Minetest logo

IRC log for #minetest-dev, 2016-11-28

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

All times shown according to UTC.

Time Nick Message
00:00 Zeno` it was never committed though... it was only open for about 10 minutes heh
00:00 Zeno` maybe travis is doing it. Very strange
00:00 paramat hmm
00:00 Zeno` the only place the branch exists is in my local repo
00:00 Zeno` just checked my 'official' and it's not there so I didn't accidentally use the wrong repo
00:01 Zeno` (wehn making the PR)
00:02 Zeno` I think github have changed how things work
00:03 Zeno` I'll look at the docs later to see if my workflow needs reviewing but it's pretty standard: make branch on my repo; push to github; open pr
00:05 Zeno` ah, ok I know what happened
00:06 Zeno` I ended up not bothering to do it the "hard" way and edited the file directly using the edit button
00:06 Zeno` which, of course, would make a branch I guess
00:06 Zeno` :3
00:07 paramat aha
00:08 sofar if you 'edit' a file on the github pages it makes a branch
00:11 Zeno` makes sense now
00:13 Zeno` ok, after reading IRC logs I'm ok with #4814 now (when sfan5's request has been added)
00:13 ShadowBot https://github.com/minetest/minetest/issues/4814 -- Limit light_source in the engine by juhdanad
00:19 paramat great
00:22 Zeno` I still don't understand why https://github.com/juhdanad/minetest/​blob/2455294701dac117b398b2218ad119c1​e8000b1c/src/voxelalgorithms.cpp#L230
00:22 Zeno` is necessary, but I'll look later
00:25 Zeno` I mean the <= (I know there is enough space reserved), but why an extra "slot"?
00:26 paramat sofar can i merge game#1422 ? see fixer's comment
00:26 ShadowBot https://github.com/minetest​/minetest_game/issues/1422 -- Fix ItemStack method call for keys. by sofar
00:27 Zeno` LIGHT_SUN (and max light) is 15 so the valid indices are [0..14]... why is there allocation for [0..15] ??
00:30 Zeno` or have I missed something obvious? hehehe
00:36 paramat also i'll make carts rails not ground content and merge that too later
00:38 Fixer paramat: i see this https://github.com/minetes​t/minetest_game/pull/1424 and I have idea i thinked long before
00:38 Fixer paramat: i think I find what dig time we need for creative
00:39 Fixer paramat: it should be such dig time should much walking/fly speed (4 nodes/sec), right now it is slower - that means you need to stop periodically
00:39 Fixer crap
00:39 Fixer paramat: it should be such that dig time should much walking/fly speed (4 nodes/sec), right now it is slower - that means you need to stop periodically*
00:40 Fixer much?
00:40 Fixer match!
00:50 lhofhansl joined #minetest-dev
00:55 paramat i see
00:55 paramat i suggested faster but owners of creative servers requested no change
00:55 paramat it's too easy to accidently break a node behind
00:56 paramat i find it's about right as it is
00:56 Fixer that was my reasoning, because their reasoning was subjecting, but this one at least based on something
00:56 Fixer subjective
00:58 paramat at least now everything is dug at the 'speed of dirt' which is a little faster than the speed of light
00:59 Fixer wait, please also check rail dig speed, i feel it was very slow
01:00 Fixer at least few days ago
01:02 Fixer noticably slow at the time of cart merge
01:05 Tmanyo joined #minetest-dev
01:06 Fixer just tested, like 2x slower than stone
01:07 Fixer for all types of rail
01:10 paramat i know
01:11 paramat possibly due to 'dig immediate'
01:13 fling joined #minetest-dev
01:14 paramat so later i'll merge a fix for rail 'ground content' and games 1416 1424
01:14 Fixer joined vanessas server once more, still no 0 0 0 bug, joined another just for good luck - no 0 0 0 bug, very nice
01:14 Fixer sleep time
01:15 Fixer i will retest 4061 once devs finally review it
01:15 Fixer it was very good once I tested it before some new chanes
01:16 Fixer lava caves have life
01:17 rubenwardy joined #minetest-dev
01:18 paramat game#1425 trivial
01:18 ShadowBot https://github.com/minetest​/minetest_game/issues/1425 -- Carts: Make rails 'is ground content = false' by paramat
01:20 diemartin joined #minetest-dev
01:22 lhofhansl Updated #4811 with last fix for zoom, should be good to go :)
01:22 ShadowBot https://github.com/minetest/minetest/issues/4811 -- Optimize/adjust blocks/ActiveObjects sent at the server based on client settings. by lhofhansl
01:22 paramat will merge games 1416 1424 1425 in a moment
01:30 paramat merging
01:34 paramat complete
01:34 paramat i hope we can get to <= 10 PRs in game
01:36 rubenwardy paramat, that would be cool
01:37 paramat well we could close game#1393 but not sure it has enough disapproval ShadowNinja ?
