Minetest logo

IRC log for #minetest-dev, 2022-07-04

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

All times shown according to UTC.

Time Nick Message
00:01 MTDiscord <GreenXenith> basically string interpolation
00:01 schwarzwald[m] I should really find something useful to PR, but I've had trouble getting the confidence or the time.
00:02 Zughy[m] well, there are a lot of bugs ?
00:02 MTDiscord <GreenXenith> See if theres any smaller bug issues
00:03 Zughy[m] #10037 is very annoying, for instance
00:03 ShadowBot https://github.com/minetest/minetest/issues/10037 -- Physics override not working properly when speed is set to 0 while moving
00:03 schwarzwald[m] The dunning-kruger effect must not work on me because I know I have absolutely no idea what I'm doing and then do it wrong anyway. ?️
00:05 Zughy[m] and I guess #12403 is easier
00:05 ShadowBot https://github.com/minetest/minetest/issues/12403 -- `is_visible = false` doesn't work on players
00:05 Zughy[m] (yes, I'm pushing my reports)
00:05 MTDiscord <GreenXenith> whats your opinion on the element syntax above, zughy
00:06 MTDiscord <GreenXenith> and should I bother using something like it for the main menu
00:07 Zughy[m] can't really say, I'm a simple mind, I see a formspec, I want to cry
00:07 MTDiscord <GreenXenith> thats fair
00:07 MTDiscord <GreenXenith> Speaking of though, how would a "better" GUI system even work
00:07 MTDiscord <GreenXenith> the string based system structure-wise isnt awful
00:08 MTDiscord <GreenXenith> the only other established systems are usually tag-based ([ht,x]ml)
00:10 MTDiscord <GreenXenith> meh, pointless exercise for now. Ill get back to main menuing.
00:13 schwarzwald[m] > I see a formspec, I want to cry
00:13 schwarzwald[m] Zughy might've accidentally hinted at one possible solution. It's not necessary for modders to ever manually write formspecs. They could be in a binary format for that matter, although I think there's a lot of value in them being human readable. They could be generated using tooling and stored in their own file format.
00:14 schwarzwald[m] Even the existing formspec language could be used that way, and nobody would ever have to look at formspec code again. But if the tooling stops being maintained, nothing is lost, as the format is simple to modify by hand if necessary, or write new tooling for.
00:17 Zughy[m] related: #6527
00:17 ShadowBot https://github.com/minetest/minetest/issues/6527 -- Formspec replacement
00:20 YuGiOhJCJ joined #minetest-dev
00:29 YuGiOhJCJ joined #minetest-dev
00:36 YuGiOhJCJ joined #minetest-dev
00:47 Soni joined #minetest-dev
01:13 Baytuch joined #minetest-dev
01:53 Baytuch joined #minetest-dev
03:40 Baytuch joined #minetest-dev
04:00 MTDiscord joined #minetest-dev
04:04 cheapie joined #minetest-dev
04:07 erle joined #minetest-dev
04:15 Baytuch joined #minetest-dev
04:37 BuckarooBanzai joined #minetest-dev
04:38 Krock joined #minetest-dev
05:30 Baytuch joined #minetest-dev
05:42 calcul0n joined #minetest-dev
05:45 calcul0n joined #minetest-dev
06:14 independent56 joined #minetest-dev
06:20 erle joined #minetest-dev
09:25 independent56 joined #minetest-dev
09:27 independent56 joined #minetest-dev
09:40 MTDiscord <luatic> I had considered this too but ultimately rejected it. I went with S-expressions instead: {"label", {1,2}; "foo bar"}
10:01 independent56 joined #minetest-dev
10:09 calcul0n_ joined #minetest-dev
11:08 independent56 joined #minetest-dev
11:12 appguru joined #minetest-dev
11:18 MTDiscord <Warr1024> Of course, what we need for a formspec replacement is yet another bespoke syntax instead of something like JSON or Lua that we already have serializers for.
