Minetest logo

IRC log for #minetest-dev, 2024-11-10

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
00:06 sfan5 I don't understand why you are insisting on spending time on perfecting code that will literally be deleted after release
00:14 fluxionary_ joined #minetest-dev
00:21 sfan5 20:19 <+Krock> How shall we proceed with the Collision bugs/PRs?
00:21 sfan5 imo: revert for release, afterwards merge the unit tests, then try again with the whole idea
00:25 hwpplayer1 joined #minetest-dev
03:18 [MatrxMT] <grorp> +1
05:00 MTDiscord joined #minetest-dev
05:21 SFENCE joined #minetest-dev
05:59 whosit joined #minetest-dev
06:17 YuGiOhJCJ joined #minetest-dev
06:24 SFENCE joined #minetest-dev
08:03 Sharpman joined #minetest-dev
08:40 ainema joined #minetest-dev
09:22 Warr1024 joined #minetest-dev
09:47 Warr1024 joined #minetest-dev
11:33 sfan5 merging #15400, #15404, #15405 in like 30m
11:33 ShadowBot https://github.com/minetest/minetest/issues/15400 -- Revert "Fix collisions with long dtime, in particular with bouncing" by SmallJoker
11:33 ShadowBot https://github.com/minetest/minetest/issues/15404 -- Re-fix CAO mesh lighting with shaders disabled by sfan5
11:33 ShadowBot https://github.com/minetest/minetest/issues/15405 -- Update credits for 5.10.0 by sfan5
12:02 MTDiscord <luatic> sfan5: ok. if i come up with what i consider a cleaner fix for #15404 in time i might propose that.
12:02 ShadowBot https://github.com/minetest/minetest/issues/15404 -- Re-fix CAO mesh lighting with shaders disabled by sfan5
12:03 sfan5 i thought the release date was today
12:08 MTDiscord <josiah_wi> The axisAlignedCollision unit tests have a bug/typo. The dtime is supposed to be calculated by the function under test, and it's checked after every test, but the code author initialized the dtime to the expected value before every call.
12:15 MTDiscord <luatic> Note that the dtime is used both as input parameter for a maximum dtime and as output value for the dtime of a collision if one occurred
12:16 [MatrxMT] <Zughy> We are releasing today, yes
12:18 [MatrxMT] <Zughy> So please merge it
12:19 MTDiscord <josiah_wi> Thanks, I'll make a note to check it and fix the docstring.
12:30 MTDiscord <josiah_wi> This time I've found an actual copy-paste error. The last two unit-tests for the -X direction are duplicates of each other. I assume one of them is supposed to test something new, but the author forgot to change the test values.
12:30 MTDiscord <josiah_wi> I diffed the test cases to make sure I wasn't reading it wrong.
13:05 Desour joined #minetest-dev
13:09 ROllerozxa I can get the basics of the fastlane metadata done for the release unless anyone else really wants to
13:10 sfan5 sure, PRs welcome
13:32 YuGiOhJCJ joined #minetest-dev
13:33 ROllerozxa https://github.com/minetest/minetest/pull/15411
13:34 ROllerozxa skipped screenshots for this PR, most important for 5.10 is updating the name and just getting an icon for the app on F-Droid
14:33 cx384 joined #minetest-dev
15:05 vampirefrog joined #minetest-dev
15:41 MTDiscord <luatic> a bit late but i'd like to get #15412 in if possible
15:41 ShadowBot https://github.com/minetest/minetest/issues/15412 -- Work around driver bug causing shaders to not work by appgurueu
15:51 sfan5 personally I would have just went with the "assume large number" workaround
16:04 MTDiscord <luatic> sfan5: you're right tbh, there won't be > 256 uniform names realistically
16:04 MTDiscord <luatic> and it seems to be a bit more of a pain to do the """proper""" workaround than anticipated
16:08 MTDiscord <luatic> reverted to the simpler workaround
16:23 sfan5 we could even do 1023 to be absolutely very safe
16:31 MTDiscord <luatic> 255 seems very safe already (who would have a 255 character long identifier?) ¯_(ツ)_/¯
16:34 MTDiscord <luatic> merging #15412 in 10m then
16:34 ShadowBot https://github.com/minetest/minetest/issues/15412 -- Work around driver bug causing shaders to not work by appgurueu
16:34 sfan5 we're not using java or ProjectContractChargingPeriodProjectAccountReferenceVMFactoryBuilderStrategy would be realistic
16:35 MTDiscord <luatic> sfan5: i mean even that is just a measly 76 characters. it still fits on my screen :P
16:35 sfan5 i did measure it and was disappointed it's not even 100 chars, yes
16:46 SFENCE joined #minetest-dev
16:47 SFENCE_ joined #minetest-dev
16:47 sfan5 release today?
