Minetest logo

IRC log for #minetest-hub, 2021-03-30

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

All times shown according to UTC.

Time Nick Message
00:34 don_flymoor joined #minetest-hub
03:11 VanessaE anyone awake?  I really need someone else to take a look at my biome_lib code
03:12 VanessaE not just to help me find the cause of mapblocks occasionally (and thankfully, rarely) ending up unpopulated by plants, but also to test the non-minetest_game option (clone, apply this patch over master, see the bottom of API.txt, https://gitlab.com/VanessaE/biome_lib/-/snippets/2097514)
03:14 VanessaE also I wouldn't mind if someone could compare generating speed at 353ca0cb or older, against how it is now at HEAD (for me, HEAD definitely feels faster)
03:15 VanessaE second opinions and all that.
03:37 MTDiscord <Ronoaldo> Hi
03:37 MTDiscord <Ronoaldo> you need help debugging is that it?
03:42 VanessaE yeah
03:43 VanessaE the "unpopulated" part is debugging, the others are just that I need someone who can test things and give me an objective opinion
03:44 VanessaE (actually, I suppose I'm really asking for a SUBjective opinion on the speed question ;) )
03:45 MTDiscord <Ronoaldo> (ok, the unpopulated happens with the patch applied or without it?)
03:45 VanessaE the patch I mentioned is unrelated
03:45 VanessaE this is three separate, unrelated questions/issues
03:45 MTDiscord <Ronoaldo> ok, sorry now I'm lost \o/
03:46 VanessaE heh
03:46 MTDiscord <Ronoaldo> let me see if I got 1) check from HEAD if I can reproduce some unpopulated blocks 2) check if head is faster than 353ca0cb and 3) peer review and test the patch --- I got them right?
03:48 VanessaE if you're inclined to help, try this: clone plantlife and biome_lib (https://gitlab.com/VanessaE/plantlife_modpack and https://gitlab.com/VanessaE/biome_lib). Install.  Create a new world (minetest_game, preferably with the flat mapgen), and enable those two mods.  Test, observe how fast it is.  Now roll biome_lib back to commit 353ca0cb, delete your map.sqlite, and go again
03:48 MTDiscord <Ronoaldo> Perfect! I'll give it a got
03:48 MTDiscord <Ronoaldo> *go
03:49 VanessaE (to me 353ca0cb is much slower than HEAD, which is what I'm hoping is the case for others)
04:00 MTDiscord joined #minetest-hub
04:12 MTDiscord <Ronoaldo> I'm not very good at testing... I used my notebook in single player and it kind of felt the same to me 😦
04:12 VanessaE huh
04:13 MTDiscord <Ronoaldo> My test mode was not ideal tho, I randomly walked with fast mode
04:13 VanessaE well as long as 353ca0cb definitely isn't faster :)
04:13 VanessaE the difference might not be too apparent in singleplayer I guess
04:14 MTDiscord <Ronoaldo> Do you have some map or seed that could trigger the most of biome generation?
04:14 VanessaE the gaol is to reduce apparent lag as everyone else sees it, if and when some chowderhead goes off into neverland and starts thrashing the mapgen to make new terrain
04:15 VanessaE well I used the flat mapgen, which happens to make lots of grass
04:15 VanessaE seed...um
04:15 VanessaE *checks*
04:15 MTDiscord <Ronoaldo> I used 1 with flat mapgen, and behave like the crowderhead but as I as also the other player I probably just saw myself lagging myself
04:16 MTDiscord <Ronoaldo> I'm not sure what I typed actually makes sense ¬¬
04:16 VanessaE 5791653907587404959
04:16 VanessaE there we go
04:16 MTDiscord <Ronoaldo> cool
04:16 MTDiscord <Ronoaldo> I'll reset things, testing the 353 first this time
04:16 VanessaE it makes sense :)
04:17 VanessaE except it's cHowderhead, not cRowderhead ;)
04:17 MTDiscord <Ronoaldo> lol thanks!
04:18 MTDiscord <Ronoaldo> I was feeling crowded
04:18 VanessaE heh
04:23 MTDiscord <Ronoaldo> so I can't paste the seed ...
04:23 VanessaE split into 4 digit groups so you can type it in:  5791 6539 0758 7404 959
04:23 VanessaE (well 4's and a 3)
04:33 MTDiscord <Ronoaldo> ok, now I could see that HEAD was indeed faster
04:33 VanessaE oh good
04:33 MTDiscord <Ronoaldo> on shutdown, from master, I noticed a higher time, and picked this message:   2021-03-30 01:31:59: INFO[Main]: Executing shutdown hooks [biome_lib] Stand by, playing out the rest of the aircheck mapblock log (there are 10175 entries)...