11:42 independent56 joined #minetest-dev
12:07 kilbith https://www.youtube.com/watch?v=16M3PJu1zfw
12:07 kilbith Camera API (WIP)
12:07 kilbith this is a static camera, not the traditional CameraFPS with mouse events
12:24 kilbith I'm tempted to make a lower level API to draw on-screen, without going through the HUD and the GUI; it could be something like: camera:draw(x, y, w, h)
12:25 kilbith right on the screen, everything in pixels
12:25 kilbith opinions?
12:51 erle joined #minetest-dev
13:01 kilbith nobody cares?
13:07 proller joined #minetest-dev
13:08 Taoki joined #minetest-dev
13:16 Zughy[m] I do, the whole camera system is cool, but I don't have enough knowledge to answer the HUD thing
13:17 Zughy[m] the Lua implementation seems clean though
13:22 rubenwardy v-rob was woking on such a thing, as a client-side API
13:22 rubenwardy a server-side API would be too slow for real time things, it would end up just being a way to make dynamic images to send
13:25 kilbith uh no, there are no dynamic images involved
13:26 kilbith each camera is drawn inside their respective viewport
13:26 kilbith so it's realtime rendering of the scene
13:27 kilbith besides, there's already a client-side camera API which is half-baked
14:01 Fixer joined #minetest-dev
14:03 MTDiscord <MisterE> So this would be good for cut scenes? What other usecases?
14:08 Baytuch joined #minetest-dev
14:10 kilbith that: https://www.youtube.com/watch?v=yGZHYut1gik
14:11 kilbith camera:get_camera_texture() inside entity def
14:15 proller joined #minetest-dev
14:26 Baytuch joined #minetest-dev
14:32 independent56 joined #minetest-dev
15:01 proller joined #minetest-dev
15:25 independent56 joined #minetest-dev
15:28 rubenwardy reviewed and fixed #7865  could use another look
15:28 ShadowBot https://github.com/minetest/minetest/issues/7865 -- Add crosshair support for Android by srifqi
15:28 rubenwardy s/could use/needs/
15:38 MTDiscord <MisterE> kilbith: so it could allow overlays and holograms? Yeah, id say theres interest
15:40 MTDiscord <MisterE> In your code in the video, I don't see a reference to the player object. So how does it control just 1 player's camera?  Or, maybe I'm still not understanding what it does :|
15:41 kilbith > so it could allow overlays and holograms? Yeah, id say theres interest
15:42 kilbith no, it means that a scene can be rendered in real-time onto an entity
15:42 kilbith > So how does it control just 1 player's camera?
15:43 MTDiscord <MisterE> Oh! So, real time portals? Non-euclidean spaces? Mirrors? Star field end portals?
15:43 kilbith you can add a new camera object with Camera(), it's not tied to an ObjectRef
15:43 kilbith unless you attach() it
15:43 kilbith possibly mirrors, yes
15:43 rubenwardy Is a camera a type of Object? Would be good to reuse the movement interpolation code
15:43 rubenwardy and attachments
15:44 rubenwardy just not necessarily serialisation
15:44 MTDiscord <MisterE> kilbith: the blog isn't out yet. Do you want to write up a short paragraph about it? Or save it for another month?
15:45 independent56 joined #minetest-dev
15:45 kilbith moment
15:45 kilbith I finish a meeting
15:47 independent56 joined #minetest-dev
16:05 independent56 joined #minetest-dev
16:14 independent56 joined #minetest-dev
16:16 kilbith so
16:16 kilbith Camera is an independant object such as VoxelManip or ItemStack
16:17 kilbith you can create it at init time or whatever
16:18 kilbith you can create as much cameras as you want
16:19 kilbith MisterE[m]: what are your expectations for the blog?
16:19 kilbith a technical in-depth explanation? the use-cases?
16:20 MTDiscord <MisterE> Conversational tone, an intro for general consumption that describes progress. Semi formal
16:20 MTDiscord <MisterE> A screenshot would be great if appropriate
16:20 Pexin can each camera exist in a separate coordinate space? so they dont see the same geometry?
