Minetest logo

IRC log for #minetest-dev, 2020-10-04

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

All times shown according to UTC.

Time Nick Message
03:55 Foz joined #minetest-dev
07:14 nore concerning the spectre-style attacks, IIRC they are highly timing-sensitive attacks, so simply not having precise time() functions is enough to avoid the attack
07:15 nore (and by timing-sensitive, I mean something on the order of micro- or nano- seconds)
07:16 nore (what firefox did was reducing the precision of timers to 20µs)
07:34 mizux joined #minetest-dev
08:00 ShadowNinja joined #minetest-dev
08:32 arthur joined #minetest-dev
08:38 arthur joined #minetest-dev
08:45 arthur joined #minetest-dev
08:50 huichilobos joined #minetest-dev
08:58 calcul0n joined #minetest-dev
09:04 huichilobos joined #minetest-dev
09:28 MTDiscord4 joined #minetest-dev
09:33 ShadowNinja joined #minetest-dev
09:33 ShadowNinja joined #minetest-dev
09:33 ircSparky_ joined #minetest-dev
09:35 pgimeno nore: yes, and in JavaScript they didn't have a precise source of time, so they used a thread that increased a counter and used that. The JIT would allow this.
09:36 pgimeno Without JIT there's no way for that to happen; any counter loop is going to be too slow to measure cache misses.
10:09 Wuzzy joined #minetest-dev
10:11 Wuzzy hey rubenwardy, u there? i have seen ur msg but this link <https://github.com/orgs/minetest/teams/engine/discussions/25> is a 404 error
10:16 sfan5 yes discussion are not visible from outside
10:17 sfan5 Wuzzy: https://a.uguu.se/VswMoyPm7SY0_.png
10:18 Wuzzy oohh secret discussions... illuminati confirmed. ?
10:18 Wuzzy anyway, are there any other people proposed for issue handling?
10:19 Wuzzy I am happy to volunteer to help you with the issue cleanup and whatother work needs to be done here. is there a ruleset for this?
10:31 sfan5 no idea
10:35 Wuzzy wtf? there is no system behid closing/opening issues, and labels? O_o
10:35 Wuzzy that cant be right
10:41 Fixer joined #minetest-dev
10:42 sfan5 that's not it
10:42 sfan5 I just don't remember if or where it would be written down
10:44 pgimeno the ruleset for non-core devs might be different than that for core devs; the idea is new so maybe it hasn't been decided yet
11:16 Krock http://irc.minetest.net/minetest-dev/2020-10-03#i_5738423
11:17 Krock rule suggestions are welcome, if this turns out to work well
11:29 Sharpik joined #minetest-dev
11:38 Krock rubenwardy: can "getPackagesFromURL" be removed? it seems to be handled in Lua now
11:39 Krock "getDownloadURL" too
11:48 rubenwardy I don't recall those functions
11:48 rubenwardy Oh, in content
11:48 rubenwardy Probably
12:12 turtleman joined #minetest-dev
12:28 thomas_irc7 joined #minetest-dev
12:38 Krock -> #10463
12:38 ShadowBot https://github.com/minetest/minetest/issues/10463 -- Remove unused functions reported by cppcheck by SmallJoker
12:40 Krock will merge #10405, #10360 and #9079 in 15 minutes
12:40 ShadowBot https://github.com/minetest/minetest/issues/10405 -- Fix short 180 degree rotation when using set_bone_position by appgurueu
12:40 ShadowBot https://github.com/minetest/minetest/issues/10360 -- Fix/Add First Person Attachments by Jordach
12:40 ShadowBot https://github.com/minetest/minetest/issues/9079 -- Minimap in HUD, minimap showing texure, control minimap modes from Lua by EvidenceBKidscode
12:42 Krock the third will take a moment to rebase but it'll be fine
13:08 Krock merging
13:19 Krock testing "rebase"d build
13:25 Krock done
15:05 rubenwardy Krock, sfan5, Wuzzy: https://dev.minetest.net/Git_Guidelines#Issue_and_Pull-Request_Management
15:06 rubenwardy not sure what specific rules there should be
15:06 Wuzzy "If a pull request or an issue does not get a response from its author within one month (when requiring more details), it is closed"
15:06 Wuzzy ooff that would be a lot of PRs getting killed
15:07 rubenwardy well, presumably you'd go to [Possible Close] first
15:07 rubenwardy what length would you reckon?
15:07 Wuzzy idk
15:07 Wuzzy oldest open pr = Apr 30 2017!!!!
15:07 Krock PRs can be left open for longer. it's only stating the lower limit
15:07 Krock mainly those which need adoption
15:08 rubenwardy yeah, and that also only applies to authors, not us
15:08 rubenwardy as in - a PR shouldn't be closed if it's not reviewed