16:47 sfan5 does anyone want to comment on the description text in #15411?
16:47 ShadowBot https://github.com/minetest/minetest/issues/15411 -- Add Fastlane metadata for F-Droid by rollerozxa
17:00 sfan5 either way I'll import the translations now
17:15 sfan5 commits -> https://github.com/minetest/minetest/compare/ci
17:17 SFENCE_ But we plan/have a consensus on drop no shaders build in the future. Am I right?
17:18 rubenwardy Yes
17:18 sfan5 yes
17:18 [ Shaders are slower
17:19 MTDiscord <luatic> Report that properly please
17:20 sfan5 SFENCE_: merging #15303 before release would allow you to produce a signed build, right?
17:20 ShadowBot https://github.com/minetest/minetest/issues/15303 -- Support generation of working Xcode project for signature purposes on MacOS by sfence
17:22 [ I would but that would require using github
17:23 SFENCE_ sfan5: Not exactly. But with it merged I will try to go thtought notirization process and it I success, signed build will be avaliable.
17:23 sfan5 so let's merge it then
17:25 SFENCE_ :)
17:27 MTDiscord <luatic> [_: don't worry, here you go: https://pastebin.com/k7uVxBNy. fill it out and send it back :)
17:27 Krock okay. I'm here too. Checking what needs to be updated in the changelog.
17:29 SFENCE_ We still do not have updated translation files in master branch (Minetest -> Luanti rename)?
17:29 rubenwardy So the entire point of shaders is to move work to the GPU. With a shader only pipeline more work can be done on improving architecture and optimising draw calls, improving performance even on low end devices. This is ultimately why we want to drop fixed pipeline support - it's long obsolete and holding the project back
17:30 sfan5 SFENCE_: about to be done
17:31 SFENCE_ rubenwardy: Agree. Some thing, like night visions etc, are not possible/easy to be done without shaders.
17:31 SFENCE_ sfan5: nice
17:31 rubenwardy Yeah that too. Shaders make a lot possible that would be expensive or impractical with fixed pipeline
17:32 sfan5 SFENCE_: https://github.com/minetest/minetest/compare/ci changes are here, feel free to check
17:32 sfan5 I'll merge then once the CI run is done
17:35 SFENCE_ I will if train wiif connection will allows me to do it.
17:38 MTDiscord <luatic> it looks like GH does not handle race conditions on labeling well
17:39 SFENCE_ Do I see well that po//luanti.pot is not changed?
17:39 sfan5 correct
17:39 sfan5 that's only done on updatepo.sh
17:40 sfan5 (we did that weeks ago)
17:41 sfan5 src/gui/guiFormSpecMenu.cpp
17:41 sfan5 830:// Temporary fix for issue #12581 in 5.6.0.
17:41 sfan5 👀
17:41 ShadowBot https://github.com/minetest/minetest/issues/12581 -- Inventory slots in Repixture look different in 5.6.0-dev (with gui_scaling_filter = true)
17:41 sfan5 "temporary"
17:41 Krock sfan5: temporary like this comment https://github.com/minetest/minetest/blob/master/src/network/networkpacket.h#L110-L111
17:41 SFENCE_ Ok. I miss that.
