Time  Nick       Message
00:27 Unarelith  rubenwardy, #7396 could be great in 5.0.0, it's not even in the milestone and is only missing one approval
00:27 ShadowBot  https://github.com/minetest/minetest/issues/7396 -- World config: Make depends easier to read by HybridDog
00:27 Unarelith  well, #7377 too if it's ready, it's really trivial
00:27 ShadowBot  https://github.com/minetest/minetest/issues/7377 -- minetest.log: Add mod origin by HybridDog
01:23 Unarelith  rubenwardy, I edited #7971, your changes are made, only testing is required now
01:23 ShadowBot  https://github.com/minetest/minetest/issues/7971 -- FormSpec: support custom size and spacing for slots in list by Quent42340
01:27 rubenwardy nice work
01:35 Unarelith  ty rubenwardy
01:35 Unarelith  I'll need your help to make this PR better though
01:43 Unarelith  did paramat add a config option for stars when he decided to hide them under that fake horizon? I'm making a skyblock game and the sky has lost all its beauty at night
03:49 paramat    Unarelith i didn't, sorry. i made a forum post and asked for opinions and there seemed to be very little need for an option
03:49 Unarelith  I missed your post :/
03:49 Unarelith  I'm currently making a skyblock game for 5.0.0 ^^'
03:50 Unarelith  Well, it won't be ready for 5.0.0 anyway so it's not a
03:50 Unarelith  *big deal, but could be really nice to have an option
03:50 paramat    doing it properly may be quite involved, and depends on how the sky api evolves, so i don't want to rush that into 5.0;0
03:50 Unarelith  yep you're right
03:51 Unarelith  I opened an issue about this anyway
03:51 paramat    Jordach's work may be able to allow deep space
03:51 Unarelith  is he still working on it?
03:52 paramat    yes
03:52 paramat    he's reworking his custom sky stuff
03:53 Unarelith  oh nice
03:55 Unarelith  btw paramat, please take a look to #7396, it's trivial, not in 5.0.0 milestone, but it's a really interesting PR missing only one approval
03:55 ShadowBot  https://github.com/minetest/minetest/issues/7396 -- World config: Make depends easier to read by HybridDog
03:57 Unarelith  (well, I'm assuming the conflict isn't a big one, but I can't really see)
04:05 paramat    hmm
04:10 paramat    it's simple, but also low priority, i can't yet get inspired to review it sorry
04:43 Unarelith  ok np, I overestimated the goal of this PR anyway ._.
08:51 Unarelith  btw nerzhul, can you take a quick look at #7971 please? if you can tell me what you think would be the best I'll do it now, else I'll wait for rubenwardy
08:51 ShadowBot  https://github.com/minetest/minetest/issues/7971 -- FormSpec: support custom size and spacing for slots in list by Quent42340
09:37 nerzhul    i'm at work no time for a such huge pr now sorry
09:40 Unarelith  huge? well I think I might have a different definition of huge, but np I understand
09:42 nerzhul    in terms of feature, and the formspec part is not the easier to review, :)
09:44 Unarelith  yep, I agree, but this PR is certainly not huge :')
09:58 crazyR     I've noticed in the latest dev update that entities dont show in game. they are there just not visible. is this a known bug?
10:00 Unarelith  crazyR, just tried with latest commit and everything works as expected
10:01 Unarelith  are you trying to join a specific server?
10:02 crazyR     its happening on my server, Creatively Survival
10:02 crazyR     it was fine before i ran an update. and i only noticed it when i was informed by a player
10:03 crazyR     item frames are the most obvious. but also world edit markers dont show neither
10:05 Unarelith  well, just tried to join Glitchtest since it has mobs, and they're displayed correctly, which build are you using exactly?
10:06 p_gimeno   maybe it's a certain type of entities only?
10:06 Unarelith  could be
10:07 Unarelith  crazyR, minetest version should be in game window title bar
10:08 crazyR     Unarelith https://github.com/minetest/minetest/commit/08610aad7e70ff2787b7a799af4d40536a56a582
10:08 crazyR     let me run an update now and see if still does it.
10:12 crazyR     Seems to have fixed it. sorry for wasting your time
10:14 p_gimeno   yeah, worldedit markers are working fine for me
10:14 Unarelith  nerzhul, and do you have time to review the smallest PR ever? :p #7976
10:14 ShadowBot  https://github.com/minetest/minetest/issues/7976 -- HUD: Font shadow added on item stacks for better readability. (Really trivial) by Quent42340
10:24 Unarelith  hmm just looked closer to the current version, it actually has a small shadow
18:03 Krock      nerzhul: that's an insane long path from ServerEnvironment::kickAllPlayers until the player finally gets deleted.. is that normal?!
18:03 nerzhul    in terms of player removal workflow ?