15:08 Wuzzy yeah, that would be very very rude
15:08 rubenwardy I do feel like it's unfair to neglect a PR, then a year later ask for changes then close it. But there's not too much alternative, hopefully the waiting time is lowered
15:09 Wuzzy but yeah when a PR is basically dead (reaction from nobody, tons of conflicts, different codebase), well, it is already dead, closing it would be only a formality
15:10 jas_ joined #minetest-dev
15:10 Wuzzy ironically, the oldest open PR also has 1 approval ? https://github.com/minetest/minetest/pull/5680
15:10 rubenwardy I'm about to review it
15:10 rubenwardy I can't maintain this pace though, this weekend will be a one-off
15:11 rubenwardy I would like to start spending consistent amounts of time reviewing
15:11 rubenwardy rather than random sprints
15:11 Wuzzy yay
15:22 Wuzzy rubenwardy: how's your work on mainmenu going? can you maybe explain how far you've come, or what the general idea of your work is?
15:22 rubenwardy I'm blocked by design
15:23 rubenwardy I have prototypes with customisable layouting (ie: allowing you to change whether the menu is a tabset / vertical tabs / or classic main menu)
15:23 rubenwardy and also with a seperate game menu (ie: first thing you see, replacement to game bar)
15:24 rubenwardy My current design idea is to use a task bar analogy - with the start menu throwing up the game menu, and each icon on the task bar being a game. A more understandable version of the game bar
15:24 rubenwardy Needs validation, really
15:24 rubenwardy For example, should a game-centric design even be done?
15:26 rubenwardy
15:26 rubenwardy 1. Is there a version of hash_node_position in C++?
15:27 rubenwardy 2. What's the minimum size that a const ref should be used? My IDE is suggesting `for (const v3s32 &pos : `
15:30 Wuzzy what is this prototype thing about? I dont think i understood it
15:32 rubenwardy I've done a mock up and a prototype. The mock up is entirely fake, it's basically like Libreoffice Draw.  The prototype is working code which is not review quality
15:33 Wuzzy ah, so it's basically a throwaway prototype?
15:33 rubenwardy it can be salvaged
15:33 Wuzzy ok, i thought there was already significant work on code done...
15:33 rubenwardy therre has been, it's just all mixed together and would need splitting out and refining
15:34 Wuzzy hrm
15:34 Wuzzy whats the point of this layouting thing anyway?
15:34 Wuzzy sounds complicated
15:34 rubenwardy Layouting?
15:34 Wuzzy rubenwardy> I have prototypes with customisable layouting (ie: allowing you to change whether the menu is a tabset / vertical tabs / or classic main menu)
15:34 rubenwardy ah right
15:35 rubenwardy well, it's not actually that complicate - it's just having a separate table for building the formspec
15:35 Wuzzy i mean, i dont see the usecase here
15:35 Wuzzy oh, interesting
15:35 rubenwardy it allows the mainmenu to be experiemented with without breaking existing things
15:36 Wuzzy ah you mean you basically have the same buttons/features/whatever in the menu, just arranged differently?
15:36 rubenwardy yeah
15:36 Wuzzy ok i see how this can simplify things
15:36 Wuzzy did you throw away all existing main menu code or did you build on it?
15:36 rubenwardy build on it
15:36 Wuzzy hrm. i believe anything except a full restart is doomed. but i might be wrong
15:37 Wuzzy plus, the controls settings menu is still hardcodded i believe (no lua)
15:37 Wuzzy ok thinking about it, this modular system actually seems smart. it would save so much time on dev when fooling around ?
15:38 DS-minetest joined #minetest-dev
15:43 rubenwardy gosh, it's so hard to spell neighbour without the u
15:43 Wuzzy lol
15:44 Wuzzy so if u understood u correctly, u want to do it in 2 steps?
15:44 Wuzzy step 1 is to build the modular system
15:44 Wuzzy and step 2 is then to fool around and test designs and find the "right" one?
15:49 MTDiscord <L​one_Wolf> What you're doing atm sounds good ruben, in the case of a game-centric design
15:50 rubenwardy I'm not sure the order of which to do it
15:51 rubenwardy for the current design, perhaps PR 1 would be changing the main menu layout to be tab based, using the modular system
15:51 rubenwardy the modular system is only a few hundred lines, it's adding a member to the tabview and calling methods on the member
