Minetest logo

IRC log for #minetest-dev, 2016-11-27

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

All times shown according to UTC.

Time Nick Message
00:00 Fixer nice
00:01 MrIbby joined #minetest-dev
00:01 lhofhansl yep :) So ABMs are also processed "optimally" with the client's settings in mind.
00:04 lhofhansl AcidNinjaFWHR: With an old client the server falls back to the current behavior (no gain in packet traffic)
00:04 AcidNinjaFWHR great
00:09 AcidNinjaFWHR git clone --depth 1 https://github.com/lhofhansl/minetest.git minetest_range
00:09 AcidNinjaFWHR i'm gonna test
00:09 AcidNinjaFWHR :p
00:13 lhofhansl AcidNinjaFWHR: cool. The branch is range_adjust.
00:13 Tmanyo joined #minetest-dev
00:15 lhofhansl AcidNinjaFWHR: You can best see what's going on when you turn the fog off and then adjust the view range. The next time you move the new setting is sent to the server.
00:23 paramat sorry for all the confusion
00:23 AcidNinjaFWHR still compiling (slow computer) ^^
00:24 AcidNinjaFWHR lhofhansl, do you have a server running it?
00:24 lhofhansl no public server, no
00:27 lhofhansl It'll also do this in Singleplayer mode
00:29 Fixer try mod profiler to see if any effect
00:33 AcidNinjaFWHR is profiler integrated?
00:33 AcidNinjaFWHR I just have to activate through the command /profiler ?
00:34 lhofhansl You mean the integrated one F5/F6? That won't tell you the difference I think, but that's a good idea perhaps.
00:38 octacian joined #minetest-dev
00:40 lhofhansl If you haven't cloned before you'll also get #4724, which retrieves block of the sphere around you rather than a cube (saving about 1/2 of the block - those in the corner that you cannot see anyway)
00:40 ShadowBot https://github.com/minetest/minetest/issues/4724 -- Improve number of blocks sent and occlusion culling. by lhofhansl
00:40 hmmmm heh
00:41 hmmmm the original author of the lighting PR figured it out
00:41 hmmmm i was going to point out that it could overflow but i thought light_source was ANDed by 0x0f... guess not
00:41 AcidNinjaFWHR joined #minetest-dev
00:43 sfan5 by the way unit tests are still broken
00:43 sfan5 i think SN broke them with the fs::<something> fix
00:43 AcidNinjaFWHR https://paste.kde.org/prz3i2091
00:44 AcidNinjaFWHR dunno what you can do with that
00:44 AcidNinjaFWHR wanna me to compare these values to the master branch?
00:45 lhofhansl AcidNinjaFWHR: that'd be cool. But I'd expect them to be the same :)
00:46 AcidNinjaFWHR yeah, considering i'm playing alone xD
00:46 lhofhansl :)
00:48 AcidNinjaFWHR recompiling…
00:51 AcidNinjaFWHR https://paste.kde.org/phn4sm8ux
00:52 AcidNinjaFWHR I have no idea what to look for in that
00:53 lhofhansl hmmm.... Me neither. Is that the exact same scene?
00:53 AcidNinjaFWHR well I just fly around a bit maybe a bit further (because it generated a nyan)
00:54 AcidNinjaFWHR but I guess we should do a stress test (multiple players, with more actions)
00:55 lhofhansl That'd be the thing to do. I'm player with my son with this :)  He still has the old client and I have the new one.
00:55 AcidNinjaFWHR not sure it will be enough to catch a difference
00:56 AcidNinjaFWHR profiler doesn't display bandwidth used
00:56 lhofhansl I logged the block range the server will send to the client. (there's a log line one can uncomment in clientiface.cpp)
00:57 lhofhansl You can see what's its doing if you turn the fog off.
00:58 Tmanyo joined #minetest-dev
01:04 AcidNinjaFWHR lhofhansl, i forwarded a port
01:05 AcidNinjaFWHR if you liked to test
01:05 AcidNinjaFWHR server is at home.fwhost.eu or 88.164.45.194
01:06 lhofhansl Heh cool... Lemme join in for a second.
01:07 lhofhansl connection timed out
01:07 lhofhansl default port 30000?
01:10 AcidNinjaFWHR yes
01:10 AcidNinjaFWHR maybe the dns isn't active yet, try the ip
01:11 lhofhansl that works it seems
01:12 lhofhansl "Media..."
