Time  Nick       Message
01:32 paramat    Wuzzy please could you compile latest MT and see if #8075 is now fixed?
01:32 ShadowBot  https://github.com/minetest/minetest/issues/8075 -- Hudbars displays twice in 5.0
01:33 Wuzzy      OK, soonish
01:42 paramat    thanks :) i'm just eager to find out
01:55 hecks      paramat: re other channel; I've singled you out for the matrix incident, don't do *that* ever again
01:55 rubenwardy don't do what?
01:56 hecks      split hairs over vector/matrix math performance while not understanding the subject very well
01:57 hecks      anyway I'll go rebase 7626 now if there's still a chance to merge it because it's holding my work up as well
02:03 hecks      Also tell me what should I write in the license blob for new files if I don't give a heck about any credit
02:04 hecks      eg. leave blank, pretend celeron wrote it, or do something else
02:07 rubenwardy it's typical to allocate copyright to celeron if you don't want credit
02:11 paramat    on irc last night you were complaining about me partially based on untruths http://irc.minetest.net/minetest/2019-02-10#i_5493285 http://irc.minetest.net/minetest/2019-02-10#i_5493292 since 25th sept the delay on your PR has nothing to do with me, just clarifying that
02:12 paramat    smalljoker agreed we should avoid unnecessary rotation calculations, it wasn't just me who was concerned about that
02:21 paramat    your PR also broke mods, we're generally not breaking mods in 5.0.0, so i was right to object to that
02:23 rubenwardy #7626
02:23 ShadowBot  https://github.com/minetest/minetest/issues/7626 -- Entity lighting fix by hecktest
02:23 hecks      well it doesn't break mods anymore (still waiting for a demonstration of one affected entity in an existing mod), at the cost of another legacy oddity to scratch your head at while reading the docs
02:24 paramat    yes, it doesn't break mods now
02:39 paramat    i don't need to understand the subject well when there are unnecessary matrix calculations being done for the large majority of objects that don't need it. it's very reasonable to be concerned about that, to not be is sloppy
02:45 paramat    i reread the PR, i was very reasonable, had vallid concerns, no regrets at all. so i will do *that* again ;) please don't hold a grudge over a very minor disagreement
11:12 ANAND      I tried to cross-compile the 5.0.0-dev and got errors pertaining to std::mutex, MutexAutoLock, and std::thread
11:12 ANAND      mutable std::mutex m_mutex;
11:12 ANAND      oops
11:13 ANAND      /src/util/container.h:118:15: error: ‘mutex’ in namespace ‘std’ does not name a type mutable std::mutex m_mutex;
11:13 ANAND      /src/script/cpp_api/s_base.h:151:7: error: ‘thread’ in namespace ‘std’ does not name a type std::thread::id m_owning_thread;
11:13 ANAND      /src/client/mesh_generator_thread.h:78:3: error: ‘MutexAutoLock’ was not declared in this scope MutexAutoLock lock(m_mutex);
11:14 ANAND      The exact same errors for every single occurrence
11:14 ANAND      Am I missing something?
11:14 ANAND      Forgot to say, cross-compiling for Windows
11:15 ANAND      I could just grab a ready-made build, but I wanted to know if this can be fixed.
11:16 sfan5      your toolchain needs C++11 thread support
11:28 p_gimeno   how many threads would a mutable mutex mute if a mutable mutex could mute threads?
11:49 ANAND      sfan5: Ok, how do I do that? :)
11:50 sfan5      install a newer compiler
11:50 ANAND      Newer version of MinGW?
11:52 sfan5      yes
11:54 ANAND      Alright
12:25 ANAND      sfan5, I had installed MinGW through apt, and I'm already on the latest version for Ubuntu 18.04, which is 5.0.3-1
12:25 ANAND      Even the latest supported version doesn't support mutex and stuff?