01:37 ShadowBot https://github.com/minetest​/minetest_game/issues/1393 -- Add sample texture with cape by Rogier-5
01:59 lhofhansl There's also https://github.com/minetes​t-mods/playeranim/issues/9
02:00 lhofhansl At least for me playeranim is broken now. For sofar it's working fine. Can somebody double check?
02:00 lhofhansl Latest git version of playeranim
02:15 ShadowNinja paramat: I agree, wiki material.
02:16 paramat great
02:21 paramat will close it
02:27 sofar lhofhansl: perhaps you can screenshot your "wrong arm" so people can see what the bug looks like?
02:47 lhofhansl I put a screen on thread for https://github.com/minetes​t-mods/playeranim/issues/9
02:48 kaeza joined #minetest-dev
03:19 paramat joined #minetest-dev
03:24 hmmmm joined #minetest-dev
03:44 Zeno` looks normal to me
03:44 Zeno` heh
03:44 Zeno` i quite like it. maybe it could become a feature
03:53 jin_xi joined #minetest-dev
04:08 VargaD joined #minetest-dev
04:13 kaeza joined #minetest-dev
04:27 Hunterz joined #minetest-dev
04:27 paramat #4812 is high priority, reviews appreciated
04:27 ShadowBot https://github.com/minetest/minetest/issues/4812 -- Allow restricting detached inventories to one player by sfan5
04:48 lumidify joined #minetest-dev
06:07 Hunterz joined #minetest-dev
06:11 lhofhansl left #minetest-dev
06:39 nrzkt joined #minetest-dev
07:00 MrIbby joined #minetest-dev
07:29 Zeno` est31, I don't think the performance label is /really/ applicable hehe
07:30 Zeno` I'm fine with the PR now that I've looked through it and yours as well
07:30 est31 well its basically due to abuse of an engine feature
07:30 MrIbby joined #minetest-dev
07:30 est31 detached inventories werent designed to be per player
07:31 Zeno` yeah I guess... put it back if you like... or just merge :D
07:31 Zeno` preferably just merge it I guess
07:32 Zeno` I guess it *is* performance
07:32 Zeno` but every other time I've seen the label removed it's used more for optimisations
07:32 Zeno` *shrug*
07:32 Zeno` s/removed/used
07:33 Zeno` why is travis failing that unit test... didn't I see a PR fixing that recently?