17:49 Krock Meeting in 11 minutes. Mostly milestone topics for the release.
17:52 grorp joined #minetest-dev
17:53 SFENCE_ From 19:05 to about 19:40 I will be going from the train station to my flat. So if there will be something on me in the meeting, I will answer later.
17:54 Krock alright. thanks.
17:55 sfan5 merging #15303, #15411 in 11m
17:55 ShadowBot https://github.com/minetest/minetest/issues/15303 -- Support generation of working Xcode project for signature purposes on MacOS by sfence
17:55 ShadowBot https://github.com/minetest/minetest/issues/15411 -- Add Fastlane metadata for F-Droid by rollerozxa
17:59 [ luatic: https://paste.sr.ht/~noisytoot/ecff9297771e47fc08c1eb8ece7972250a49fc6c (raw: https://paste.sr.ht/blob/257123e82e16ff616a86c48803a5c141aca0bacd)
18:00 sfan5 [: is it intentional that you are running a debug build?
18:02 [ sfan5: It's because that's what I had compiled. I can try with a release build of 5.10.0-rc1, but since it also happens on a release build of 5.6.1 (from Debian) I expect it'll be the same.
18:04 sfan5 looking at https://www.notebookcheck.com/Intel-Graphics-Media-Accelerator-4500MHD-GMA-X4500MHD.9877.0.html your gpu was barely suitable for basic gaming back in the day. with it being old enough to buy beer (in germany) now I think that's just the answer.
18:04 Krock Pinging those who reacted to the discussion announcement: Desour Zughy
18:04 Desour o/
18:04 sfan5 looks like you should be able to enjoy quake 3 however
18:05 grorp hi
18:06 [MatrxMT] <Zughy> hi
18:08 [MatrxMT] <Zughy> I guess we're ready for a release?
18:08 sfan5 yes
18:08 [MatrxMT] <Zughy> what to do with https://github.com/minetest/minetest.github.io/pull/311?
18:08 Krock the translations are updated, and the milestone empty.
18:09 sfan5 we should merge that at release time the latest
18:09 sfan5 and then ping c55 to actually hook up luanti.org
18:09 Krock luanti.org looks just fine as-is  8)
18:10 Krock though this PR is technically not bound to the binary release
18:10 [MatrxMT] <Zughy> I see there's no draft for a 5.10 entry in the blog. But frankly we don't have enough manpower to maintain it..
18:10 sfan5 could simply copy the forum post there?
18:10 [MatrxMT] <Zughy> we should probably look for people willing to keep it up to date
18:10 [MatrxMT] <Zughy> (for the future)
18:12 cx384 Hi, I was just fixing up a PR. Maybe someone wants to review it when the feature freeze is done. #15415
18:12 ShadowBot https://github.com/minetest/minetest/issues/15415 -- Support LuaVoxelManip in core.spawn_tree by cx384
18:12 Krock The blog also needs an update of the name in a few places - at least in medium-term.
18:12 sfan5 time to close out the release then?
18:12 [MatrxMT] <Zughy> +1
18:12 Krock what does "close out" mean in this context?
18:12 sfan5 finish
18:13 Krock sure.
18:13 [ sfan5: Luanti is playable on it and it's the most recent GPU that works without blobs on RYF hardware.
18:13 Krock conveniently, the Android version is now generated automatically.
18:15 sfan5 proceeding then, as soon as the wiki page loads
18:15 Krock testing the win32 builds on windows...
18:17 Krock hmm. The VS 2019 portable builds did exit with code 1, but are marked as passed nonetheless.
18:17 sfan5 that's normal. it's the unit tests
18:17 Krock ah yes. I just saw those. A few about translations.
18:18 sfan5 the wiki page could use a few updates. some things are now out-of-date. might do that tomorrow.
