github.com/minetest/minetest/issues/152 -- Locked furnaces, doors, signs |
| 15:37 |
celeron55 |
well i suppose it would be blog#152 (no idea if the bot understands that though) |
| 15:38 |
MTDiscord |
<warr1024> well I guess there's our answer |
| 15:39 |
celeron55 |
(frankly the bot could just output the title based on the link) |
| 15:45 |
[MTMatrix] |
<Zughy> On Matrix we see nothing with # + the number, so the link is preferable |
| 15:47 |
MTDiscord |
<warr1024> If the # thing works right, then the bot SHOULD post the link, along with the summary. 聽Do you not see the bot messages on Matrix? |
| 15:47 |
[MTMatrix] |
<Zughy> Nope |
| 15:48 |
[MTMatrix] |
<Zughy> luatic: I'd avoid using the MTG texture pack for the cover. Go with something nice like RPG16 |
| 15:51 |
[MTMatrix] |
<Warr1024 (@warr1024:matrix.org)> Okay, that IS weird. 聽Feels like it might have been an oversight in config, e.g. the bot was filtered out without considering that consequence. |
| 15:52 |
[MTMatrix] |
<Warr1024 (@warr1024:matrix.org)> Oh, or maybe the discord/IRC bridge bridges NOTICE as well as PRIVMSG but the Matric/IRC one only does PRIVMSG? |
| 15:52 |
[MTMatrix] |
<Warr1024 (@warr1024:matrix.org)> Yeah, checking it in an actual IRC client, the bot sends notices, so that might be the difference |
| 15:54 |
Warr1024 |
That name suffix business is weird too 馃 |
| 15:56 |
MTDiscord |
<rubenwardy> https://github.com/minetest/blog/pull/152 |
| 15:56 |
MTDiscord |
Command sent from Discord by rubenwardy: |
| 15:56 |
MTDiscord |
!title |
| 15:56 |
ShadowBot |
Add 5.9.0 release blog post by appgurueu 路 Pull Request #152 路 minetest/blog 路 GitHub |
| 15:57 |
MTDiscord |
Command sent from Discord by rubenwardy: |
| 15:57 |
MTDiscord |
!title https://github.com/minetest/blog/pull/152 |
| 15:57 |
ShadowBot |
MTDiscord: Add 5.9.0 release blog post by appgurueu 路 Pull Request #152 路 minetest/blog 路 GitHub |
| 15:59 |
MTDiscord |
<warr1024> Works. 聽Got the title on IRC, got the embed on Matrix, got both on Discord but it's not terribly obtrusive. |
| 16:17 |
pgimeno |
NOTICE messages are typically not read by bots, to avoid bot loop problems, but maybe the bridges should make exceptions in that case |
| 16:17 |
pgimeno |
https://github.com/minetest/blog/pull/152 -- Add 5.9.0 release blog post by appgurueu |
| 16:17 |
pgimeno |
blog#152 |
| 16:18 |
pgimeno |
I'll provide that until ShadowBot supports it |
| 16:20 |
|
Desour joined #minetest-dev |
| 16:21 |
|
luatic joined #minetest-dev |
| 16:28 |
luatic |
Zughy: done |
| 16:37 |
MTDiscord |
<luatic> rubenwardy: updated |
| 16:42 |
rubenwardy |
ok fixed some things, looks good |
| 16:45 |
luatic |
alright thanks. will probably merge in 15m. |
| 16:52 |
MTDiscord |
<nrz0> merging #14793 it has been extenvely tested |
| 16:52 |
ShadowBot |
https://github.com/minetest/minetest/issues/14793 -- refacto: rework the GUI element handler function by nerzhul |
| 16:52 |
pgimeno |
https://github.com/minetest/minetest/pull/14793 -- refacto: rework the GUI element handler function by nerzhul |
| 16:53 |
MTDiscord |
<nrz0> it's a big refactor on the handling of each GUI element, no issue expected, in case of you have the ref, the three pass should have seen all issues 馃檪 |
| 16:53 |
Noisytoot |
Zughy: The bot posts links as a NOTICE. If the Matrix bridge doesn't support relaying notices then it should be fixed. |
| 16:55 |
Krock |
pgimeno: chain reaction |
| 16:58 |
Noisytoot |
RFC 1459 only specifies that automatic replies shouldn't be sent in reply to NOTICEs, and relaying to matrix isn't an automatic reply |
| 16:58 |
luatic |
nrz0: out of curiosity, is there a reason you write "refacto"? |
| 17:00 |
MTDiscord |
<nrz0> luatic, wdym ? the title is bad for you ? i'm not english native, for me it's a refactoring of the GUI handlers |
| 17:00 |
Desour |
@nrz0: but why not "refactor"? |
| 17:00 |
Noisytoot |
You wrote "refacto" instead of "refactor" |
| 17:00 |
luatic |
nrz0: i would usually expect to see "refactor", with a trailing "r" :) |
| 17:01 |
luatic |
publishing the blog post now |
| 17:01 |
MTDiscord |
<nrz0> ok, it'll try to don't forget the R next time 馃槃 |
| 17:01 |
MTDiscord |
<nrz0> maybe because in fresh we said, i refacto the code 馃槃 |
| 17:03 |
Krock |
in france, the programs don't run. they turn. |
| 17:04 |
|
SFENCE joined #minetest-dev |
| 17:08 |
MTDiscord |
<athozus> they can at least walk |
| 17:08 |
MTDiscord |
<athozus> (in French) |
| 17:10 |
luatic |
5.9.0 release blog post: https://blog.minetest.net/2024/08/12/5.9.0-released/ |