07:34 Zeno` est31, all fixed and ready to merge
07:34 Zeno` #4812
07:34 ShadowBot https://github.com/minetest/minetest/issues/4812 -- Allow restricting detached inventories to one player by sfan5
07:34 Zeno` bbiab
07:35 lumidify joined #minetest-dev
08:01 yang2003 joined #minetest-dev
08:09 Darcidride joined #minetest-dev
08:13 lumidify joined #minetest-dev
08:42 Zeno` merging #4814
08:42 ShadowBot https://github.com/minetest/minetest/issues/4814 -- Limit light_source in the engine by juhdanad
08:47 Zeno` finished
08:49 Zeno` for the record, I did not *insist* on the assert being changed to SANITY_CHECK, I was backing up sfan5's request (which I agreed with)
08:49 Zeno` my comment should not have been enough to stop it being merged if another core dev disagreed
08:50 Zeno` github is messed up
08:50 Zeno` I see now that the comment by sfan5 was actually no longer relevant, but it is still showing in the PR
08:51 Zeno` it makes reviewing more difficult than it should be. Stupid github
08:51 Zeno` still showing as relevant*
08:51 Zeno` anyway, no issue anyway
08:54 nrzkt joined #minetest-dev
08:54 Zeno` will also merge #4812
08:54 ShadowBot https://github.com/minetest/minetest/issues/4812 -- Allow restricting detached inventories to one player by sfan5
08:54 Zeno` nrzkt, the change to UNORDERED_MAP can be a trivial PR/merge if you want to change it after I merge 4812
08:55 nrzkt Zeno`: i'm at work i cannot but you are free to do it :)
08:56 Zeno` nrzkt, I just mean that it should not block the PR being merged
08:56 Zeno` (I don't think)
08:56 nrzkt i don't know, but adding those things with PR adding the containers could be a good practice to prevent forget the performance gain :)
08:56 Zeno` the change to UNORDERED_MAP can be a very quick trivial (change it and merge without approval) after this
08:57 Zeno` Ok I won't merge
08:57 Zeno` sfan5, do what you feel is best and merge after you make the decision :)
08:57 Zeno` nrzkt, you have a good point
08:58 Zeno` I just think we have been getting "bogged down" with trivial things "blocking" things being merged
08:58 Zeno` e.g. see my comments about #4814
08:58 ShadowBot https://github.com/minetest/minetest/issues/4814 -- Limit light_source in the engine by juhdanad
08:58 Zeno` my comment was never meant to stop the PR being merged :/
08:58 nrzkt Zeno`: yes i know, i'm not blocking the PR but if the change could be embeded by commiter, good point :)
08:59 Zeno` it would have been a very different question if I had said "NO DON"T MERGE THIS UNTIL it's not assert!!!!!!!!!!!!!!!"
08:59 Zeno` heh
08:59 Zeno` yep, cool
09:00 Zeno` I generally shout pretty loudly if I disagree with something strong enough that I'd be upset if it was merged
09:00 Zeno` so silly little comments are just that... comments
09:38 proller joined #minetest-dev
10:30 Hunterz joined #minetest-dev
10:51 lumidify joined #minetest-dev
11:26 Fixer joined #minetest-dev
12:15 Darcidride joined #minetest-dev
12:41 sfan5 merging #4812 (the UNORDERED_MAP change should be done in a different commit)
12:41 ShadowBot https://github.com/minetest/minetest/issues/4812 -- Allow restricting detached inventories to one player by sfan5
12:41 sfan5 also merging game#1413 then (counts as trivial fix)
12:41 ShadowBot https://github.com/minetest​/minetest_game/issues/1413 -- Avoid "leaking" creative inventories to other players by sfan5
12:48 nrzkt sfan5: thanks for the merge
12:51 Zeno` yeah, the unordered map in a different PR is what I suggested
12:51 Zeno` but... I may have caused a larger problem
12:51 Zeno` kinda checking master now
12:53 Zeno` in my defense, it worked before!
12:53 Zeno` hmm
12:55 Zeno` init packet not recognised by server
12:55 Zeno` wtf caused that?!
12:55 Zeno` I tested the PR I merged and it was fine
12:56 Zeno` and now it's not working?
12:56 Zeno` Shara, are you awake?
12:58 Zeno` sfan5, nrzkt, can you connect to redcat?