16:20 MTDiscord <MisterE> Make an issue against the blog
16:20 kilbith I'll make a video instead
16:21 MTDiscord <MisterE> That will be grent
16:21 MTDiscord <MisterE> *great
16:21 kilbith Pexin: examples?
16:21 Pexin lots. but simple ones?  ...uhh
16:22 kilbith I intend to make the Camera coords system 1:1 with world coords
16:22 Pexin well basically so you dont need to carce out a region
16:22 Pexin carve
16:24 MTDiscord <MisterE> How long do you think the blog submission would take you? More than two days we cant wait for but if its before that we can make that work. Ideally by the end of today though. Otherwise it can go into next month's post.
16:25 MTDiscord <MisterE> as for details, if you think that the average person interested in minetest development would be interested, you can put it in.
16:28 rubenwardy technically it's July already, so wouldn't be incorrect in next month's post :D
16:29 erle is it already merged?
16:29 erle i think teasing features should only happen if the users have some easy way to experience them (i.e. compile git HEAD)
16:30 kilbith MisterE[m]: can it be tomorrow?
16:30 MisterE[m] erle: No, the blog is for letting people know what devs are working on.
16:30 erle ah i see
16:30 erle i thought it was a “what's new” thing
16:30 MisterE[m] kilbith: We can try for that kilbith
16:30 MisterE[m] erle: That too erle
16:30 MisterE[m] It just has to be described as what it is
16:32 rubenwardy I'm hesistant to put super immature stuff there if it's still a long time away
16:33 erle hmm, hmmm, maybe i can get cora to write something about how to place structures very fast (the trick is to make the minetest engine place decorations for you that unpack into buildings later, but i think she figured out a bit more about what is slow and what is fast)
16:33 MisterE[m] erle: Ruben and I were discussing having special how-to articles. That would do nicely.
16:34 rubenwardy content marketing
16:34 erle well it is also help for using the engine. most structure placement code in mods is garbage in one way or the other.
16:34 MTDiscord <MisterE> Write it, submit it as a PR or an issue, and we will review and post.
16:35 erle sometimes the engine is at fault (it still can't place giant schematics IIRC), sometimes people make mapgen with a code flow that reminds me of spaghet
16:35 MTDiscord <luatic> erle: TBH that belongs on the minetest_doc project rather than a "blog"
16:35 erle i'd rather say it could go into the rubenwardy book maybe?
16:35 rubenwardy detailed modding tutorials like that owuld be better on minetest_doc, yeah
16:35 rubenwardy or my book
16:35 MTDiscord <MisterE> Depends how its written?
16:36 erle anyways, the blog surely could link to new book chapters
16:36 rubenwardy the blog should be fairly high level rather than technical info
16:36 rubenwardy ooh, could add the new book chapter to the blog post
16:36 rubenwardy https://rubenwardy.com/minetest_modding_book/en/quality/translations.html
16:37 kilbith rubenwardy: the PostProcessing thing by x2048 is a super immature as well, yet mentionned in the blog
16:37 kilbith -a
16:37 MTDiscord <MisterE> As is jordach's mapgen
16:37 rubenwardy true. That has a lot of political buy in
16:38 rubenwardy Jordach's map gen probably shouldn't have been in the blog post, would be lucky to ever see that :D
16:38 kilbith btw that feature will be financially backed by new company I work in
16:38 rubenwardy cameras or post processing?
16:38 kilbith thats why I have time to develop it
16:38 kilbith cameras
16:39 MTDiscord <luatic> rubenwardy: found a typo "player’s lang,uage"
16:39 rubenwardy thanks
16:39 MTDiscord <MisterE> Jordach's mapgen can be used today tho, if you build a server form source.