04:33 VanessaE that's pretty normal
04:34 MTDiscord <Ronoaldo> cool
04:34 VanessaE 1 chunk is 125 entries
04:34 VanessaE so that ^ would be around 9 or 10 entries as reported by 353ca0cb or earlier
04:35 MTDiscord <Ronoaldo>  grep 'Stand by' *.log -A 1 353.log:[biome_lib] Stand by, playing out the rest of the aircheck mapblock log 353.log-(there are 12 entries)... -- HEAD.log:[biome_lib] Stand by, playing out the rest of the aircheck mapblock log HEAD.log-(there are 10175 entries)...
04:35 VanessaE it shouldn't take more than a few tens of seconds on a slow machine
04:35 MTDiscord <Ronoaldo> I have the verbose log of both runs, if they help somehow?
04:35 VanessaE nah, don't need it
04:36 VanessaE I didn't add any debug statements related to the faster code
04:36 MTDiscord <Ronoaldo> it ware a few tens of sec. indeed... I'm not sure my machine is "fast", but it's not slow (Ryzen 7 3700U with 16G RAM)
04:37 VanessaE I figured in normal use, even if someone goes off and tortures the mapgen, they won't last very long before they either get bored and sign off, or they find an area to play in, either way the mapgen will probably settle back down to idle before the next regular shutdown
04:38 VanessaE so that hook usually won't have anything to do
04:38 MTDiscord <Ronoaldo> makes sense
04:38 MTDiscord <Ronoaldo> oh, I see... the second time I waited less time before shutting down, that is why it took more time right?
04:39 VanessaE man I can't math
04:39 MTDiscord <Ronoaldo> neither do I...
04:39 VanessaE 10000 odd entries now is 80ish under the old code
04:39 VanessaE that's a LOT
04:40 MTDiscord <Ronoaldo> I am considering to write a small script to test before-after and record
04:40 MTDiscord <Ronoaldo> would that be useful?
04:40 VanessaE during my tests I've never managed to get more than 50 or so "old" entries stacked up (typically about half that)
04:40 VanessaE do you just have a really fast flight configured or something? :)
04:40 MTDiscord <Ronoaldo> maybe?
04:41 VanessaE my machine must be slower than I thought then :)
04:41 MTDiscord <Ronoaldo> I do have some waaaay to far view distance (or something) in my conf
04:41 VanessaE Phenom II X6 1055T, 2.8 GHz (it's 6 cores but of course MT can only use one when Lua is involved)
04:41 MTDiscord <Ronoaldo>   doubletap_jump = false aux1_descends = false menu_last_game = minetest autojump = true screen_w = 1291 screen_h = 945 remote_port = 30000 name = pedro address = mercurio.ronoaldo.net fast_move = true keymap_cmd_local = KEY_PERIOD maintab_LAST = local enable_fog = true noclip = false viewing_range = 500 keymap_decrease_viewing_range_min = KEY_MINUS keymap_increase_viewing_range_min = KEY_PLUS mute_sound = false free_move = false
04:41 MTDiscord pitch_move = true fixed_map_seed = 5791653907587404959 selected_world_path = /home/ronoaldo/Software/minetest/minetest-5.4.0-ronoaldo-linux/bin/../worlds/biomeworld mainmenu_last_selected_world = 3 mg_name = flat world_config_selected_mod = 3 language = pt_BR fullscreen = false enable_damage = true creative_mode = false mgflat_spflags = nolakes,nohills,nocaverns enable_waving_plants = false fsaa = 0 anisotropic_filter = false mip_map = false
04:41 MTDiscord enable_waving_water = false tone_mapping = false trilinear_filter = false bilinear_filter = false leaves_style = fancy enable_waving_leaves = false enable_shaders = false node_highlighting = box mg_flags = light,nodecorations,ores,nocaves,biomes,nodungeons
04:41 VanessaE I guess a Ryzen 7/3700 is like 3-4x as fast
04:42 Fixer joined #minetest-hub
04:42 MTDiscord <Ronoaldo> I am terrible with hardware so I don't know for sure ... I barelly know the CPU version or something
04:42 VanessaE you might want to limit your paste size when you're in the "IRC" channels there on the discord
04:42 MTDiscord <Ronoaldo> oops sorry
04:43 VanessaE anyway, as long as the new code feels faster, it's doing its job :)
04:43 MTDiscord <Ronoaldo> it does for my tests now
04:45 VanessaE oh yeah... if I let it get way behind and then quit without letting it catch up,
04:45 VanessaE (there are 21839 entries)...