12:59 sfan5 works fine
12:59 sfan5 as of current HEAD
12:59 Zeno` weird
13:00 Zeno` I am getting this
13:00 Zeno` http://dpaste.com/1PA8BPQ
13:00 Zeno` cleaning and rebuilding all
13:01 Zeno` before I merged everything was fine. This is weird, it's making my head explode
13:03 nrzkt redcat ? what is that
13:03 Zeno` yep, I cannot connect
13:04 Zeno` 51.254.216.93, nrzkt
13:05 nrzkt i'm at work :p
13:05 Zeno` nrzkt, go home then
13:05 Zeno` :)
13:07 Zeno` nrzkt, I just called your boss and he has given you the day off with full pay
13:13 nrzkt xD
13:30 yang2003 joined #minetest-dev
13:41 Zeno` joined #minetest-dev
13:49 Zeno` ok I can connect before your PR/commit sfan5
14:01 Darcidride joined #minetest-dev
14:05 xerox123 joined #minetest-dev
14:36 hmmmm joined #minetest-dev
14:42 proller joined #minetest-dev
14:49 red-001 joined #minetest-dev
14:49 red-001 joined #minetest-dev
14:53 Fixer i rarely get that area even before on just test ii during high network load iirc
15:01 jin_xi joined #minetest-dev
15:08 lumidify joined #minetest-dev
15:11 proller joined #minetest-dev
15:24 Fixer_ joined #minetest-dev
15:25 Fixer_ I don't have that problem reported by Zeno
15:29 blaze joined #minetest-dev
15:44 Void7 joined #minetest-dev
15:53 Darcidride joined #minetest-dev
15:55 Hunterz joined #minetest-dev
15:59 Krock joined #minetest-dev
16:01 Kray joined #minetest-dev
16:02 Calinou joined #minetest-dev
16:19 sfan5 (random information for future people) this commit breaks non-ascii TTF font display with irrlicht 1.9 https://github.com/zaki/irrlicht/commit/​c6fd504cf418c8a9f7e9d372aad92e2c6566f03a
16:21 xerox123 joined #minetest-dev
16:22 nore joined #minetest-dev
16:32 proller joined #minetest-dev
16:49 hmmmm joined #minetest-dev
17:14 garywhite joined #minetest-dev
17:14 Fixer impressed by those flows in caves
17:15 turtleman joined #minetest-dev
17:21 Fixer https://i.imgur.com/Vjcij0k.png
17:23 Foghrye4__ joined #minetest-dev
17:32 rubenwardy joined #minetest-dev
17:40 AcidNinjaFWHR joined #minetest-dev
17:48 hmmmm it's cute but not worth the cpu time
17:48 hmmmm those caves are the worst thing to happen to minetest
17:49 hmmmm everybody's server got 50x laggier
17:49 sfan5 are those the v7 caves?
17:51 hmmmm well they were originally part of v5 but yea
17:52 sfan5 as nice as v7 might be, i have to say that i also prefer the v6 caves
17:52 hmmmm paramat added them back when re-implementing v5 and i separated it out from the mapgen so others mapgens can use it too
17:52 hmmmm so now it's being used everywhere
17:52 hmmmm there are definitely more cave algorithms out there, just that nobody really implemented any because there are so many more important things to work on
18:05 jomat joined #minetest-dev
18:09 crazyR_ joined #minetest-dev
18:20 AlexINF joined #minetest-dev
18:26 Fixer v7
18:26 Fixer hmmmm: most of people don't go that deep
18:27 Fixer ok, fine, remove caves, trees, just leave the desert and 3 blocks
18:27 Fixer not all caves have those flows
18:28 Fixer just few
18:28 Fixer max_lag was pretty good
18:29 Fixer no need to panic
18:30 Fixer that was test of flow, those long flows just stay that way after initial flow
18:31 sofar I wonder if new mapblocks can be iterated until stable so that after mapgen finishes with them they don't cause more mapblock updates?
18:32 Fixer i've played at least hour with flows, observed those mapblock updates and they were pretty low, nothing unusual
18:33 nrzkt joined #minetest-dev
18:33 AcidNinjaFWHR salut nrzkt
18:34 Fixer i feel like caves is not worst part
18:34 nrzkt hi AcidNinjaFWHR
18:35 Fixer mapgen surface generation with those biomes feel much slower in terms of jitter, in caves it is pretty quite and no jitter
18:36 AcidNinjaFWHR hi all
18:37 hmmmm sofar, the problem is, define stable
18:37 AcidNinjaFWHR hummm I haven't tested v7 caves what's so specific about it?
18:38 hmmmm I mean if you can define that, then by all means put it into transformLiquids() - we can mark a block as being liquid stable then and any attempts to update it should get ignored in the transforming liquid queue
18:42 Fixer but it is already stable, at least from what I seen flows are stable everywhere
18:43 Fixer i mean they are flowing after mapgen, but they finish and stabilize
18:43 hmmmm don't tell me "it's already stable"
18:44 Fixer visually
18:44 hmmmm tell me the algorithm to give me a yes or no answer to the question "is it stable"
18:44 Fixer why? is 4061 not satisfiying? water takes all cpu?
18:45 hmmmm what is 4061
18:45 Fixer nevermind, eat time!