16:40 MTDiscord <luatic> MisterE: You could feature my revival of the bone override PR I guess
16:40 MTDiscord <luatic> #12388
16:40 MTDiscord <MisterE> I think, that as long as very WIP features are known to be so, they should go in the blog. Maybe with lower priority than more likely to be merged soon feature
16:40 ShadowBot https://github.com/minetest/minetest/issues/12388 -- Extend bone override capabilities by appgurueu
16:41 MTDiscord <luatic> Like most PRs, it is in ready for review limbo
16:41 rubenwardy sure
16:41 rubenwardy sure @MisterE
16:41 rubenwardy Well, the average PR open time is plumetting recently at least
16:41 rubenwardy *cough* #7865 *cough*
16:42 ShadowBot https://github.com/minetest/minetest/issues/7865 -- Add crosshair support for Android by srifqi
16:42 MisterE[m] Something of the form ”kilbith is working on a PR that implements ...”
16:42 rubenwardy The camera PR is on the roadmap at least, and is pretty awesome
16:42 MTDiscord <luatic> in terms of likeliness to be merged, #12353 has been discussed during the last meeting, but I don't think fixing physics makes for an exciting blog post ?
16:42 ShadowBot https://github.com/minetest/minetest/issues/12353 -- Fix acceleration by appgurueu
16:42 rubenwardy so as long as the tech works out it doesn't seem unlikely
16:43 MisterE[m] To point is to generate interest in development, and inform everyone of what others are working on, so they know to follow the development.
16:43 kilbith OpenMW is doing that
16:43 kilbith 0 A.D. as well
16:43 MisterE[m] Doing what?
16:44 kilbith blog marketing
16:44 MisterE[m] Yeah. I specifically got the idea from Veloren
16:44 erle i actually think the physics thing is a decent thing for a blog post
16:44 erle you should probably ask fleck if it started in physics class lol
16:45 MisterE[m] Well, people, you have 1.5 days or so to make submissions for this month's post
16:45 erle given he was still in school when he noticed it was wrong
16:45 MisterE[m] https://github.com/minetest/blog/issues
16:45 MisterE[m] submit them there
16:48 rubenwardy I've wanted a blog for years but didn't want to be the main person to do it, that's how it never gets updated regularly   *looks at the twitter*
16:49 rubenwardy Content marketting is a bit different than a blog with monthly/weekly updates. The idea behind content marketting is to appear more in search results
16:50 rubenwardy for example, we could make an article about how to get started with making a game in Minetest, and then it could become a top result when someone searches "how to make voxel game" or whatever
16:52 erle given the state of google i would not count on it
16:53 Pexin seo is infernal black magic
16:53 rubenwardy The monthly blog posts are great for communication with users though, it allows them to know we're doing stuff without crawling through github
16:53 independent56 joined #minetest-dev
16:53 erle i noticed that for some software i wrote its homepage is not even on the first page of search results – however, the first search results *link* to that page. so clearly, google is using something other than page rank and serving something else than useful results in this case.
16:53 erle that is true
16:54 kilbith there's need to focus on marketing this summer at least, after what happened to MC
16:54 erle what happened to MC?
16:55 erle i maintain that marketing in the sense of generating hype is bullshit, while marketing in the sense of providing useful information is not.
16:55 kilbith https://www.reddit.com/r/Minetest/comments/vjqxwz/with_microsofts_latest_decision_to_police_chat_on/
16:55 MisterE[m] they implemented global chat monitering, even on anarchy servers, not sure about singleplayer. They if you say something they dont like, they can permaban you
16:56 MisterE[m] spelling go brr
16:56 erle LOL
16:56 erle knowing microsoft, i bet if you are banned you can't even play on your private server or something
16:56 MisterE[m] Minecraft is now a "safe Space"
16:57 MisterE[m] e.g. no Freedom of speech
16:58 MisterE[m] erle: this is correct IIUC
16:58 ROllerozxa "i bet if you are banned you can't even play on your private server or something" yep, pretty much. no ability to play on multiplayer servers that rely on the mojang authentication servers
16:58 schwarzwald[m] May I have permission to triage at least some parts of the Coverity report? The bugs I've been PRing fixes to haven't been marked yet.
16:58 erle hilarious, but it has nothing to do with minetest, except for minetest being excactly unlike it
16:59 erle schwarzwald[m] why do you need permission?
17:01 schwarzwald[m] erle: Because I can't triage the report without permission.