01:13 lhofhansl Arrghh... Need to go anyway... Off to a party :)
01:14 lhofhansl left #minetest-dev
01:15 AcidNinjaFWHR hehe
01:35 AcidNinjaFWHR Here is a video of the lhofhansl range_adjust branch: https://youtu.be/P4lwPsFqaqk
01:52 paramat thanks
01:54 AcidNinjaFWHR it's okay i'm not even sure it's helping anyone xD
01:55 AcidNinjaFWHR have you noticed the "bug" of light when going back to loaded parts of the map when it's night
01:55 AcidNinjaFWHR not sure if it's related
01:55 paramat yes that's normal, happened before
01:56 AcidNinjaFWHR ok :)
01:56 turtleman joined #minetest-dev
02:05 paramat game#1406 doesn't seem right
02:05 ShadowBot https://github.com/minetest​/minetest_game/issues/1406 -- Make drop logic working better for non-8-step-plants by Thomas--S
02:06 paramat i'll merge game#1179 game#1408 game#1412 in 30 mins or so
02:06 ShadowBot https://github.com/minetest​/minetest_game/issues/1179 -- Add simple fast inventory by rubenwardy
02:06 ShadowBot https://github.com/minetest​/minetest_game/issues/1408 -- Carts: Increase power rail acceleration to help climbs by paramat
02:06 ShadowBot https://github.com/minetest​/minetest_game/issues/1412 -- Carts tweak by SmallJoker
03:19 paramat ok merging those 3 in a moment
03:24 yang2003 joined #minetest-dev
03:26 octacian joined #minetest-dev
03:28 MrIbby joined #minetest-dev
03:31 paramat now merging
03:39 paramat merged
04:12 sofar you know what, paramat?
04:12 sofar I'm thinking I actually want powerrail to accelerate more
04:12 sofar here's my thinking
04:12 sofar I was experimenting with the max distance a cart would go
04:12 paramat =S
04:13 sofar turns out I can get them to go 54 nodes horizontal if they go max speed
04:13 sofar since max speed is only 7m/s, that's pretty far
04:13 sofar so that feels ok
04:13 sofar but, it turns out you need 10 powerrails to even get to 7m/s
04:13 sofar that's a LOT of powerrail to get to max speed
04:14 paramat hmm
04:14 sofar and it will be very difficult to make long powered rail sections where you don't have to punch
04:14 sofar I'm seeing right now that you need 2 power rail per 20 to maintain speed (roughly)
04:15 sofar that's an awful lot
04:15 paramat reduce slowing by friction?
04:15 sofar I think the speeds are ok, even if everything is 7m/s
04:15 sofar but the most elegant solution would be to make powerrail accelerate more
04:16 paramat i'm fine with that
04:16 sofar 54 nodes sounds just good for roll out
04:16 sofar I mean, if it's 20 more nodes that would be ok, but I'd rather increase the speed then to e.g. 8m/s
04:17 sofar (which incedentally is what MC has for train speed)
04:17 paramat by the way i feel the handbrake could be stronger, like equal to brake rails, it's so weak as to be almost useless
04:18 sofar yeah, we should also add a handbrake sound :)
04:18 sofar squeeeEEEEEEEEE
04:19 paramat it's almost unnoticeable when used
04:19 paramat heh
04:26 hmmmm joined #minetest-dev
04:30 hmmmm joined #minetest-dev
04:49 garywhite joined #minetest-dev
04:59 MrIbby joined #minetest-dev
05:04 sofar so, cart accel 6 gives you 6 power rail to get to max speed
05:04 sofar that seems OK
05:29 paramat joined #minetest-dev
05:29 paramat 6 is ok with me
05:46 Foghrye4_ joined #minetest-dev
05:52 sofar I'm quite pleased with the cart sound... it does sound really nice to me
05:56 Eater4 joined #minetest-dev
06:09 kaeza joined #minetest-dev
06:10 kaeza game#1416
06:10 ShadowBot https://github.com/minetest​/minetest_game/issues/1416 -- Carts: Take creative mode into account. by kaeza
06:33 lumidify joined #minetest-dev
07:13 VanessaE ok, those "detached inventory leak" changes have gone live on my servers.
07:51 AcidNinjaFWHR joined #minetest-dev
08:16 Krock joined #minetest-dev
08:23 Hunterz joined #minetest-dev
08:31 jin_xi joined #minetest-dev
08:37 AcidNinjaFWHR Hi
08:37 AcidNinjaFWHR I noticed inconsistenies between minetest_game and ruberwardy/email mod
08:38 AcidNinjaFWHR the inventory displays 2 tabs [Crafting][Email] (maybe because it detected the email mod)
08:38 AcidNinjaFWHR but then once you click on email, you can't come back to the inventory window xD
08:39 AcidNinjaFWHR (both pulled from master, because I wanted to update minetest_game to current master) :)
08:41 VanessaE sfan5: the "leaked inv" patch is broke.  Either that or my patch for unified inventory.  unified_inv no longer works
08:42 VanessaE I wonder if AcidNinjaFWHR's issue ^^^ is the same thing
08:42 AcidNinjaFWHR yep
08:42 AcidNinjaFWHR using unified_inventory as well.
08:43 VanessaE AcidNinjaFWHR: my copy of u_i is patched for the "leaked inventory" thing, yours is not.
08:43 AcidNinjaFWHR i could try that.