18:45 AcidNinjaFWHR #4811 is almost ready for merge :P
18:45 ShadowBot https://github.com/minetest/minetest/issues/4811 -- Optimize/adjust blocks/ActiveObjects sent at the server based on client settings. by lhofhansl
18:47 hmmmm oh no come on
18:47 hmmmm that's PR is certainly going to make things worse in terms of latency
18:49 hmmmm actually nevermind, i misread something
18:49 hmmmm looking good
18:51 Fixer wat? when I played with it feel much better especially at low vranges, less jitter
18:51 hmmmm so I guess with this PR, the client request blocks concept is officially dead
18:52 hmmmm sorta wonder how celeron feels about 4811
18:54 AcidNinjaFWHR What does mean "One approval" label on the issue?
18:55 garywhite Wait a minute, #4812 actually fixes the 0,0,0 bug?
18:55 ShadowBot https://github.com/minetest/minetest/issues/4812 -- Allow restricting detached inventories to one player by sfan5
18:55 Fixer yes
18:55 garywhite Hallelujah
18:55 Fixer so many people asking it
18:56 Wayward_One AcidNinjaFWHR, it means one dev officially approves of the PR
18:56 AcidNinjaFWHR ah ok
18:56 AcidNinjaFWHR thanks Wayward_One
18:56 Wayward_One np
18:56 sofar I didn't understand how directly it does, though, I suppose the client has a hard time with some data that it receives from the server about detached inventories?
18:58 Fixer i had big problems with this bug, wait times up to 15 minutes on some servers, due to those detached inventories
18:58 Fixer now it is ok
18:58 AcidNinjaFWHR Awesome
18:59 sofar is it perhaps receiving *every* detached inventory?
19:01 garywhite So that means that  if I go pull the newest git of MT & minetest-game, that it'll be fixed?
19:01 Fixer sofar: every
19:01 Fixer sofar: not only mine but everybodies
19:01 Fixer red cat will update later
19:02 VanessaE my servers are already updated
19:02 VanessaE (as is upstream unified_inventory, which was also affected)
19:02 Wayward_One garywhite, the server will need it too
19:03 sofar well if you run this on your server, please send feedback to the github PR
19:03 garywhite Exactly, so if I get the latest minetest and minetest-game on github for my server, it will fix the issue? (I don't have it personally, but my server does lag, so I wouldn't be surprised)
19:04 Fixer garywhite: update engine & game & unified_inv
19:04 Fixer mods need fix too
19:05 garywhite So basically it'll most likely exist until all mods are corrected?
19:05 Fixer no
19:06 Fixer only some mods cause it to be that heavy
19:06 garywhite So what mods need fixing then?
19:06 Fixer sfan said: mt_game creative (fixed), armor, bags, unified_inventory (fixed)
19:07 garywhite ok
19:07 AcidNinjaFWHR I'd have to retry but I think it doesn't work with rubenwardy/email mod
19:07 Fixer i did some grepping through lots of mods and it is mostly those
19:08 Fixer garywhite: only some servers are affected, and they can easily update those mods and it will fix the problem
19:08 Fixer vanessa already fixed it
19:08 Fixer nice clean login
19:09 Taoki joined #minetest-dev
19:09 AcidNinjaFWHR oh thank you VanessaE :)
19:10 VanessaE you're welcome
19:17 lumidify joined #minetest-dev
19:22 Thomas-S joined #minetest-dev
19:27 AlexINF left #minetest-dev
19:29 Calinou joined #minetest-dev
19:32 xunto joined #minetest-dev
19:50 paramat joined #minetest-dev
19:52 sfan5 Fixer: bags was part of unified_inventory and thus is fixed already, stu released a new version of armor with the fix already
19:52 paramat hmmmm #4061 is almost ready you might be interested
19:52 ShadowBot https://github.com/minetest/minetest/issues/4061 -- Add loaded blocks to transforming_liquid-queue similar to updateLiquid by MillersMan
19:53 Fixer sfan5: nice, i sorted that least few hours ago to see what things need check: http://pastebin.com/raw/pW8Vb5S6
19:53 paramat celeron55 we're interested in your opinion of #4811 since it addresses your issue #1088
19:53 ShadowBot https://github.com/minetest/minetest/issues/4811 -- Optimize/adjust blocks/ActiveObjects sent at the server based on client settings. by lhofhansl
