Minetest logo

IRC log for #minetest-dev, 2018-02-05

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

All times shown according to UTC.

Time Nick Message
00:37 turtleman joined #minetest-dev
00:51 peetcamron left #minetest-dev
01:05 paramat game#2027 now ready for review
01:05 ShadowBot https://github.com/minetest​/minetest_game/issues/2027 -- Add marram grass for coastal sand dunes by paramat
04:08 Lone-Star joined #minetest-dev
05:01 paramat #7017
05:01 ShadowBot https://github.com/minetest/minetest/issues/7017 -- Lua_api.txt: Document 'wielditem' visual in object properties by paramat
05:04 paramat will merge #6999 and 7017 in 5 mins
05:04 ShadowBot https://github.com/minetest/minetest/issues/6999 -- Item entity: Prevent motion in ignore nodes by paramat
05:12 rud0lf joined #minetest-dev
05:13 paramat merged
05:13 rud0lf left #minetest-dev
05:13 Sokomine joined #minetest-dev
07:37 Darcidride joined #minetest-dev
08:00 QwertyCool joined #minetest-dev
10:24 kilbith joined #minetest-dev
11:13 Fixer joined #minetest-dev
11:23 Beton joined #minetest-dev
12:22 YuGiOhJCJ joined #minetest-dev
12:39 Icedream joined #minetest-dev
13:44 ThomasMonroe joined #minetest-dev
14:09 ThomasMonroe joined #minetest-dev
14:16 nerzhul merging #7013
14:16 ShadowBot https://github.com/minetest/minetest/issues/7013 -- Add minetest.is_player by HybridDog
14:26 antims joined #minetest-dev
14:26 antims joined #minetest-dev
14:49 VanessaE find_nodes_in_area(): area volume exceeds allowed value of 551368   <--- someone remind me...why does such a limit exist, again?
14:50 VanessaE and why is this limit and the reasoning for it not documented in lua_api.txt?
14:54 VanessaE the particular use I have in mind would only return a table with a few hundred entries at most (more likely a couple dozen in practice)
14:57 celeron55 https://github.com/minetest/minetest/pull/5277
14:57 VanessaE yeah, read that
14:57 celeron55 limiting the area seems to be a rather silly way to limit the output size
14:57 VanessaE but...um, 551k ain't the size of a u32
14:57 celeron55 but that's what it's doing
14:59 VanessaE in any case, I'm trying to figure the best way to resolve a crash in moretrees.  it uses ^ that to search for date palm trunks
14:59 VanessaE (hence the dozens to hundreds max)
15:00 VanessaE (I didn't write the date palm code, so I don't know what the search size is offhand)
15:04 VanessaE if it were up to me, I'd have it limit the size of the returned table by omitting random entries
15:05 VanessaE (as opposed to crashing, or returning only the first or last part of the table)
15:46 FourFire left #minetest-dev
16:34 paramat joined #minetest-dev
16:35 paramat VanessaE yes perhaps that limit is not a great idea. it's set to a default mapchunk plus 1 node of padding
16:36 paramat will add to docs request
16:37 paramat could you add an issue about rethinking that limit?
16:37 VanessaE if you come up with a better solution, I'd appreciate it if it gets backported to 0.4.x.
16:37 VanessaE I'll let you make an issue.
16:40 paramat after all, in some situations a larger search may be ok, as long as we document that large volumes can be intensive
16:41 paramat you should make the limit issue :]
16:42 VanessaE I'm kinda busy, and honestly I'm a little down on making issues these days (any project, not just MT)
16:45 paramat i'll support greatly raising or removing the limit
16:45 paramat ok
16:45 paramat will attend
16:47 Jordach joined #minetest-dev
17:30 paramat #7021
17:30 ShadowBot https://github.com/minetest/minetest/issues/7021 -- Remove or greatly raise volume limits on 'find nodes in area (under air)'
17:47 paramat #6477 is updated, simple PR, will review
17:47 ShadowBot https://github.com/minetest/minetest/issues/6477 -- Repositioning formspec item in delete world dialog by srifqi
17:48 paramat +1
17:51 cx384 joined #minetest-dev
17:59 VanessaE left #minetest-dev
18:00 VanessaE joined #minetest-dev
19:03 paramat #7009 for close?