15:51 rubenwardy PR 2 would be replacing the game bar, and making the main menu game-centric
15:51 rubenwardy idk
16:07 MTDiscord <L​one_Wolf> Pr 1 first sounds good
16:40 thomas_irc7 joined #minetest-dev
17:21 sfan5 <rubenwardy> 2. What's the minimum size that a const ref should be used? My IDE is suggesting `for (const v3s32 &pos : `
17:22 sfan5 in the clang-tidy configuration I have excluded all irrlicht vector types from the "turn into reference suggestion"
17:23 Krock constness is only an optimization for larger data types
17:24 Krock 3x 4 bytes won't make a difference there
17:38 sfan5 Wuzzy: what do you mean by "There are no speed optimizations (to my knowledge) for simple entities that never move"?
17:38 Wuzzy wait, there are?
17:38 Wuzzy ok
17:38 Wuzzy so...
17:39 sfan5 I am asking which optimizations you think are missing
17:39 Wuzzy if an entity A moves all the time, and entity B never moves, which entity eats more CPU?
17:40 sfan5 assuming they are physical that still depends on what their velocity and acceleration
17:40 Wuzzy entity B has velocity and acceleration 0
17:41 Wuzzy lets further assume both entities are non-physical
17:41 Wuzzy i want to know if the act of moving eats more CPU
17:41 sfan5 if they are not physical it's irrelevant if the entity moves or not
17:41 Wuzzy ok...
17:42 sfan5 if physical, |vel|=0 |acc|=0 will consume no resources for collision detection
17:42 Wuzzy nice
17:42 Wuzzy so velocity and acceleraion always eat same CPUamount, no matter hte value (if non-physical)?
17:43 Krock it consumes resources for other entities that may collide with it
17:43 Wuzzy i said its non-physical
17:43 Wuzzy so there should be no collision
17:44 sfan5 yes that's the case (why wouldn't it?)
17:44 Wuzzy ah ok
17:44 Wuzzy i thought there would be a possible optimization by skipping speed calculations for "static" (nonmoving) entities
17:44 Wuzzy looks like i have no idea what i am talking about
17:44 Wuzzy ok
17:44 Wuzzy I have no suggestions for optimizations then
17:45 sfan5 there are possible suggestions when you have an entity affected by gravity
17:45 sfan5 because collision detection will run every step
17:45 Wuzzy but i notice that minetest really doesnt like it if there is a large amount entity
17:45 Wuzzy i also like sorcerykid's entity PR
17:45 Wuzzy with all those movement functions
17:46 Wuzzy will be very very helpful for mods
17:48 sfan5 yes I think it's good too
17:48 sfan5 (conceptually)
18:18 Fixer_ joined #minetest-dev
18:30 search_social joined #minetest-dev
18:43 MTDiscord <j​osiah_wi> https://github.com/minetest/minetest/pull/10464
18:46 sfan5 if you fill in the PR description it'll have good chances of being merged
18:53 MTDiscord <j​osiah_wi> Oh I must have missed that. I'll fix it right away.
18:53 MTDiscord <j​osiah_wi> Um, what "description"?
18:54 MTDiscord <j​osiah_wi> Looks like I answered all the questions about the PR on the description box.
18:56 sfan5 I'd remove the "Add compact, short information about your PR for easier understanding:" and the unused todo list
18:58 MTDiscord <j​osiah_wi> Updated. :+1:
18:59 MTDiscord <L​one_Wolf> I just delete all that and format it myself lol
19:04 MTDiscord <j​osiah_wi> The template is very handy imo, and it's there for a reason.
19:06 MTDiscord <j​osiah_wi> Also would Minetest consider participating in this event? https://hacktoberfest.digitalocean.com/
19:15 rubenwardy The contributions tend to be too low quality
19:15 MTDiscord <a​ppguru> Yeah, someone posted an article concerning that in the discord
19:16 rubenwardy I got a hacktoberfest t-shirt two years ago from minetest repos
19:16 rubenwardy I use it as pjamas
19:16 rubenwardy I mean, the worst that happens is a few PRs are marked as invalid
19:16 MTDiscord <L​one_Wolf> The coloring doesn't look very nice
19:34 TC01 joined #minetest-dev
20:18 thomas_irc7 joined #minetest-dev
20:29 thomas_irc7 joined #minetest-dev
21:58 nore I added support for plurals inside my translations PR, but it seems like ngettext doesn't work everywhere... does someone know what the problem might be?
21:59 nore (I'll go to bed now, so I'll look at the answers tomorrow, night all)
23:06 TC01 joined #minetest-dev

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