18:23 sfan5 release commits are ready. if you say the win32 build works we can go ahead
18:24 sfan5 merging website#311 now btw
18:24 sfan5 web#311 ?
18:24 ShadowBot https://github.com/minetest/minetest.github.io/issues/311 -- Initial rename of Minetest to Luanti by rollerozxa
18:24 sfan5 yeah
18:25 MTDiscord <jordan4ibanez> Does opengl_debug get enabled in dirty and debug builds?
18:26 sfan5 debug builds yes. "dirty" has zero effect on anything
18:27 ROllerozxa oo website rename time
18:27 MTDiscord <jordan4ibanez> When you do the stock cmake . -DRUN_IN_PLACE=1 does that put it into debug mode?
18:27 ROllerozxa is celeron55 around this evening to migrate the landing page to the new domain?
18:28 sfan5 no, debug mode is not the default
18:30 Krock Windows has serious problems with download priorities. It's still trying to join a local network hosted devtest server.
18:30 Krock tested low bandwidth connections, I guess.
18:31 grorp github needs a shrug emoji reaction
18:31 Krock dynamic shaders work too. Ready to go.
18:32 Krock *shadows
18:32 sfan5 commits pushed. waiting for CI to produce builds now
18:33 Krock got the GLSL warning but couldn't spot anything unusual
18:34 grorp the "failed to retrieve uniform information" one?
18:35 Krock yes.
18:37 Krock Intel HD 4000, on Windows 10.0.19045  Judging by the update queue that's probably 21H2 or 22H1
18:38 sfan5 the latter
18:40 Krock So 22H2 might be still one of the updates that fail when Windows is booted through Grub
18:40 rubenwardy edited the play store listing, waiting for CI to finish. Listing change will be published at the same time as update
18:42 grorp I'll prepare a forum post if nobody else has
18:42 rubenwardy missing a blog post, but can come later
18:48 SFENCE_arch joined #minetest-dev
18:49 SFENCE_arch I am here.
18:49 cx384 hi
18:56 sfan5 windows and macos uploaded to gh
18:58 sfan5 minetest_android_deps tagged
18:59 sfan5 anything missing or should I proceed?
19:01 SFENCE_arch I remeber this was renamed in cmake files in some PR to adriod_deps or somethink like it.
19:01 sfan5 🤷
19:02 sfan5 repo name has not changed
19:04 sfan5 continuing.
19:05 MTDiscord <luatic> I should be able to base a blog post on grorp's RC1 forum post and the changelog
19:05 grorp I already have a forum post ready :)
19:06 grorp ah I misread
19:07 Topic for #minetest-dev is now Luanti (fka Minetest) core development and maintenance. Chit chat goes to #minetest. https://dev.minetest.net/ https://irc.minetest.net/ https://github.com/minetest
19:07 rubenwardy apks uploaded and app sent to play store review
19:07 sfan5 I'll edit the website and leave the rest to you
19:08 rubenwardy launchpad importers seem to be struggling, there's no 'start import'
19:09 grorp how do you pin forum topics?
19:09 rubenwardy done
19:09 rubenwardy you may need to be a moderator
19:10 rubenwardy there's a hammer drop down and "change to sticky"
19:11 rubenwardy assuming the forum loads
19:12 MTDiscord <luatic> your optimism is misplaced
19:12 MTDiscord <jordan4ibanez> Well put it in the correct storage compartment
19:17 [MatrxMT] <Zughy> Didn't Buckaroo offered to handle the forum themselves?
19:17 [MatrxMT] <Zughy> *offer
19:17 sfan5 okay I'm finished
19:18 MTDiscord <luatic> tomorrow would have been 11.11. 😔
19:19 Desour let's do 6 major releases and 11 minors in one day, so we can release luanti 11.11 tomorrow
19:20 Desour if someone who can do this has time, I'd like a contentdb grep: `waving *= *3`
19:21 MTDiscord <luatic> Desour: wrong paste?