19:53 ShadowBot https://github.com/minetest/minetest/issues/1088 -- Block send range is not large enough for high-end but is already unnecessarily high for low-end computers
19:53 Fixer sorted that in last few hours*
19:54 sfan5 everything with :get_player_name() definitely need a fix
19:54 sfan5 stuff like e.g. lottmobs.trader_inventory = minetest.create_detached_i​nventory(unique_entity_id, move_put_take) probably doesnt
19:55 paramat the worst offenders for lava flows in tunnels are mgv5 and mgvalleys as they place lava sources in tunnels, mgv7 doesn't
19:56 Fixer sfan5: does this need fixing? https://github.com/minetest-mods/global_​exchange/blob/22f9f394c16dd7147d1eb20121​48bb5bc5a0e9a6/exchange_machine.lua#L175
19:56 sfan5 no because it's just one
19:56 Fixer ah ok, nevermind
19:56 sfan5 (player name appears nowhere)
19:56 Fixer you can use my list if you want
19:56 paramat heh mtgame has 8 PRs
19:57 paramat thought i was in the wrong repo for a moment
19:58 lumidify joined #minetest-dev
20:01 Fixer is not you really need to get close to flow to activate? it is not like there millions of flows all around
20:01 paramat Fixer is game#1420 fixed now?
20:01 ShadowBot https://github.com/minetest​/minetest_game/issues/1420 -- Error when digging chest with "key" (bluish one) assigned to that chest
20:02 Fixer paramat: fixed
20:03 paramat thanks
20:19 Fixer found trivial(?) bug: in creative - place boat in inventory, place it, go in a boat, remove boat from inventory, dismount it with mouse click - no boat in inventory
20:21 paramat seems not a bug
20:21 paramat unless i misunderstand
20:22 VanessaE sounds like you don't get the boat back if you don't already have one in the inventory.
20:22 turtleman joined #minetest-dev
20:24 Fixer yes
20:25 lhofhansl joined #minetest-dev
20:26 paramat you dismounted but didn't dig it
20:27 paramat please can anyone review #4061 ? well tested and looks good
20:27 ShadowBot https://github.com/minetest/minetest/issues/4061 -- Add loaded blocks to transforming_liquid-queue similar to updateLiquid by MillersMan
20:27 Fixer while i was in boat i pointed mouse at it and left click
20:29 VanessaE paramat: how does that play alongside the liquid queue tending not to "empty" at a sane rate (ask Zeno`)?
20:31 paramat ah
20:31 paramat VE not sure
20:32 Fixer i think he had some debug output with numbers of blocks in queue
20:32 Fixer he posted them very long time ago
20:32 VanessaE https://daconcepts.com/vanessa/ho​bbies/minetest/stats-weekly.html  Zeno` says the "sawtooth" pattern in the memory usage graph is likely the liquid queue.
20:33 VanessaE just making sure that's kept in mind.
20:33 Fixer VanessaE: why not just output the number and monitor it?
20:33 Fixer millersman knows how to do it
20:33 VanessaE Fixer: "output the number"?  well, I'm sure he or someone else can cover that.
20:33 VanessaE that graph is all I personally have to go by.
20:34 Fixer VanessaE: wonder what happens if you disable flow completely and remeasure?
20:34 VanessaE and you know how Zeno` is about memory usage/leaks ;)
20:34 Fixer yeah
20:34 Fixer millersman can provide numbers
20:41 Fixer sfan5: there are other bag mod used by some servers: https://github.com/cornernote/minetest-bags
20:42 sfan5 oh
20:42 Fixer it has local bags_inv = minetest.create_detached_inventory(​player:get_player_name().."_bags",{
20:45 Fixer VanessaE: you can also try compiling it with 4061 and comparing to older results
20:53 VanessaE I'll pass for now
20:56 Fixer sfan5: can you open an issue for that mod? I have hard time writing what he needs to fix, or maybe made a howto post on forum about 4812
20:58 paramat Fixer maybe this code will fix your boat issue? https://github.com/minetest/minetest_game/com​mit/788ae97ba17891aea44c81adc269c88ea6f18c90
20:59 Fixer hard to say, that fix above was meant to fix increasing number of carts after each cart dig
20:59 Fixer i think
21:00 paramat it works for carts
21:01 paramat tested, your bug doesn't occur for carts
21:01 Fixer yeah, just checked too
21:02 paramat could you open an issue mentioning the tested fix?