17:03 schwarzwald[m] Maybe only admins can triage it? I would hope there's a non-admin role that can triage but not upload builds.
17:24 independent56 joined #minetest-dev
17:33 independent56 joined #minetest-dev
17:42 olliy joined #minetest-dev
18:09 independent56 joined #minetest-dev
18:17 independent56 joined #minetest-dev
18:31 independent56 joined #minetest-dev
18:52 independent56 joined #minetest-dev
19:17 proller joined #minetest-dev
19:19 panwolfram joined #minetest-dev
19:26 independent56 joined #minetest-dev
19:35 independent56 joined #minetest-dev
19:38 independent56 joined #minetest-dev
20:08 independent56 joined #minetest-dev
20:11 independent56 joined #minetest-dev
20:35 independent56 joined #minetest-dev
20:57 independent56 joined #minetest-dev
21:00 TheCoffeMaker joined #minetest-dev
21:05 independent56 joined #minetest-dev
21:20 behalebabo joined #minetest-dev
21:23 independent56 joined #minetest-dev
21:26 independent56 joined #minetest-dev
21:27 Zughy[m] Friendly reminder that if anyone wants to help with confirming bugs, there are still 16 of them that need confirmation
21:27 Zughy[m] https://github.com/minetest/minetest/issues?q=is%3Aissue+is%3Aopen+label%3A%22Unconfirmed+bug%22+-label%3A%22Action+%2F+change+needed%22+-label%3A%22Possible+close%22+-label%3AAndroid+-label%3AmacOS
21:35 erle #11829 is very easy to confirm. i have put the repro instructions in the issue.
21:35 ShadowBot https://github.com/minetest/minetest/issues/11829 -- Client hangs on startup if too many texture modifiers are used in a game or mod
21:37 erle #11532 is part of some parcour courses and player builds
21:37 ShadowBot https://github.com/minetest/minetest/issues/11532 -- Similar to sneak glitch, with different block configuration
21:38 erle as is #11171 btw
21:38 ShadowBot https://github.com/minetest/minetest/issues/11171 -- New move sneak glitches through blocks despite sneak_glitch = false
21:39 celeron55 it almost seems like we should add more glitches to the movement code, not remove them
21:40 erle well, in both cases, i definitely don't think it's worth it to remove them – the way you have to set up the nodes is very specific and i have only ever seen it made deliberately as part of a parcour or hidden base
21:41 erle i.e. “sneakjump in the corner in the basement to get to the secret room”
21:41 erle walking on the side of walls is stupid though
21:45 rubenwardy celeron55: that is what the code wants to do, after all
21:55 olliy1or joined #minetest-dev
22:05 MTDiscord <FatalError> Will we ever have SSCSM
22:05 MTDiscord <FatalError> I know this topic has been discussed since well... forever
22:09 independent56 joined #minetest-dev
22:09 MTDiscord <FatalError> But will it even happen?
22:10 MTDiscord <FatalError> We could just make it a completely benign system
22:10 MTDiscord <FatalError> Hell, you could make your own custom language for it
22:10 MTDiscord <FatalError> Lua imitation meat
22:14 MTDiscord <GoodClover> Making a new programming is an entire seperate beast. Not exactly an option anyone's want to take.
22:16 MTDiscord <FatalError> Well if we're concerned about security
22:16 MTDiscord <FatalError> And apparently just neutering lua is somehow "just not secure enough"
22:18 MTDiscord <FatalError> It really couldnt be that hard to gaurentee a secure enviorment
22:18 MTDiscord <FatalError> We have 90% of the framework involved in this, we could even use open source libraries or hell, even reintegrate hacked clients into the API
22:21 MTDiscord <FatalError> I mean its possible to clear the enviorment in lua, so why cant we just do that, and explicitly add safe functions
22:21 independent56 joined #minetest-dev
22:21 MTDiscord <FatalError> We can also do this within the client's API itself that way we can guarantee safety without risking hacks
22:28 independent56 joined #minetest-dev
22:34 panwolfram joined #minetest-dev
22:34 erle FatalError what do you want to use CSM for that would not be solved by “make a better CSM interface in the main menu and host them on CDB” ?