19:21 MTDiscord <luatic> ah this is supposed to be a regex
19:22 Desour idk the exact syntax x)
19:22 sfan5 will push https://github.com/minetest/minetest/commit/9d90c972a6217c5906445f21791d34665e37386d once ci completes
19:22 Desour (reason for this: I want to see how many mods are using liquid waving for non-liquid drawtype nodes)
19:24 hwpplayer1 joined #minetest-dev
19:24 sfan5 we should also prioritize #14964 this time for real
19:24 ShadowBot https://github.com/minetest/minetest/issues/14964 -- SDL: Use scancodes for keybindings by y5nw
19:24 sfan5 so we can stop disabling sdl every release
19:25 MTDiscord <luatic> sfan5: agreed
19:25 Krock I'd favour platform-specific printable character lookups over that approach
19:25 Krock I've yet to find an application that relies on scancodes
19:25 MTDiscord <luatic> didn't we agree a while ago that we ultimately want to use scancodes?
19:25 sfan5 don't all games do that?
19:26 MTDiscord <luatic> it makes sense because you want things to be in the right position rather than to have the right label on them
19:26 MTDiscord <luatic> "WASD" for example
19:26 [MatrxMT] <Zughy> We should change the website link in the repo, but we should also get rid of that fucking joke that is luanti.org at the moment
19:26 MTDiscord <luatic> a neo or dvorak user won't mind that being.. something else so long as it's in a comfortable position
19:26 MTDiscord <luatic> tbh i really don't see why luanti.org can't just be a redirect to minetest.net
19:27 MTDiscord <luatic> or the other way around with luanti.org properly set up (whichever is easier for celeron to do)
19:27 ROllerozxa well it doesn't seem like c55 is here right now
19:27 Krock hmm. I don't have any particular reference of a game where it would be obvious which way it functions.
19:27 ROllerozxa (I mentioned his username here earlier to see if he was available)
19:29 grorp left #minetest-dev
19:29 Krock there's mixed use of characters. some need to be arranged in a specific way (WASD), and others may hold a specific purpose (/KC)
19:30 Krock s/specific purpose/literal meaning/
19:32 [MatrxMT] <y5nw> sfan5: Is it also possible to look at #15152? Using scancodes would open usecases for binding different keys to the same action (e.g. left and right shift for sneaking, since there is no single "Shift key" scancode), and having a reworked key change menu based on formspecs makes this feature easier than working with hardcoded GUI elements and stuff even if the menu is still written in C++.
19:32 ShadowBot https://github.com/minetest/minetest/issues/15152 -- Rework guiKeyChangeMenu by y5nw
19:33 sfan5 first things first
19:35 [MatrxMT] <y5nw> Krock: allowing both scancodes and keycodes would not really be an option unless we separate things cleanly since it can become a source of conflicting keybindings. With the AZERTY layout the Z key is on the same position as W on QWERTY/QWERTZ so it would be bound to both moving forward and zooming.
19:36 rubenwardy Krock: unfortunately we can't just bind to / on all platforms
19:37 rubenwardy That would require SDL3
19:37 rubenwardy instead it'll be the key in that position, not based on the character value
19:37 rubenwardy The user can always rebind
19:37 rubenwardy It's either do this or not ship with sdl2
19:38 sfan5 realistically we can't use sdl3 either because it's not even released yet
19:38 rubenwardy Yeah exactly, it's not an option
19:39 sfan5 for the platforms where we control the binaries we could theoretically use it, but eh
19:39 Krock rubenwardy: isn't the 3rd option to have compatibility code on our own? or does SDL2 not provide enough information?
19:40 sfan5 AFAIK SDL2 literally does not give us enough information to be compatible to the old way, which is why we're in this entire situation
19:40 Krock y5nw: indeed. we'd have to use one notation or the other to avoid conflicts - or display them well.
19:40 rubenwardy Yeah exactly. We found this out during the 5.9.0 feature freeze
19:40 Krock ah okay. thanks for the explanation.