04:45 VanessaE heh
04:45 VanessaE took about 20s for it to finish.
04:46 MTDiscord <Ronoaldo> yeah, but this test was just an edge case tho, on real world usage this would not happen at shutdown
04:46 VanessaE (maybe 30, I didn't time it closely)
04:46 VanessaE yeah, that was flat mapgen, flying at full speed and not looking back or stopping, then quitting without pausing more than a moment
04:46 VanessaE worst possible case
04:47 VanessaE 20-30s to process 21839 action entries is pretty good
04:47 VanessaE :)
04:47 MTDiscord <Ronoaldo> indeed
04:47 MTDiscord <Ronoaldo> it's like 1000 entities/s
04:47 VanessaE entries
04:48 MTDiscord <Ronoaldo> entries makes more sense ....
04:49 MTDiscord <Ronoaldo> ok, out of curiosity, from here https://gitlab.com/VanessaE/biome_lib/-/snippets/2097514#LC123
04:49 MTDiscord <Ronoaldo> why not initialize all the default replacement tables like the final ones (c3, c4 and c5)?
04:50 VanessaE because string.split() returns a table, {foo, bar, baz, meh} if given "foo,bar,baz,meh" as in input, and those last three can take the tables in that form
04:50 MTDiscord <Ronoaldo> c1 and c2 are initialized with if/else blocks but the c3-c5 ones are initialized with an oneliner
04:51 MTDiscord <Ronoaldo> hmm, c1 and c2 are string:bool "maps" right?
04:52 VanessaE but the first two need tables in the form of {["foo"] = true, ["bar"] = true,.....} so that I can just do  test = "foo" if table[test] then....   i.e. if the element specified by 'test' is present at all in the table
04:53 VanessaE (like at like 139 of the snippet)
04:53 MTDiscord <Ronoaldo> ok
04:53 VanessaE line*
04:53 MTDiscord <Ronoaldo> got it
04:54 MTDiscord <Ronoaldo> the only "edge" case I see is if strings.split("default:dirt, default:air") is used, second entry will have a leading space
04:54 MTDiscord <Ronoaldo> is a trim needed?
04:54 VanessaE string:bool maps..  um..  maybe.  :)  there's a name for this kind of {["foo"] = true, ...} table but I can't remember it
04:54 VanessaE I think spaces are stripped-off?
04:55 VanessaE * `include_empty`: boolean, default: `false`
04:55 VanessaE I assume that trims whitespace as well as things like foo,,,bar
04:56 VanessaE (and if it doesn't, well, the user should remove the spaces from the minetest.conf entry ;) )
04:56 luk3yx VanessaE: It doesn't look like it does trim whitespace
04:56 VanessaE ok.
04:56 VanessaE maybe I can just use that string.trim() call
04:57 VanessaE https://github.com/minetest/minetest/blob/master/doc/lua_api.txt#L3235-L3244
04:57 VanessaE maybe not
04:58 MTDiscord <Ronoaldo> I tought that was a lua stdlib function. It's defined by minetest right/
04:58 MTDiscord <Ronoaldo> ?
04:58 luk3yx Both string.trim and string.split are defined by Minetest
04:59 luk3yx They're defined in https://github.com/minetest/minetest/blob/6c9be39db0d8ae2759cc8d6d5c4b952d13cf4f64/builtin/common/misc_helpers.lua#L171-L208
05:01 VanessaE local function trim(s)
05:01 VanessaE local s1=string.gsub(s, ", ", ",")
05:01 VanessaE return string.trim(string.gsub(s1, " ,", ","))
05:01 VanessaE end
05:01 VanessaE that ought to cover it.
05:01 luk3yx And string.trim doesn't work?
05:01 luk3yx You'd have to run it over every list entry in split
05:01 VanessaE it just says it drops leading and trailing whitespace
05:02 luk3yx Say `for k, v in ipairs(my_list) do my_list[k] = v:trim() end`
05:02 VanessaE not if I trim the string before splitting it.
05:02 MTDiscord <Ronoaldo> the issue is that the split will produce intermediate values that need to be trimmed
05:02 VanessaE e.g. string.split(trim(s)) (trim being the above funt)
05:03 luk3yx So if you wanted to use string.trim you'd have to iterate over the result of string.split
05:03 VanessaE no
05:03 luk3yx But otherwise making your own trim function is probably easier
05:03 VanessaE string.gsub will perform its substitution across the WHOLE string
05:03 MTDiscord <Ronoaldo> what about local s = strings.split ; then use biome_lib.default_water_nodes[s(i)] = true
05:03 VanessaE not just the first match.