08:43 VanessaE (not unless you applied my pull request :) )
08:43 AcidNinjaFWHR https://github.com/minetest-tec​hnic/unified_inventory/pull/84 ?
08:43 AcidNinjaFWHR i'll try right now :)
08:44 VanessaE *scratched head* my copy of minetest_game suddenly lacks the necessary patch
08:46 AcidNinjaFWHR no that's not directly related
08:46 VanessaE damn it.  sfan5, merge conflict
08:47 AcidNinjaFWHR VanessaE, you have the same behavior I have?
08:48 VanessaE I don't use the email mod you're using (I don't think).  rather, I'm just getting the default inventory formspec instead of unified_inventory's.
08:48 AcidNinjaFWHR once I clicked the email top button (over the inventory) I see the mail window, but the close button then closes both the mail window and the inventory window and when I press i again, it shows both formspecs with the mail window over the inventory one.
08:48 VanessaE weird
08:49 AcidNinjaFWHR that too yes.
08:49 AcidNinjaFWHR the inventory is the simple one
08:50 AcidNinjaFWHR humm VanessaE
08:51 AcidNinjaFWHR don't we have to add ourselves the priv "ui_full" ?
08:52 VanessaE that's offtopic for here, but that feature only applied if UI is configured to default to "minimal" mode
08:52 VanessaE applies*
08:52 AcidNinjaFWHR oh, that doesn't change anything.
08:52 AcidNinjaFWHR ah ok
08:54 VanessaE ok, looks like rubenwardy's UI patch #58 fixes my issue.
08:54 ShadowBot https://github.com/minetest/minetest/issues/58 -- Small fix to socket.cpp for Linux users by Canisaur
08:54 VanessaE shaddup bot :P
08:57 AcidNinjaFWHR oh well...
09:01 AcidNinjaFWHR I took the mod from your repo, I still don't have unified_inventory shown
09:03 AcidNinjaFWHR brb later I have to do painting (wall and ceiling…) Bye
09:04 nrzkt joined #minetest-dev
09:30 AcidNinjaFWHR joined #minetest-dev
10:02 red-001 joined #minetest-dev
10:23 red-001 #4642
10:23 ShadowBot https://github.com/minetest/minetest/issues/4642 -- Allow the join/leave message to be overridden by mods. by red-001
10:36 juhdanad joined #minetest-dev
10:49 Player_2 joined #minetest-dev
11:00 AlexINF joined #minetest-dev
11:10 sfan5 VanessaE: what broke?
11:10 sfan5 also i tested the patch before and it worked for me
11:12 VanessaE sfan5: nevermind the "broke" part.  it was just a sfinv conflict.  rubenwardy had a PR against UI that I forgot about.  that said, when I tried to apply the "leak" patch to mt_game, it threw a merge conflict
11:12 sfan5 so it's applied correctly now?
11:13 nrzkt joined #minetest-dev
11:13 VanessaE nope: http://pastebin.ubuntu.com/23542478/
11:14 VanessaE (I knmow I probably could have done `git merge sfan5-patch-1` as well)
11:14 VanessaE -m
11:14 sfan5 the sfinv patch broke stuff
11:16 sfan5 fixed the pr, please git fetch or whatever else
11:16 VanessaE sure
11:16 VanessaE applies clean.
11:19 VanessaE it'll go into effect the next time one of the servers cras--er, restarts. :)
11:21 sfan5 umm
11:21 sfan5 oh
11:21 sfan5 those many inv slots you have look horrible with a small window
11:21 VanessaE yeah, I know.
11:21 VanessaE MT doesn't scale the background image
11:22 VanessaE (or rather, split+tile it)
11:22 sfan5 also it looks like the fix applied correctly
11:22 VanessaE it's not on VE-Building yet.
11:22 VanessaE :)
11:22 VanessaE now it is ;)
11:22 sfan5 no more bags or refill detached invs from other people
11:23 VanessaE oh yes, the UI part went in a few hours ago.  that reboot just now brings in the mt_game one.
11:25 sfan5 now we just need Fixer to verify that it fixes the bug
11:46 lumidify joined #minetest-dev
11:57 AcidNinjaFWHR joined #minetest-dev
12:02 proller joined #minetest-dev
12:06 Fixer joined #minetest-dev
12:10 blaze joined #minetest-dev
12:20 juhdanad joined #minetest-dev
12:22 Fixer testing vanessa
12:22 VanessaE ok
12:22 Fixer Hi!
12:23 VanessaE hi
12:24 Fixer first try sucessful
12:26 Krock R
12:27 Fixer can you guys login into daconcepts.com:30008? I need some more people to lag it down
12:27 Fixer i don't experience 0 0 0 bug right now
12:28 VanessaE try 30000 instead, if your machine can handle it
12:28 VanessaE that one is far more likely to experience that bug
12:28 Fixer VanessaE: amazingly I had no 0 0 0 bug on 30000
12:28 VanessaE huh.