19:03 ShadowBot https://github.com/minetest/minetest/issues/7009 -- Guest mode
19:07 Foz joined #minetest-dev
19:10 H4mlet joined #minetest-dev
19:11 fwhcat joined #minetest-dev
19:12 Foz joined #minetest-dev
19:17 Wuzzy joined #minetest-dev
19:17 Wuzzy https://github.com/minetest/minetest/pull/7019
19:33 sofar looks great imho
19:33 VanessaE +1
19:33 VanessaE (finally a Wuzzy proposal I agree with ;) )
19:34 Wuzzy s/proposal/PR/
19:35 sofar it's a pretty big pr, which is surprising
19:38 QwertyCool joined #minetest-dev
19:38 QwertyCool hey everyone
20:05 Gael-de-Sailly joined #minetest-dev
20:17 numzero joined #minetest-dev
21:02 Fixer joined #minetest-dev
21:34 Foz joined #minetest-dev
21:36 Beton joined #minetest-dev
21:39 rubenwardy sofar: talking about PRs...
21:40 sofar lol
21:40 sofar I haven't paid much attention recently, I know there's a few I should look at
21:52 Wuzzy I'm not happy with the practice of closing almost-but-not-quite-finished PRs. this could seriously harm contributor morale …
21:53 rubenwardy Which in particular?
21:53 Wuzzy unless a PR is very old and very hard to integrate, maybe
21:53 Wuzzy none in particular. its just a general observation. its like ... no activity for 1 month and you're dead
21:53 rubenwardy The high number of open PRs is probably putting of new contributors
21:54 Wuzzy yes, this, too
21:54 Wuzzy woah
21:54 Wuzzy 30 PRs are open but ONLY "Rebase Needed"
21:54 rubenwardy Part of me wishes I could work on Mt full time, so I could adopt more stale PRs
21:54 Wuzzy this is frustrating
21:55 rubenwardy Yeah, rebase needed is silly as you don't need it for reviewing
21:55 rubenwardy Aal
21:55 Wuzzy woah the oldest rebase needed from 2015… *sobs*
21:57 Wuzzy well if you want to work on MT fulltime why dont you do it?
21:57 rubenwardy Money and carrer
21:58 rubenwardy Uni takes up most of my time currently
21:58 rubenwardy And during the summer I'm working to build experience for a graduate job
21:58 paramat joined #minetest-dev
21:58 Wuzzy if you had the free choice, would you go full MT? Or do you think other things are more interesting to you?
21:59 sfan5 working on MT full-time sounds pretty terrible to be honest
21:59 Wuzzy hhahahhahahahaha
22:00 Wuzzy why? at least as core dev you don't need to go through the tiresome PR process, right?
22:00 Wuzzy or does it apply to everyone?
22:00 sfan5 partially
22:00 celeron55 as a core dev working full time on MT, you'd be forced to run the entire PR process
22:00 celeron55 :D
22:00 rubenwardy If money wasn't an issue for the rest of my life, I probably would
22:01 sfan5 often enough it's still required to bug someone else to give your pr a quick look and approve it
22:01 sfan5 but you don't need two (other) people to look at it
22:01 rubenwardy Idk actually, I don't find mt development fun
22:01 rubenwardy But I like the project
22:02 Wuzzy ouch. i can see why people would hate to do that. the PR process is really time-consuming and a drain on morale. :(
22:02 Wuzzy especially when discussions bog down on whether to put a comma after the 7th word in the 334th line XD
22:03 celeron55 in the end, i would probably do it
22:03 rubenwardy Tbh, you seem to choice issues which are mostl
22:03 celeron55 but very, very ironically
22:03 rubenwardy Likely to discend into bikeshedding
22:04 sfan5 which color would an official Minetest bike shed have though?
22:04 Wuzzy gray
22:05 sfan5 like Irrlicht's GUI?