05:04 MTDiscord <Ronoaldo> spaces are ilegal in node names right?
05:04 VanessaE yep.,
05:04 VanessaE but they're legal in table keys.
05:04 MTDiscord <Ronoaldo> ok, full replace before split fixes that
05:04 VanessaE so I have to pack the string before splitting it into a table.
05:05 MTDiscord <Ronoaldo> what is pack the string?
05:05 VanessaE as in pack out the spaces
05:05 VanessaE compact
05:05 MTDiscord <Ronoaldo> ok, got it... why not just plain replace(" ", "") from the cfg value then split?
05:06 VanessaE eh?
05:06 VanessaE oh, string.gsub can use an empty string as the replacement?
05:07 MTDiscord <Ronoaldo> sorry, I'm new to lua but most string subs allow that, It may not be the case 😦
05:07 VanessaE it can.
05:07 VanessaE > =string.gsub("foo bar baz", " ", "")
05:07 VanessaE foobarbaz2
05:07 VanessaE that simplifies it a bit :)
05:07 MTDiscord <Ronoaldo> cool 🙂
05:07 MTDiscord <Ronoaldo> ok, so edge case fixed then.
05:08 MTDiscord <Ronoaldo> LGTM
05:09 VanessaE there, reload the split
05:09 VanessaE er
05:09 VanessaE there, reload the *snippet*
05:10 VanessaE (I guess you have to re-download it, git reset biome_lib to HEAD, and re-apply it)
05:12 MTDiscord <Ronoaldo> I did not applied the patch, was just reading the diff 🙈
05:12 VanessaE ok :)
05:13 VanessaE thanks for catching that though :)
05:14 MTDiscord <Ronoaldo> ok, so 2/3 help, it's 2am now I'll go sleep before I start typing stranger things
05:14 VanessaE good night :)
06:19 Kimapr1 joined #minetest-hub
06:24 Kimapr joined #minetest-hub
06:50 hlqkj joined #minetest-hub
07:08 scr267_ joined #minetest-hub
07:44 Kimapr1 joined #minetest-hub
08:00 ShadowNinja joined #minetest-hub
09:39 calcul0n_ joined #minetest-hub
09:53 celeron55 offtopic but interesting https://helplbrysavecrypto.com/
09:56 sfan5 I thought "crypto" meant cryptography, weird ;)
10:09 VanessaE it still does :P
10:10 calcul0n__ joined #minetest-hub
10:15 MTDiscord <appguru> Is configuration being a pain again?
10:37 olliy joined #minetest-hub
10:56 rubenwardy Cryptocurrency does seem more like a security than a currency
10:56 rubenwardy Blockchain bros are so cringe
11:06 rubenwardy https://twitter.com/SaishyKitty/status/1368999942061236232
11:15 Kimapr joined #minetest-hub
12:46 DS-minetest joined #minetest-hub
14:23 hlqkj joined #minetest-hub
14:30 ronoaldo joined #minetest-hub
16:37 Krock no idea how the camera works but I'm getting there sooner or later. https://krock-works.uk.to/u/rotations-2021-03-30_18.36.15.mp4
16:56 Krock https://krock-works.uk.to/u/rotations-2021-03-30_18.56.01.mp4
16:56 Krock how about that
17:03 Krock uhm it's kinda glitchy when colliding with stuff. not sure if depending it on the movement direction is a good idea
18:03 rubenwardy What is that?
18:04 Fixer joined #minetest-hub
18:09 Krock rubenwardy: alternative 3d view, using the same rotation and control mechanisms as seen in other games
18:10 Krock but since there's no "focus lock" for actions (dig, place, punch), it's just pure eye candy at best
18:19 Peppy joined #minetest-hub
18:19 MTDiscord <Benrob0329> That seems like it could be rather helpful for certain types of games, I feel like adding intractability would be worthwhile
20:12 homthack joined #minetest-hub
21:22 MTDiscord <Jordach> Krock: one flaw, it applies with first person attachments
21:22 MTDiscord <GreenXenith> Is that really a flaw?
21:22 MTDiscord <GreenXenith> Sounds like a feature™️
21:31 MTDiscord <Jordach> dewpends
21:31 MTDiscord <Jordach> depends even
21:31 MTDiscord <Jordach> IRL you sidestep
21:32 MTDiscord <Jordach> not rotate your body 90 degrees
21:32 MTDiscord <Jordach> if it only applies to local players that's fine
22:37 Jordach joined #minetest-hub
22:40 panwolfr- joined #minetest-hub
22:42 Calinou_ joined #minetest-hub
22:42 erstazi joined #minetest-hub

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