12:28 Fixer yeah
12:28 Fixer never had any problems on it apart from low fps
12:29 VanessaE that's surprising, for sure.
12:30 Fixer VanessaE: i need more players on 30008 to make sure it is gone for me (at least 5 players, with stuff in their inventories I guess)
12:32 VanessaE there, that adds four more.
12:33 VanessaE (as soon as client instance #4 signs in)
12:33 ShadowBot https://github.com/minetest/minetest/issues/4 -- cppcheck warnings
12:34 VanessaE shaddup, ShadowBot :P
12:34 Krock VanessaE, check out pull/issue hashtag 1
12:34 VanessaE I can add perhaps four more, but I have something else going on on my box, don't want to risk a system crash.
12:34 Krock so many references to it
12:35 VanessaE hashtag?  HASHTAG?!   blasphemer!  It shall always be called "pound", "hash", or "number" :)
12:35 Fixer "rock"
12:36 VanessaE paper :P
12:36 Fixer where are you? i just logined and see people log out
12:36 VanessaE there are four of me (three "testuser" and one as my usual account) on 30008.
12:37 Fixer VanessaE: make sure they have stuff in inventories i guess
12:37 VanessaE here's two more
12:37 VanessaE they all do
12:38 VanessaE fok that's six of me now.  I don't dare launch any more.
12:38 VanessaE -f
12:45 VanessaE ok, that's eight now (two were on the wrong server and all had their view range set too high).
12:45 FirePowi joined #minetest-dev
12:46 DI3HARD139 joined #minetest-dev
12:48 VanessaE there, ten client instances.
12:48 Fixer one more try
12:50 VanessaE keep trying as much as you want, the server's mostly idle anyway
12:50 VanessaE it would be useful if others here would do the same
12:50 VanessaE my box is maxed out, I can't safely launch any more clients.
12:50 Fixer VanessaE: i've tried like 10 times and I had no 0 0 0 bug, positive sign
12:51 Fixer thanks for your help, you can logout now, I will periodically visit this server during the day, but I think bug is fixed
12:51 VanessaE ok.
12:52 VanessaE great, now the system stats is gonna look like it crashed :)
12:53 Fixer it crashed?
12:54 VanessaE nope
12:54 VanessaE just a shitload of signoffs all at once.
12:55 VanessaE in the stats tracker, the result looks very much like when the server crashes or is restarted.
12:55 Fixer VanessaE: when you massively logged out, I logined and some mapblocks were not loading, i did spawn at correct place, i did another reconnect and it went even better
12:55 Fixer all tests were done under heavy network load
12:55 Fixer i was downloading some stuff in parallel
12:56 VanessaE interesting.
12:56 VanessaE maybe it's just a QoS issue?
12:56 Fixer nono
12:57 Fixer bug happened even if i did not download anything in parallel
12:57 VanessaE hm
12:57 Fixer i did it for maximum bug effect
12:57 Fixer since it needs to download those inventories with slower speeds
12:57 VanessaE at least you spawned at the right place.
12:57 Fixer yes, everything looks right
12:57 Fixer 0 0 0 bug is gone for now
12:58 VanessaE so what was ultimately the cause then?
12:58 Fixer VanessaE: lots of detached_inventories of other players downloaded by player
12:59 Fixer VanessaE: so much stuff that it blocked other stuff
12:59 VanessaE ah
12:59 sfan5 well problem no.2 is that the network stack gets super slow in some cases
12:59 VanessaE I thought that might be
12:59 Fixer VanessaE: after patch only your detached_inventory is sent to you
13:04 Shara Can't wait to get this fix on RC.
13:04 Shara Thanks all for looking at it. :)
13:05 Fixer Shara: did you fixed that 18 light Lamp?
13:05 Shara Fixer: I adjusted the light value in the mod, but not sure if more is required
13:05 Fixer ah ok
13:06 Fixer does it look yellow with diagonal black lines?
13:06 juhdanad So you can not test my PR on RC anymore :(
13:06 Shara Yellow with diagonal black lines?
13:06 Shara juhdanad: If all it was was the lightsource being set to 18 for that node, surely it can be duplicated easily?
13:07 Fixer never mind, i will install it in localhost
13:07 juhdanad Fixer: yes, the texture of the node is just like you described.
13:08 Shara I did not check texture of node, just edited the mod code
13:08 Fixer juhdanad: those lamps were located next to the waterfall btw, i was walking near them
13:09 juhdanad Fixer: yes, the water flooded the lamps which caused a light update.
13:10 lumidify joined #minetest-dev
13:26 proller joined #minetest-dev
13:30 Fixer i will try flooding this lamp
13:30 VanessaE <Fixer> That seems to be oka~&*@$%!#^&%$^&!@#$@! SYSTEM ERROR
13:38 juhdanad joined #minetest-dev
13:41 turtleman joined #minetest-dev
13:57 Fixer juhdanad: did not crash for me even without that patch
13:59 juhdanad Fixer: on RC the bud is fixed server-side.