22:05 Wuzzy because half of the HUD is gray :D
22:05 Wuzzy like my soul
22:05 celeron55 the thing about money is, though, that it always comes from somewhere, and always there will be someone having expectations in that end
22:06 Wuzzy we still haven't found a solution to combat bikeshedding, really :( its not just my issues that degrade into this
22:06 celeron55 gray is the most colorful color; it has all the colors!
22:07 celeron55 and equal amounts at that
22:07 nerzhul rubenwardy, make MT have strong Unittests is some funs and permits to catch anoying bugs sometimes :p
22:08 nerzhul when i get some useless time to code in C++ (these days i'm fully coding in golang) i like to add UT to mt
22:09 rubenwardy Making tests is boring and basically rewardless
22:11 Wuzzy boring yes, but rewardless, nope
22:11 sfan5 rewardless yes, but it does have a purpose and helps the project
22:13 Wuzzy how is a tested codebase not a reward?
22:14 Wuzzy but mt is pretty stable already, why do i complain? :D
22:14 rubenwardy There's no thanks
22:15 celeron55 it would be nice to have an interactive changelog where people could thumbs up stuff
22:17 celeron55 i don't know how many users would thumbs up unit tests, but maybe someone
22:20 celeron55 nobody bothers to thank anyone directly these days, but people like to click reaction buttons like there's no tomorrow
22:20 nerzhul Wuzzy but MT has bugs, like every software
22:21 Wuzzy interactive changelog? we call this "forum"... xD
22:21 celeron55 well, i wish the forum actually supported this
22:21 celeron55 maybe in 10 years i'll have a MT-specific platform which supports all these things i want
22:21 celeron55 an*
22:22 Wuzzy voting on changes seems a bit strange. under-the-hood changes are doomed to be unpopular even if they are very important as not everyone will understand them
22:22 celeron55 not voting
22:23 Wuzzy well ok not really voting, thumbs upping. anyway its the same here, too
22:23 Wuzzy it might give *some* insight but its no silver bullet
22:24 celeron55 maybe we should write to unit test messages who made the test, so you can give a shout to them when it catches your new bug you were about to commit
22:24 celeron55 8)
22:33 book` joined #minetest-dev
22:36 QwertyCool is anyone here anywhere close to taking on reworking GUI/moving away from formspec? I'd really like to contribute but not sure where to start.
22:37 sfan5 i don't think that is beyond the idea stage
22:37 sfan5 moving away from the mess that are formspecs is a quite huge thing
22:39 sofar everyone would applaud it
22:39 QwertyCool the reason i ask is because initially i wanted to make scaling better for 4k screens (then scaling in general) but i looked at the dicsussions around scaling and formspec and realised that there's likely to be an overhaul so it's discouraging me from working on adding to formspec's convolution
22:39 sofar I think you can probably implement it side-by-side with the old api
22:40 QwertyCool but im preety novice to c++ so just understanding the 3000+ lines that is formspec let alone recoding it seems pretty daunting by myself
22:41 sofar one point of notice is that if you do redo the formspecs, you probably want to redo the HUD, and of course all the mainmenu stuff as well
22:42 QwertyCool actually the first change that i realised that would need to be made would be updating settings changes without restarting the client, ie, font_size because thats the part that doesnt yet scale
22:45 QwertyCool sofar, if written properly do you think far less code would be needed? or about the same? or more?
22:46 sfan5 more
22:46 QwertyCool but seperated ?
22:46 sofar it really needs to be an improvement
22:46 sofar so, it would be more code
22:46 QwertyCool i thought we wanted to move away from big files
22:47 sofar split it into a hundred tiny files, I could care less
22:47 sofar it will be more code, most likely, in the end :)
22:47 sfan5 microservices
22:50 sofar I'm one of those developers that doesn't mind files with 5kloc
22:50 sofar etags is for whimps
23:04 celeron55 QwertyCool: "there's likely to be an overhaul" has been the status for years
23:05 celeron55 feel free to improve a bit, your work won't be lost for a long time
23:05 rubenwardy It won't be in 0.5 at leaat
23:07 VanessaE celeron55: so it'll be lost, it'll just take a while? :P
23:08 celeron55 literally everything will be lost
23:08 celeron55 just wait enough!