18:04 Krock      yes. it gets pushed through at least two event queues
18:05 Krock      actually three. currently following the path of PEER_REMOVED
18:06 Krock      finally reached Server::DeleteClient. this feels like a jump n' run game
18:17 Krock      okay, the register_on_leaveplayer issue in singleplayer is actually caused by that the server connection closes before it can disconnect the peer
18:17 Krock      i.e. ConnectionSendThread::sendPackets is not called
18:20 Unarelith  is anyone preparing an extensive documentation based on `lua_api.txt`? the wiki is really outdated so it could be nice to start writing future doc now
18:21 rubenwardy https://pauloue.github.io/lua_api/
18:23 Unarelith  wow nice
18:24 jas_       krock also when hosting a multiplayer world from the gui
18:24 jas_       (sorry, for what it's worth)
18:25 Krock      jas_: very same issue
18:25 Krock      but thanks for confirming
18:29 nerzhul    Krock our player init & removal model is crazy yes
18:41 p_gimeno   do I still need to bump the protocol version if something changes? or does it now go into a general "5.0.0 release" protocol?
18:41 Krock      (refactor time)
18:43 Krock      p_gimeno: depends which features we want to add to 5.0-dev. here I'd say you can just leave the protocol version. It's meant to be unstable anyway (in terms of compat)
18:44 p_gimeno   ok, so I'll wait for a direct request to bump it
19:03 Krock      but consider adding a line to the networkpacket.h changes list anyway
19:04 Krock      it's what usually happens when adding a feature which requires a new proto version
19:08 nerzhul    rubenwardy can't we host that on github pages or gitlab pages ?
19:08 nerzhul    p_gimeno :for current release you can
19:08 nerzhul    generally we did this once per release
19:08 rubenwardy It is on GitHub pages
19:09 nerzhul    rubenwardy ofc but not ours, then not maintained directly by our pipelines
19:09 p_gimeno   thanks nerz
19:09 rubenwardy Sigg
19:09 rubenwardy *sigh
19:09 rubenwardy It's not ready yet
19:09 rubenwardy Try searching for things
19:09 p_gimeno   Krock: will do
19:09 nerzhul    was this done by a -hub user ? i don't know about that gh username
19:20 stujones11 nerzhul, it looks like you broke the android build, again #7978
19:20 ShadowBot  https://github.com/minetest/minetest/issues/7978 -- Lastest android build fails
19:21 nerzhul    stujones11: it's a test to see if you are alive
19:21 nerzhul    but yeah i forgot again to update the android make file
19:23 stujones11 I thought that but there may be other problem, too
19:23 stujones11 why the undefined ref to u32Tof32Slow, etc?
19:24 nerzhul    not me
19:24 nerzhul    it's Krock !
19:24 Krock      no it was nerzhul! surely
19:24 * Krock    hides
19:25 nerzhul    :D
19:25 nerzhul    i will push the android makefile fix in 1 hour i'm busy now
19:25 Krock      oh well, the file is missing here https://github.com/minetest/minetest/blob/master/build/android/jni/Android.mk#L242
19:25 Krock      nerzhul: can do that for you
19:26 nerzhul    Krock don't hesitate: )
19:26 nerzhul    both client & server file
19:26 stujones11 okay, I will re-test once completed
19:26 nerzhul    i hope at a point i can finis the cmake switch
19:27 Krock      stujones11: after adding src/util/ieee_float.cpp, is there any other error?
19:27 stujones11 yes, that would be great!
19:27 nerzhul    both Krock and my activerobject managers cpp files
19:27 Krock      ok
19:27 nerzhul    that should do the trick
19:35 p_gimeno   Krock: could you please take a look at #7973? unit tests are currently failing in release mode due to a GCC bug and a too permissive check, and the fix is fairly simple (I don't dare say trivial but close)
19:35 ShadowBot  https://github.com/minetest/minetest/issues/7973 -- Fix unit test in release mode (consequence of -ffast-math)
19:36 Krock      https://pastebin.com/raw/M9e8fdxh will merge in 15 minutes unless it breaks the build for stujones11
19:37 Krock      eeks. that's awfully sorted
19:41 Krock      p_gimeno: so simply skipping the test as if nothing happened?
19:41 Krock      I'd try to add at least a line which states that the test cannot be executed due to the compiler flags
19:42 p_gimeno   if the compiler does not have a working IEEE implementation, you can't test against a known-good IEEE implementation
19:42 p_gimeno   or the system
19:43 p_gimeno   the general tests that precede that point are performed either way
19:46 nerzhul    what is the problem ?
19:49 p_gimeno   nerzhul: the unit test is failing in GCC in release node because of -ffast-math
19:49 p_gimeno   the new float serialization test
19:50 nerzhul    i see
19:50 p_gimeno   it's a bogus error, the code works as intended, but (1) we forgot to check for is_iec559, and (2) GCC lies to us about its IEC559 support
19:51 nerzhul    is this due to a precision loss ?
19:51 p_gimeno   no, it's because isnan() reports false for nans in -ffast-math
19:51 nerzhul    in which case ? is this not fixeable ?