13:59 juhdanad *bug
14:00 Fixer juhdanad: i've tried in singleplayer, i had no problems with it
14:00 Fixer only some people get this crash
14:00 juhdanad Maybe it depends on the compiler.
14:02 juhdanad When you place a lamp, it declares an array in the stack with 16 elements, and then assigns a value to the 18-th element.
14:12 lumidify joined #minetest-dev
14:34 proller joined #minetest-dev
15:30 Karazhan joined #minetest-dev
15:38 AlexDPR joined #minetest-dev
15:54 octacian joined #minetest-dev
15:58 Karazhan joined #minetest-dev
16:22 hmmmm joined #minetest-dev
16:24 hmmmm joined #minetest-dev
16:37 lumidify joined #minetest-dev
16:47 lumidify joined #minetest-dev
17:06 lumidify joined #minetest-dev
17:08 blaze joined #minetest-dev
17:14 lumidify_ joined #minetest-dev
17:26 Thomas-S joined #minetest-dev
17:36 Fixer will try this https://github.com/minetest/minetest/pull/4811
17:39 sfan5 pushing http://sprunge.us/ROWS?diff in 10 mins
17:50 OldCoder joined #minetest-dev
17:51 nrzkt sfan5, dead link
17:52 sfan5 works fine here, here's it https://github.com/minetest/minetest/commi​t/bb06d377a163b2d168c9d327ad38b871132fa8ea
17:54 nrzkt okay :)
17:58 AlexDPR left #minetest-dev
18:10 paramat joined #minetest-dev
18:23 ElectronLibre joined #minetest-dev
18:30 paramat i might merge the new biome system within a day, details: game#1022
18:31 ShadowBot https://github.com/minetest​/minetest_game/issues/1022 -- Default: New biomes, improve biome points by paramat
18:33 sfan5 paramat: won't that cause weird "biome edges" with old maps?
18:34 Krock ^ needs backwards compatiblity, if possible
18:34 paramat yes, but the biome system was not stable
18:34 paramat compat is not possible
18:35 paramat this is why i've built up a cycle's worth of changes to be merged in 1 commit, less breakage
18:43 lumidify joined #minetest-dev
18:46 * Krock tells paramat that game1419 should be a pull to speed up the process
18:48 Taoki joined #minetest-dev
18:56 octacian_ joined #minetest-dev
18:59 hmmmm mm
18:59 hmmmm biome system?
19:00 Foghrye4__ joined #minetest-dev
19:01 hmmmm paramat:  ??? i assume there's a corresponding PR in the core?
19:01 hmmmm but i don't see anything
19:01 garywhite joined #minetest-dev
19:04 paramat new registered biomes in mtgame
19:04 diemartin joined #minetest-dev
19:04 hmmmm o
19:05 paramat no engine changes
19:05 hmmmm that's not changing the biome system then
19:06 paramat yeah my wording was a little unclear
19:08 paramat so everyone, shall we now make mgv7 the default mapgen? i'm fed up with seeing 'lets try' videos of MT using mgv6
19:09 paramat mgv7 terrain is stable (apart from the optional, default-off floatlands)
19:10 paramat registered biomes are probably not stable yet (they may never be)
19:10 hmmmm yea sorry i never gave the floatlands a proper review
19:10 hmmmm i looked through it and all the code was honestly perfect
19:10 hmmmm you'd expect there to be at least one mistake somewhere but not this time
19:11 paramat heh cool
19:11 paramat mgv6 makes MT look fairly boring
19:12 paramat the only thing is, mgv6 always has trees nearby it's good for not scaring kiddies
19:13 paramat but we have bushes now so wood will be more available
19:17 paramat MC players tend to panic if they can't see a tree from spawn. they're conditioned to be desperate to get wood before nightfall :]
19:18 paramat i'll open an issue
19:20 paramat there is one #3655 , c55 doesn't like mgv7 but may be remembering the older version
19:20 ShadowBot https://github.com/minetest/minetest/issues/3655 -- Defaulting to Mapgen V7
19:21 Fixer on 2b2t there is no grass or trees like 500 blocks away from spawn, they are duping tnt and withers right now and gonna level that spawn for sure this time
19:24 blaze joined #minetest-dev
19:29 octacian joined #minetest-dev
19:34 rubenwardy joined #minetest-dev
19:39 rubenwardy AcidNinjaFWHR: https://github.com/rubenwardy/email/commit​/70fe6824164cfdb3fa9fde04329a34f7b2e635a9
19:40 garywhite r u kidding Fixer?