23:09 celeron55 usually, in this world, it's a matter of years before a thing basically disappears; if you make a really lasting thing, it'll take tens of years
23:09 celeron55 in that scale, a formspec improvement is a good investment
23:09 celeron55 not the best, but a fair one certainly
23:09 VanessaE celeron55: here.  *hands a mop and sponge*... you didn't just kill the joke, you liquified it :P
23:10 * celeron55 mops VanessaE out of the door
23:25 Foz joined #minetest-dev
23:26 QwertyCool i thought overhaulw as on the roadmap for 0.5 though
23:27 celeron55 i guess, 0.5 is still up for it, if it comes to it
23:27 celeron55 -,
23:28 celeron55 but nothing is happening in that area
23:28 celeron55 more of a prediction than a plan
23:29 QwertyCool ah.
23:29 QwertyCool maybe you could answer this for me, if formspec is an api of sorts, what files are calling upon it to create the pre-game settings menu, and the pause menu
23:32 celeron55 off the top of my head, no idea
23:33 celeron55 builtin/mainmenu :P
23:34 paramat 0.5 is too close to include a formspec overhaul, we want to release it in say 2-3 months and already have lots to do before release
23:35 celeron55 the pause menu is still not in Lua afaik
23:35 celeron55 ...or is it?
23:35 paramat redoing formspecs won't break client-server compat (i assume) so can be done later
23:36 paramat but starting work now is fine of course
23:36 sfan5 good luck with such a degree of backwards compat
23:36 celeron55 yeah it's in game.cpp, void Game::showPauseMenu()
23:36 sfan5 it's already lots of work to make sure mods won't break
23:39 paramat yes 0.5 is not breaking mods generally
23:39 pgimeno_ what is wrong about formspecs? the API? the textual format? the implementation?
23:41 celeron55 i'm fine with them; really they just need a proper rewrite with positioning, scaling and styling fixed
23:41 celeron55 some issues come from irrlicht's input and UI system also
23:41 rubenwardy They don't adapt to different screen sizes well
23:42 celeron55 so to answer the question: the implementation
23:42 celeron55 essentially, if there was a perfect implementation, nobody would bother to complain
23:43 celeron55 well ok, probably would; but then you'd be in some really difficult toolkit features that are really just utopia
23:44 celeron55 (my dream is a CSM-run pure Lua toolkit like i've said, but that's probably just dreaming also)
23:46 pgimeno_ well, love2d is programmed in Lua exclusively, and there are GUI libs for it, so it's not off the table
23:47 pgimeno_ I happen to be the maintainer of one such lib, after the original author went AWOL
23:48 celeron55 i'm interested in any proof-of-concepts
23:49 pgimeno_ well, with apologies for the off-topic, you need love2d to run it, but here it is: https://github.com/pgimeno/Gspot
23:49 celeron55 (i can say i certainly didn't expect a maintainer of a lua gui library to just pop in here... lol)
23:49 pgimeno_ it comes with a demo
23:50 celeron55 it assumes the entirety of love2d's API, right?
23:50 pgimeno_ yeah
23:52 celeron55 some kind of a defined subset version would probably be needed and then a CSM API and a wrapper+fill-in library in between
23:53 pgimeno_ let me say that particular lib isn't very complete, it lacks multi-line textboxes for example
23:55 celeron55 do you know of one that would support the things people currently do with formspecs?
23:57 pgimeno_ hm, luigi may be close, https://github.com/airstruck/luigi
23:57 pgimeno_ I'm not aware of any such lib supporting tables, but I haven't studied them in depth
23:57 paramat for the mediumm term i would prefer to keep formspecs and just try to improve them, instead of a complete replacement and the work and disruption. but this opinion is partly due to lack of dev time
23:57 pgimeno_ tables as in HTML-like tables
23:58 paramat i know formspecs are not great, but i fear the process of replacement more. 'try to avoid big rewrites'
23:58 celeron55 luigi looks to have some extra backend support
23:58 ThomasMonroe joined #minetest-dev

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