19:52 p_gimeno   uh, e.g. float a = 0.0f/0.0f; puts(isnan(a) ? "true" : "false"); prints false
19:53 p_gimeno   it's fixable by not running the test in the way that it's not intended to work; the fix is in #7973
19:53 ShadowBot  https://github.com/minetest/minetest/issues/7973 -- Fix unit test in release mode (consequence of -ffast-math)
19:53 nerzhul    clang doesn't have the problem ?
19:54 p_gimeno   I don't know, I haven't tried with clang
19:55 Krock      pushing patch ...
19:55 Krock      (anroid)
19:58 p_gimeno   With -ffast-math, g++ breaks many guarantees that IEC559 gives, but still reports itself as supporting IEC559 despite not being. The test relies on a working IEC559 implementation, because it uses that for comparison.
19:59 Unarelith  p_gimeno, it's not only g++, clang too
19:59 * Krock    wonders why it worked in the PR
20:00 p_gimeno   Unarelith: thanks, I didn't know whether clang also makes isnan() return false for NaNs
20:00 p_gimeno   Krock: not in release mode, maybe
20:00 p_gimeno   Krock: it only happens in release mode, because that's what enables -ffast-math
20:00 Krock      right, I also always use debug builds to test
20:04 p_gimeno   it's a gcc bug, actually, but we also neglected to check is_iec559 instead of just the serialization mode
20:08 p_gimeno   Krock: as for the message, there's none when it return without performing the test if the serialization mode is FLOATTYPE_SLOW
20:08 p_gimeno   I think this is the same case
20:09 p_gimeno   it's just that that check is incomplete, and is_iec559 is better
20:10 Krock      for the float re-implementation it's also skipping that.. alright
20:25 nerzhul    zdz)
21:05 nerzhul    Krock i just published the following PR as an addition to the DOcker doc PR: #7979
21:05 ShadowBot  https://github.com/minetest/minetest/issues/7979 -- Add a basic kubernetes deployment file by nerzhul
21:05 nerzhul    it's simply and example kubernetes configuration to use the minetest container image on a k8s if server owners want to host them on a such cluster :)
21:37 Krock      nerzhul: that .yml file should either be in misc/ or util/
21:41 nerzhul    okay ty
21:46 nerzhul    except that is this okay ?
21:47 Krock      didn't test but it looks good
21:48 Krock      +1 afterwards
21:50 sfan5      so what is this about turning F1000 into F32?
21:50 sfan5      any place it needs to/should/shouldn't be done?
21:50 Krock      anything which is written to the disk
21:51 Krock      the network however can fully be replaced
21:55 sfan5      do we have the precision problems with the disk format too?
21:55 sfan5      (or can we leave that as F1000?)
21:59 nerzhul    Krock will move nad merge asap then
22:01 Krock      sfan5: afaik there's yet no issue with it. I mentioned it because the saved data needs to be release-interchangeable
22:04 Krock      large numbers have a good precision; I also remember multiplying values by 10 or 100 in some places to improve the precision for this case
22:05 Krock      we won't get around sanitizing the values anyway
22:08 sfan5      entity position is already multiplied by 10 I think
22:09 Krock      various magic numbers for the position sending https://github.com/minetest/minetest/blob/master/src/client/client.cpp#L868-L878
22:10 Krock      but those are converted to integer values
22:34 p_gimeno   entity position is stored as a single in memory, so in the range 16384-32767 the maximum precision is 1/512 which is worse than 1/1000. Multiplying it by 10 would only be useful at a radius of about 800m around the centre.
22:35 p_gimeno   Scratch that, it would be fully utilized at that radius, but it would be useful up to about 8000m.
23:08 Unarelith  `singlenode` map gen isn't supposed to add one node to prevent the player from falling?
23:09 Unarelith  oh, no, not at all actually
23:32 VanessaE   nerzhul: are you available?
23:33 nerzhul    VanessaE: i will go to bed soon but ask if it's not too long :)
23:33 VanessaE   nerzhul: can you go over on #minetest and set Xiong straight regarding what is or is not official, and release delays on the play store?
23:33 VanessaE   (maybe check the log for the last 30 or so mins first)
23:35 nerzhul    it's a problematic user ?
23:35 VanessaE   no, just not believing the rest of us
23:35 VanessaE   since you're the play store package maintainer, maybe he'll listen to you
23:37 nerzhul    it seems i have nothing to say, it's the debate between proprietary vs free software
23:37 nerzhul    yes play is not free but it permits to have a good user base
23:37 nerzhul    and i upload the package you mentionned in the irc log, and yes 0.4.17.1-2 is 0.4.17.2 for android users
23:38 nerzhul    because i didn't triggered a useless version bump on the classic code :)
23:38 nerzhul    you can link that to him, i'm off now :)à
23:38 VanessaE   did.
23:49 p_gimeno   !tell Krock Please see https://notabug.org/pgimeno/Gists/commit/d91caa3 for a trivial fix for a compiler warning
23:49 ShadowBot  p_gimeno: O.K.
23:57 p_gimeno   !tell Krock The warning applies to src/util/ieee_float.cpp
23:57 ShadowBot  p_gimeno: O.K.