19:40 rubenwardy email had support for an early version of sfinv with a different API
19:40 Fixer no
19:41 garywhite They're blowing up the spawn :/
19:41 AcidNinjaFWHR_ joined #minetest-dev
19:56 paramat hmmmm are you still opposed to the idea of a separate mgv6.2 that has mgv6 terrain but uses the biome API and has mgv7 biomes? i remember you insisted it be an option within the existing mgv6, however it would be much easier to implement as a new mapgen, and with your 'mapgenbasic' changes a small file too
19:58 paramat in fact it would be a complete nightmare incorporated into mgv6
19:59 lumidify joined #minetest-dev
19:59 hmmmm well
19:59 hmmmm at that point you're reimplementing the same mapgen with slightly different terrain shapes
19:59 hmmmm mapgen v6 has things like mudflow and whatever that give it uniqueness
20:00 sfan5 Fixer: can you re-check that the 0,0,0 bug is fixed on ve-creative?
20:00 sfan5 if it is then the pr should be good to merge
20:00 Fixer sure, i forgot about it :)
20:00 paramat yeah the new one wouldn't have mudflow
20:00 Fixer i was just testing 4811
20:01 paramat anyway i'm hardly desperate to do this, but could do it quickly and easily as a new mg
20:02 hmmmm it's hard for me to call a mapgen that just has a different terrain shape generating function and nothing else a new mapgen
20:02 hmmmm maybe we should stop calling it a new mapgen
20:03 sfan5 call it "different" instead :P
20:03 hmmmm a mapgen consists of a terraingen and a node decorator ...?
20:03 hmmmm and all these are really just pairing different terrain generators with the same node decorator
20:03 hmmmm or something
20:03 hmmmm i dunno
20:04 paramat correct
20:04 Fixer sfan5: does number of players online affect it? or it sends more less the same amount?
20:04 sfan5 Fixer: the number of players that *have been* online affects it
20:04 Fixer sfan5: i've logined without 0 0 0 bug but there were no players
20:04 Fixer ah nice
20:04 Fixer i have no 0 0 0 bug
20:04 sfan5 nice
20:04 Fixer checked it under stressed network too
20:05 sfan5 anyone mind reviewing #4812?
20:05 ShadowBot https://github.com/minetest/minetest/issues/4812 -- Allow restricting detached inventories to one player by sfan5
20:16 sfan5 *silence*
20:17 Fixer_ joined #minetest-dev
20:18 paramat lol
20:22 Fixer_ i don't remember but without 4811, does it load blocks selectively or everything around?
20:23 Fixer_ you can't imagine how much better in forests with SFS patch
20:23 lumidify joined #minetest-dev
20:24 paramat server sends blocks within the view frustrum and out to 'max block send distace' (10 = 160 nodes)
20:24 paramat the fixed 160 nodes is the problem
20:25 Fixer_ i see 1.31+E06 vertices drawn and got 96fps... while on actual server spawn with same amount i get like 25 fps, reason?
20:25 Fixer_ difference is in first scene is big forest, in second - city spawn on top of water
20:26 lhofhansl joined #minetest-dev
20:26 Fixer_ paramat: thats without 4811?
20:27 Fixer_ reading in 4811 "the server will only send blocks that the client can actually see "
20:27 Fixer_ is not that was before too?
20:27 lhofhansl Hi... Missed the beginning about 4811 discussion. Anything I can help with :)
20:28 Fixer_ i'm playing with it right now
20:28 Fixer_ walking around forests and stuff, looking at cm: block in range
20:28 lhofhansl You should not see any change in behavior.
20:28 Fixer_ do i benefit it if I use 160 v_range that I use in game?
20:29 lhofhansl Other than active objects not being processed when outside of the visible range.
20:29 Fixer_ only people on calculators will benefit?
20:29 lhofhansl you'll benefit in that it will render correctly now :)
20:29 Fixer_ in what sense?
20:29 lhofhansl oh wait... v_range.
20:29 lhofhansl Yeah, you won't benefit in that case.
20:30 lhofhansl (with default settings)
20:30 Fixer_ block load depended on frustrum even before this, yes?
20:30 lhofhansl yes, but hardcoded to 72 degrees, now it's correct.
20:31 Fixer_ i see its effect now, it is much smoother with short v_range
20:31 Fixer_ less map block loading
20:32 Fixer_ people with calculator will love it I hope
20:32 lhofhansl yep... Less work for the server, less traffic, less work on the client for blocks you can't see anyway :)
20:32 Fixer_ since a lot of people use v_range 25 that will help a loot for servers
20:33 lhofhansl personally I don't understand how anybody can play with 25 even on a phone, but that's just me
20:33 paramat Fixer_ yes what i wrote was 'without 4811'
20:33 Fixer_ on my tablet it is around 20 fps with 25 v_range
20:34 Fixer_ or less
20:34 lhofhansl that it's another thing to look at eventually. We also have Minecraft PE and it seems to be mush smoother, so there's room for improvement.