| 17:10 |
|
SFENCE joined #minetest-dev |
| 17:26 |
luatic |
poked the fedora minetest package maintainers via mail |
| 17:33 |
luatic |
e-mail, that is, before anyone thinks i sent them a letter. |
| 18:15 |
|
YuGiOhJCJ joined #minetest-dev |
| 18:17 |
|
SFENCE joined #minetest-dev |
| 19:32 |
Mantar |
Refacto is my favorite superhero. He can make Spaghetti-man go straight |
| 19:41 |
|
SFENCE joined #minetest-dev |
| 20:51 |
ROllerozxa |
I've opened an issue at F-Droid mentioning the Android version code issue for 5.9.0: https://gitlab.com/fdroid/fdroiddata/-/issues/3333 |
| 20:57 |
[MTMatrix] |
<grorp> rollerozxa: thank you. |
| 21:06 |
sfan5 |
thought: we could default num_emerge_threads > 1 for map眉gens where it's known to be safe (singlenode only I think lol) |
| 21:09 |
celeron55 |
it will have effects if someone uses the mapgen lua environment |
| 21:10 |
rubenwardy |
#9357 |
| 21:10 |
ShadowBot |
https://github.com/minetest/minetest/issues/9357 -- Mapgen: "Unfinished" y-slices with num_emerge_threads > 1 |
| 21:10 |
pgimeno |
https://github.com/minetest/minetest/issues/9357 -- Mapgen: "Unfinished" y-slices with num_emerge_threads > 1 |
| 21:11 |
rubenwardy |
What is the issue - biomes or ores? |
| 21:11 |
rubenwardy |
if lua mapgens can't access the thread unsafe stuff then num_emerge_threads > 1 should work in singlenode right |
| 21:12 |
rubenwardy |
but there is generate_ores |
| 21:12 |
rubenwardy |
and biomemap |
| 21:12 |
rubenwardy |
presumably the latter is blank in singlenode |
| 21:13 |
celeron55 |
anything breaks where the mapgen tries to act as if all neighboring chunks are either generated or not generated. with num_emerge_threads > 1, there's a third option: a neighboring chunk is also being generated |
| 21:14 |
celeron55 |
the chunks don't see each other in that case in their half-finished states |
| 21:16 |
celeron55 |
if you design for that when writing the mapgen, it isn't necessarily an issue. if you don't, then it can easily become a problem |
| 21:16 |
Desour |
couldn't we just schedule the mapchunk generation such that never two adjacent chunks are generated concurrently? |
| 21:21 |
sfan5 |
lua has no way to tell if a neighboring chunk is being generated or not generated yet |
| 21:21 |
sfan5 |
so it can't possibly be thread unsafe |
| 21:22 |
sfan5 |
the "don't generate adjacent chunks" feels like a hack but I bet it would solve the issue engine mapgens have |
| 21:22 |
Desour |
async lua mapgens can overgenerate a 1 block thick layer, afaik |
| 21:23 |
Desour |
so, if two are adjacent, there's a conflict |
| 21:23 |
sfan5 |
this conflict doesn't depend on threading |
| 21:26 |
Desour |
well, sure, the structures generated into the overgeneration layer can intersect. but if they do, one expects the newer one to just replace the older one. if a mapgen completely fills the overgeneration layer, like the inside blocks, it does something wrong |
| 21:34 |
|
YuGiOhJCJ joined #minetest-dev |
| 21:34 |
rubenwardy |
#14963 |
| 21:34 |
ShadowBot |
https://github.com/minetest/minetest/issues/14963 -- Generate Android versionCode from Major.Minor.Patch by rubenwardy |
| 21:34 |
pgimeno |
https://github.com/minetest/minetest/pull/14963 -- Generate Android versionCode from Major.Minor.Patch by rubenwardy |
| 21:34 |
rubenwardy |
echo |
| 21:36 |
Desour |
echo |
| 21:38 |
MTDiscord |
<luatic> pgimeno: you can turn it off i think, or only leave it on for blog#id |
| 21:44 |
MTDiscord |
<mistere_123> so what are the effects of the mapgen bug? 聽Trees being cut off at mapblock boundaries? |
| 21:45 |
MTDiscord |
<mistere_123> (trees, and other decorations)? |
| 21:47 |
MTDiscord |
<luatic> i think a fully custom mapgen like SD's that doesn't rely on overgeneration could deal with it without any issues at all, but i'm not sure |
| 21:47 |
pgimeno |
@luatic my bad, I thought I did that but I was wrong |
| 21:48 |
pgimeno |
#14963 |
| 21:48 |
ShadowBot |
https://github.com/minetest/minetest/issues/14963 -- Generate Android versionCode from Major.Minor.Patch by rubenwardy |
| 21:48 |
pgimeno |
https://github.com/minetest/blog/pull/100 -- fix form typo by MisterE123 |
| 21:48 |
pgimeno |
blog#100 |
| 21:49 |
pgimeno |
fixed |
| 22:07 |
MTDiscord |
<ronoaldo> Hello! I just tested Minetest 5.9 and I would like to say a huge congratulations for the work done in the graphics part ... I got massive FPS improvements and the visuals are awesome! |
| 22:33 |
|
panwolfram joined #minetest-dev |
| 23:05 |
|
Eragon joined #minetest-dev |