19:41 rubenwardy The best way forward is to use scan code. It's the best result for users
19:48 rubenwardy Play update is live
19:48 rubenwardy Set at 10%
19:49 rubenwardy !title https://play.google.com/store/apps/details?id=net.minetest.minetest
19:49 ShadowBot rubenwardy: Error: That URL appears to have no HTML title within the first 64KB.
19:49 [MatrxMT] <grorp> huh, how was it approved so quickly
19:52 rubenwardy Yeah was expecting the listing change to take much longer
19:52 MTDiscord <jordan4ibanez> They're in here, watching
19:56 Desour they were also surprisingly fast with renaming the minetest topic on youtube to luanti, iirc
19:56 MTDiscord <luatic> can i get a wiki account so i can fix minor things in the changelog?
19:57 Desour only a matter of time until google launches its alternative to microsoft's minecraft
19:58 ROllerozxa Desour: google populates their youtube game topics from, among other places, google play game listings, so makes sense since the play listing has been updated
19:58 Desour ah, I see
19:58 ROllerozxa I made an issue on the F-Droid data repo to notify them of the release and that we have Fastlane metadata now: https://gitlab.com/fdroid/fdroiddata/-/issues/3421
19:58 sfan5 it's not clear if we should ever change the app id
19:59 ROllerozxa for android? you should not
20:00 ROllerozxa iirc even google play is still called com.android.market or something in the app ID from back when it was called android market. it's not worth it changing the ID
20:14 Sharpman joined #minetest-dev
20:42 SFENCE joined #minetest-dev
20:51 Desour ok, I don't need the grep anymore. I found a liquid-like waving nodebox: waterlily
21:06 SFENCE joined #minetest-dev
21:20 SFENCE joined #minetest-dev
22:14 [MatrxMT] <Zughy> reporting from Matrix chat:
22:14 [MatrxMT] <Zughy> "Now that the new name is shown on the client, please please please update the luanti.org website to something I can show parents. If I show them the way it is, they will think I am mad. Our director asked me already what is with it."
22:15 SFENCE joined #minetest-dev
22:15 [MatrxMT] <Zughy> I'd also like to say that I'm frankly highly annoyed by how some of us did their best to have Luanti look like a professional project with the blog post, to then see it all hijacked with that ridiculous homepage we currently have
22:18 [MatrxMT] <Zughy> I bring my arse to small and big conferences, introducing Luanti to people, with organisations reimbursing my expenses, to then.. have to deal with that? Like, seriously? What am I supposed to say to people opening luanti.org right now? Or, you know, to my sponsors
22:20 [MatrxMT] <Zughy> It's been weeks
22:20 ROllerozxa it's really not a good look
22:21 ROllerozxa I was hoping my website PR could have been merged like a week or two ago so we could have moved things over with c55 present well before the release happened
22:24 [MatrxMT] <Zughy> Also, related question: since c55 is not that active and AFAIK people like Buckaroo offered to take care of things like the forum, why exactly is not happening? I think this dictator for life thing has been damaging the project, since the dictator is barely present and there are stuff that depend on him
22:25 [MatrxMT] <Zughy> *important stuff
22:32 ROllerozxa I assume it's the user data that he's concerned about handing out or making sure that the person hosting it will be able to do it for however many years
22:32 ROllerozxa but I do think I would trust buckaroo with that. and I don't see him leaving luanti anytime soon
22:33 SFENCE joined #minetest-dev
23:04 sfan5 entrusting someone with user data is a big deal
23:04 sfan5 as for the website I don't see why it couldn't have happened sooner
23:09 SFENCE joined #minetest-dev
23:29 SFENCE joined #minetest-dev
23:32 panwolfram joined #minetest-dev
23:48 SFENCE joined #minetest-dev
23:55 hwpplayer1 joined #minetest-dev

| Channels | #minetest-dev index | Today | | Google Search | Plaintext