22:34 erle (i am collecting use cases)
22:36 independent56 joined #minetest-dev
22:36 Pexin < celeron55> it almost seems like we should add more glitches to the movement code, not remove them
22:36 Pexin don't you worry, I have plans
22:43 independent56 joined #minetest-dev
22:46 independent56 joined #minetest-dev
23:29 MTDiscord <paradust> this is one of the hardest problems. this is why it pretty much only happens in browsers, which are backed by big companies with deep pockets
23:32 MTDiscord <paradust> the attack surface is huge: Miscompiles (if LuaJIT is used), buffer overruns (in the API), hardware-level attacks (row hammer, etc), side-channel attacks (e.g. spectre), etc
23:33 MTDiscord <paradust> browsers have multiple levels of sand-boxing, and they are still regularly exploited
23:49 erle paradust and yet, there are parts of browsers that are not exploited regularly, the ones that use tested and proven strategies.
23:49 MTDiscord <FatalError> But you cant simply override the clientside environment, adding dangerous gunctions would quite literally be impossible with a clear environment, with only benine functions implemented
23:50 erle (which broadly means following LANGSEC principles, and TBL's ”principle of least power”)
23:51 erle FatalError i think it helps much more if you present a use-case. lukyx for example, wanted forms that react instantly, without lag. that is a very clear use case (and one where i would argue that browsers have shown you do not need a full-fledged scripting environment).
23:52 MTDiscord <FatalError> Set velocity and position without delay
23:53 MTDiscord <FatalError> Render objects only visible on clientside
23:53 erle lag-free velocity and position seems like a movement prediction issue, i fail to see where you need scripting there.
23:53 erle minetest already has predictive hints AFAIK
23:53 MTDiscord <FatalError> Its absolutely nessecary
23:53 erle elaborate please
23:53 MTDiscord <FatalError> Set_pos is dirt useless 90% of the time
23:54 MTDiscord <FatalError> Because lag will cause the player to snap back
23:54 MTDiscord <FatalError> Get_pos doesnt give accurate results
23:54 erle oh, i see. that seems like a different issue than what i thought.
23:55 erle how exactly are you using set_pos that lag makes it weird?
23:55 erle are you setting the position several times per second or so?
23:55 MTDiscord <FatalError> Delay breaks shit, it always does. Same issue with hud, if I want to update a hud waypoint without delay, i fucking cant
23:55 MTDiscord <FatalError> Itll look like shit
23:55 MTDiscord <FatalError> The use cases are basically endless
23:55 erle i have made the xmaps mod, so i am aware of the difficulties
23:56 erle FatalError do you have a test case for the set pos thing? are you *sure* it is not anticheat related?
23:57 MTDiscord <FatalError> Literally nothing to do with anticheat
23:57 erle i have had snap-backs, but since the whole physics thing i assume it is a difference in server and client prediction
23:57 MTDiscord <FatalError> Its not
23:57 erle like, going in a minecart in a straight line, getting snapped
23:57 MTDiscord <FatalError> Not what im talking about
23:58 erle okay, so how do you use set_pos?
23:58 MTDiscord <FatalError> I dont because itll break everything
23:58 erle okay how do you *want* to use it?
23:58 erle like what's the use case?
23:58 MTDiscord <Jonathon> sscsm is useful for nice clientside prediction of things or making interactions less laggy, etc. at the end of the day you should always trust/verify things on the server side
23:58 erle if you want to set the player position each frame, DO NOT DO THIS
23:59 MTDiscord <FatalError> OH WOW REALLY?
23:59 MTDiscord <FatalError> You mean I cant use setpos where I want to?
23:59 MTDiscord <FatalError> Because itll break?
23:59 MTDiscord <FatalError> Because of delay?
23:59 MTDiscord <FatalError> CRAZY ALMOST LIKE I WAS SAYING THAT
23:59 MTDiscord <FatalError> ANWAY
23:59 erle i am saying a different thing

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