20:35 lhofhansl I still wanna look into why occlusion culling seems to be so ineffective (in wireframe mode - thanks again for that idea, btw) one can see lot's of blocks that should be obscured, but are shown anyway.
20:36 Fixer_ yes
20:36 Fixer_ especially deeep ones
20:39 milllersman joined #minetest-dev
20:39 proller joined #minetest-dev
20:58 milllersman paramat: #4061 is updated and in the final state (or at least close to)
20:58 ShadowBot https://github.com/minetest/minetest/issues/4061 -- Add loaded blocks to transforming_liquid-queue similar to updateLiquid by MillersMan
20:59 Fixer very nice
20:59 Fixer i will try it too
21:00 Fixer major water bugfix
21:00 paramat ok
21:01 milllersman maybe there are a few style issues but otherwise it should be ok
21:01 paramat btw mgv7 floatlands cause a lot of falling water columns, could be a good place to test
21:04 milllersman i've been using https://github.com/minetest/mine​test/files/101456/Just.Test.zip which is IMHO a very good test-case
21:05 Fixer it has insanely buggy stuck liquids
21:05 Fixer extreme case
21:05 Fixer only problem for me were gups in previous versions
21:06 Fixer i will compile it today
21:06 milllersman i've seen one location in that world where the scan finds nodes that will not flow away after it, only after a second load... this is repeatable but seems to have a different cause
21:06 Fixer if it is rare, don't bother
21:08 milllersman Fixer: The gaps should be fixed now... They weren't an issue of detecting water that needs to flow but more of preventing water falling/flowing away when the source-node wasn't loaded yet
21:08 paramat ^ hmmmm
21:09 Fixer milllersman: for me gaps presented two problems: 1) people who swim up may end up at one 2) unnecessary calculations when gap approaches slop %)
21:09 Fixer slope*
21:10 milllersman i have three commits in my PR, one small unrelated fix, one that prevents the gaps and one for adding loaded blocks
21:12 Fixer paramat: i thinked a bit about keys... how I'm gonna remember them? o_0 they are just "key", if I have like 10 of them, that will be hillarious picking it up
21:13 Fixer i have an idea
21:13 Fixer we should have ability to NAME them
21:14 paramat i'm not getting involved with keys. the ninj has commented about this issue
21:14 Fixer where?
21:16 torgdor joined #minetest-dev
21:20 paramat PR thread
21:20 Fixer ah
21:32 Fixer compiling 4811 and 4061 in one package, kinda not cool, but wth
21:50 sfan5 paramat> btw mgv7 floatlands cause a lot of falling water columns
21:50 sfan5 am i missing something?
21:51 sfan5 i haven't seen any water at all yet in my floatlands world
21:51 Fixer i seen small pool somewhere in my world
21:52 sfan5 i've seen water in a different world, but never water that is exposed to falling down
21:54 sfan5 #4812 still requires review (i'd like to have it merged soon)
21:54 ShadowBot https://github.com/minetest/minetest/issues/4812 -- Allow restricting detached inventories to one player by sfan5
21:54 sfan5 ^ paramat, hmmmm
21:56 lhofhansl left #minetest-dev
22:06 Fixer it is so satisfying to watch those liquids go down
22:09 milllersman the not so nice part is to see that Just.Test causes a constant load of around 20k nodes in the transforming-liquids-queue, but only an average of 7-20 nodes per block, that water above empty blocks is the killer
22:12 Fixer milllersman: don't forget it has holes in the ground, that will cause unnecessary calculations
22:12 Fixer milllersman: just test is only for checking stuck liquids, it has empty bottom with looots of water
22:13 Fixer milllersman: you will rarely see such extreme numbers IRL
22:14 milllersman yes, i know... the lavafalls of Waterflow only create up to a few hundred nodes while walking or a few thousand while flying fast
22:15 Fixer i'm at just test now flying looking how stuck liquids go down, they go nice
22:15 Fixer then will check gaps in waterflow, and then fresh world, water above caves in oceans
22:17 Fixer sorry for holes in just test, it is pain in the ass to cover it with stone below
22:17 milllersman joined #minetest-dev
22:17 Fixer worldedit ooms
22:18 milllersman joined #minetest-dev
22:22 paramat sfan5 lakes in floatlands can be spaced very far apart, but can be big
22:23 paramat the water sometimes travels through tunnels to the floatland underside, then falls as a column
22:24 milllersman joined #minetest-dev
22:24 Fixer current fog seems somewhat extreme to me, what do you think? example: https://imgur.com/a/XBZ08
22:26 Fixer quite a lot is lost at the end in that particular scene
22:26 millersman joined #minetest-dev
22:28 millersman joined #minetest-dev
22:29 millersman Fixer: no problem, the plenty endless columns were good regression tests, in the first version they sometimes got stuck
22:46 Fixer flying in waterfall map
22:46 Fixer lavaflow*
22:51 Fixer I feel satisfied with 4061
22:51 Fixer it is good enough and just works
23:02 millersman joined #minetest-dev
23:04 millersman joined #minetest-dev
23:08 Fixer millersman: liquid queue looking good at fresh worlds? near oceans, rivers?