21:02 paramat (the cart code)
21:03 paramat confirmed the boat bug
21:04 Fixer this? https://github.com/minetest​/minetest_game/issues/1426
21:04 paramat thanks
21:30 Fixer sfan5: nevermind, just posted to few mods githubs, i hope i got it right
21:31 sfan5 i'm sure they'll understand
21:35 sofar paramat: another case where boat and carts don't share user interactions :)
21:35 sofar I fear we'll never make it coherent, lol
21:37 millersman joined #minetest-dev
21:39 millersman paramat: Did a small update of #4061. This also fixed the last issue where nodes didn't flow away... Maybe I was adding the wrong nodes or missed a few ones
21:39 ShadowBot https://github.com/minetest/minetest/issues/4061 -- Add loaded blocks to transforming_liquid-queue similar to updateLiquid by MillersMan
21:40 paramat looking
21:41 paramat what about the fact that liquids are 'floodable = false'? reflowscan.cpp line 99
21:42 paramat if a node is floodable it won't be a liquid source
21:43 millersman this was the issue that i fixed
21:44 paramat that line still seems to have a check for a liquid source, seems not needed
21:45 paramat see my line cooments
21:45 paramat (comments)
21:51 millersman ah sorry my fault, missed the git add before commiting, pushed again
21:51 paramat ahh
21:52 millersman the test is now wheter it is floodable (which worked before) or it is liquid flowing (which obviously wasn't but i thought that it would be)
21:54 Tmanyo joined #minetest-dev
21:55 paramat i see, your latest change isn't showing up yet in github
21:56 garywhite Hi Tmanyo
21:56 Tmanyo hi
22:02 millersman joined #minetest-dev
22:02 paramat ok i see the changes, it checks for floodable AND liquid flowing, always false
22:04 millersman joined #minetest-dev
22:04 Zeno` joined #minetest-dev
22:04 millersman paramat: sorry, i did something horribly wrong when commiting, will fix this
22:06 millersman joined #minetest-dev
22:06 paramat ok
22:08 millersman for some reason, having it always false works better in some situations :(
22:09 milllersman joined #minetest-dev
22:10 milllersman grrr, i hate my connection
22:10 paramat heh
22:11 milllersman i'll look into the whole issue again tomorrow when i'm more rested
22:13 paramat ok
22:16 AntumDeluge joined #minetest-dev
22:20 DI3HARD139 joined #minetest-dev
22:20 FirePowi joined #minetest-dev
22:23 milllersman joined #minetest-dev
22:23 red-001 joined #minetest-dev
22:27 milllersman joined #minetest-dev
22:31 DI3HARD139 joined #minetest-dev
22:36 milllersman paramat: I think I found the reason why some columns aren't properly updated, it is in fact a problem of the world: some columns consist of two flowing-nodes above each other with both having a very low level (probably zero), adding the upper one removes it but the lower one doesn't get updated
22:37 milllersman not sure whether i should test for such bugs in the world or just ignore them
22:37 paramat don't worry this doesn't have to be perfect
22:38 paramat that situation would be rare i guess
22:38 milllersman normally those columns disappear when they are loaded a second time as then the topmost node was already gone
22:39 Fixer i was already impressed with flows ingame, i hope those changes don't ruin anything
22:40 milllersman joined #minetest-dev
22:41 Calinou joined #minetest-dev
22:42 milllersman i've decided to remove the or is_flowing liquid for the following reason (correct me if i'm worng, i'm already very tired): if the node itself is a flowing liquid it will be directly added and the test will be never executed, if the node itself is a source node and the neighbour is a flowing node the neighbour is added as it will be a flowing node on top of another column and the test results only in unnecessary overhead if the
22:42 milllersman neighoubring node isn't the top node it shouldn't have a lower level so there is either no need to add it or there is something broken in the world
22:42 Fixer thanks to milllersman i'm now testing with std::cout << m_liquid_queue->size() << std::endl; to see size of queue
22:43 milllersman ^^ paramat, sounds my reasoning correct? i'm not sure if i miss something
22:44 paramat i'll have to think about it
22:49 Fixer on fresh world i got mostly below 1000
22:51 Fixer below 500 most of the time
22:52 milllersman joined #minetest-dev
22:53 Fixer anyone wants this data? i can record into debug.txt and plot in excel to see value distribution
22:53 milllersman joined #minetest-dev
22:55 milllersman joined #minetest-dev
22:55 paramat milllersman your reasoning seems correct, but i could be wrong
22:56 paramat we'd have to test the new code
22:56 Fixer now?