23:09 Fixer i see server:liquid transform 0.000X - is this it?
23:14 millersman joined #minetest-dev
23:17 millersman joined #minetest-dev
23:17 paramat added some code review
23:17 millersman joined #minetest-dev
23:18 millersman Fixer: Not sure, I have removed the debug output so the output is from somewhere else
23:19 Fixer millersman: i did some testing on just test, lava flow maps, v6, valleys, near oceans, caves underwater, falling sand underwater - looks good to me, i really like it
23:21 Fixer fps seems good, not much block updates near oceans
23:21 sofar Fixer: did you test game#1422 ?
23:21 ShadowBot https://github.com/minetest​/minetest_game/issues/1422 -- Fix ItemStack method call for keys. by sofar
23:22 Fixer i can try it
23:22 millersman joined #minetest-dev
23:22 Fixer super easy to change by hand
23:24 troller joined #minetest-dev
23:26 sofar I'm fairly sure when that broke, and yeah, I didn't test it properly enough apparently
23:28 millersman paramat: Should I add some comments explaining what's going on to the scan method itself?
23:28 millersman In short I'm just scanning all columns inside a block or immediately beneath the block (in case they were disabled by anti-gap code)
23:29 paramat some comments are helpful for non-obvious stuff
23:29 millersman In each column I'm looking for the topmost and lowest nodes in water and check whether they can flow to the side
23:30 paramat makes maintenance easier later
23:30 Fixer sofar: it fixes the issue, what is interesting, if I place another empty locked chest and try dig it with that key - it won't dig it %)
23:31 Fixer sofar: so it has sideeffect
23:31 millersman The complex stuff is just that there are many nodes on the border and have neighbours in the blocks above/below or to the side, so i'm using a lookup to not always query the map for the actual block/node
23:31 Fixer sofar: tested in creative
23:31 sofar left click with a key isn't supposed to do anything
23:31 Fixer sofar: in creative, let me try noncreative mode
23:31 paramat so yes restrained extra comments would be good
23:32 Fixer sofar: but you still can dig with left mouse botton
23:32 Fixer sofar: while holding key
23:32 millersman joined #minetest-dev
23:32 sofar could be the chest code preventing digging
23:33 Fixer could be
23:33 Fixer it reappiers
23:34 paramat i'll fix game#1419
23:34 ShadowBot https://github.com/minetest​/minetest_game/issues/1419 -- Revert rebase error in creative
23:35 Fixer sofar: i feel like it detects that it is wrong key and denies dig?
23:35 millersman joined #minetest-dev
23:38 Fixer sofar: you think tooltips not supported, how about making it sign like? player can rename the key by left clicking and it will show as tooltip you see for signs, yes, you will need to scroll through them in inv row but at least it is something...
23:38 Zeno` joined #minetest-dev
23:39 Fixer i've noticed that creative menu now set at crafting by default, bug?
23:40 paramat probably due to sfinv merge
23:41 paramat should be fixed
23:41 paramat can you open an issue?
23:41 Fixer opened
23:41 Fixer https://github.com/minetest​/minetest_game/issues/1423
23:42 sofar Fixer: can't do infotext for itemstacks either
23:42 Fixer damn :(
23:44 Fixer so only crappy option for now is left clicking to see its name in formspec..
23:46 Fixer decided to dive deep to lava lakes
23:47 Fixer activates nicely
23:48 Zeno` hmmm
23:54 millersman joined #minetest-dev
23:54 millersman paramat: Have updated #4061 again, fixed formating issues and added comments
23:54 ShadowBot https://github.com/minetest/minetest/issues/4061 -- Add loaded blocks to transforming_liquid-queue similar to updateLiquid by MillersMan
23:54 paramat ok
23:55 paramat game#1424 trivial
23:55 ShadowBot https://github.com/minetest​/minetest_game/issues/1424 -- Creative: Re-implement constant hand dig times, add comment by paramat
23:56 Shara joined #minetest-dev
23:56 paramat Zeno` can this branch be deleted? https://github.com/minetest/​minetest/tree/Update-credits
23:56 millersman joined #minetest-dev
23:57 millersman left #minetest-dev
23:58 Zeno` paramat I suppose so
23:58 Zeno` is it a branch?
23:58 Fixer minetest_game had sfan related branch?
23:58 Fixer or has?
23:58 paramat yeah saw that too
23:58 Zeno` how did it become a branch?
23:58 paramat dunno it's weird
23:59 Zeno` yeah delete it of course!
23:59 paramat odd new branches keep popping up
23:59 Zeno` it should not be a branch in the first place
23:59 Zeno` maybe github are doing maintenance or something. very odd
23:59 paramat seems a side effect of some commits

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