22:56 Fixer or new changes planned?
22:56 paramat could be tested after merge
22:56 Fixer i test todays revision before this very latest changes
22:58 paramat wwell, the current code viewable at github has an error, reflowscan.cpp L100
22:58 milllersman i've tested it locally, just double-checking that i don't push some debug code
22:58 paramat ok
22:59 Fixer yes, i'm staying at previous version before those OR changes
23:00 milllersman Yes the OR changes just introduce overhead in the queue that won't result in actual changes in the behavior (or at least shouldn't)
23:01 paramat yeah
23:05 milllersman pushed the new changes
23:09 Fixer can't not share this https://i.imgur.com/MCSH5BV.png
23:11 milllersman Fixer: wow, this is a nice place
23:12 Fixer in the middle of ocean
23:12 Fixer were i test right now
23:12 sofar Fixer: screenshot thread on forum :)
23:16 milllersman paramat: Tested it again, for me it looks good
23:17 paramat good
23:19 milllersman so that is enough for today, bye
23:26 paramat ok the new code is up and looks good
23:27 Fixer did some measuring over ocean https://i.imgur.com/mzLlwib.png
23:27 Fixer don't take it very seriously
23:28 paramat the 2 code changes today actually make no difference to behaviour, they just remove an unnecessary check, so your testing is still valid
23:29 Player_2 joined #minetest-dev
23:30 paramat but some of that may be caused by mapgen?
23:30 Fixer what exactly?
23:30 paramat are you measuring size of liquid queue?
23:31 Fixer yes, i was riding in ocean -> caves under ocean with opening into it -> flow activated -> recorded queue sizes
23:31 Fixer oceans have lots of cave entrances
23:31 Fixer if I walk over land, it is much less I believe
23:31 paramat well, most of this may be liquid queue due to mapgen, not reflow scan
23:32 Fixer i mean it outputs data each time reflow function is called
23:32 Fixer not every second
23:32 Fixer it is good enough
23:37 Fixer below 500 most of the time, shoots up to 5000 on huuuuuge lava basin and goes down rapidly
23:37 Fixer below 1000 now
23:37 paramat ah
23:37 red-001 joined #minetest-dev
23:39 paramat so you're measuring the number of nodes per mapblock added to liquid queue by reflowscan?
23:40 Fixer paramat: std::cout << m_liquid_queue->size() << std::endl; inside that function to update numbers
23:41 paramat ok, queue size
23:41 Fixer i'm flying like crazy and found very big lava cluster which boosted it to 12000 for short time and went down fast to 2000
23:41 Fixer and declined to <1000
23:42 paramat newly generated world?
23:42 Fixer yes
23:42 Fixer https://i.imgur.com/ug2zZvB.png
23:44 Fixer i will record numbers for deep caves tommorrow but from what I see it is below 3000 and often <1000, you need to be lucky to boost it higer
23:44 Fixer higher
23:45 Fixer and i have fly and fast, so thats kinda worsens it
23:46 paramat this isn't very relevant to the PR though, you're just measuring liquid queue due to mapgen, which will obviously be large
23:46 Fixer yes, but it also reflow along the way
23:46 Fixer in some cases
23:47 Fixer it has tendency to go down to 0 with time
23:48 Fixer paramat: i think hmmm m requested this numbers from him long time ago, so i wanted to see it too
23:49 Fixer if just standing it can read 0, if move mouse a bit around bumping to 500 around those lava basins
23:52 paramat the important number is 'number of nodes per mapblock added to liquid queue by reflowscan in already generated world'
23:56 Fixer i can do /emergeblock 1000x1000x1000 and just walk a little in it?
23:56 Fixer numbers should be really low

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