12:27 sfan5      gcc 5.0 is from 2014
12:28 sfan5      travis windows builds use gcc 7.1, for comparison
12:29 sfan5      ANAND: https://packages.ubuntu.com/bionic/gcc-mingw-w64 ubuntu even has a recent enough version
12:30 sfan5      hm, looks like mingw and gcc use different versioning schemes
12:30 sfan5      your build should be working then
12:40 nerzhul    sfan5: aur/mingw-w64-gcc 8.2.0-3
12:40 nerzhul    it follows the gcc version scheme :)
12:40 sfan5      not on ubuntu
12:41 sfan5      https://packages.ubuntu.com/bionic/mingw-w64-common this thing is 5.0.3, whereever that number comes from
12:44 nerzhul    reading mingws-w64_5.0.3.orig.tar.bz2 COPYING shows: "Copyright (c) 2009 - 2013 by the mingw-w64 project"
14:18 rubenwardy You also need to use the Unix threads version of mingw
19:14 VanessaE   FINALLY!
19:14 VanessaE   https://github.com/minetest/minetest/issues/8220
19:19 sfan5      oof
19:19 VanessaE   sfan5: bad?
19:19 sfan5      segfaults in lighting calculations are kinda bad
19:20 sfan5      no matter how unlikely
19:20 VanessaE   is there enough info there to track it down, without debug symbols?
19:22 sfan5      not sure
19:22 VanessaE   btw, that's a Skyblock server.
19:22 VanessaE   so, singlenode mapgen
19:25 VanessaE   (I've just noted that in the first post)
19:33 VanessaE   and added the code snippet for that server's mapgen
19:38 paramat    aha, mod error
19:39 VanessaE   ?
19:39 VanessaE   (note, I did not write that code)
19:39 VanessaE   (so don't blame me :) )
19:40 VanessaE   nevertheless, if it's really a mod bug, isn't a segfault kinda a horribly ungraceful way to die? :)
19:40 rubenwardy a mod error shouldn't cause an error there
19:40 rubenwardy and the only way a mod should be able to cause a segfault is with LVM
19:41 rubenwardy actually, not even then
19:41 VanessaE   rubenwardy: did you see the code I posted?
19:41 rubenwardy no lol
19:41 VanessaE   it's in the issue.
19:42 rubenwardy you're missing a local for vdata
19:42 rubenwardy but obviously that's not the crash
19:43 VanessaE   you mean rnd is missing it :)
19:43 VanessaE   that ain't my code.
19:43 VanessaE   I've pointed him to this issue, though.
19:43 VanessaE   (so at least he can fix whatever's wrong with his code while you guys argue with the segfault)
19:46 paramat    added comment showing the mod error
19:46 Wuzzy      nerzhul: daily Weblate poke
19:46 VanessaE   paramat: "unconfirmed bug"?  um, a segfault is still a bug, whether triggered by a mod or not....
19:48 paramat    the only blocker not being attended to is #8203
19:48 ShadowBot  https://github.com/minetest/minetest/issues/8203 -- Unzip forces all files to lowercase
19:51 * VanessaE hands Wuzzy one of those multi-language phrase translation booklets ;)
19:51 Wuzzy      paramat: no
19:51 Wuzzy      https://github.com/minetest/minetest/issues/8075#issuecomment-462426669
19:55 paramat    hmm well, debatable. the crash is caused by the mod, but yes ideally the usage of lvm should have parameter checking to avoid causing the crash. ok i'll rename the issue to request parameter checking
19:56 VanessaE   no, the crash is apparently *triggered* by the mod, but is caused by an engine fault that should have resulted, instead, in a meaningful Lua backtrace that I could then hand-off to the modder to fix without bugging you guys.
19:56 paramat    sorry wuzzy i don't think that can block release, not a serious enough bug as explained. although i realise why you would object
19:57 Wuzzy      yeah right lets release with broken statbars...
19:58 paramat    it's not an engine bug, setting calc lighting to the whole emerged area as known to be simply impossible due to how lighting works, not due to a bug, so it needs parameter checking
19:59 Wuzzy      ???????????
19:59 Wuzzy      Hudbars displays twice in 5.0 #8075
19:59 ShadowBot  https://github.com/minetest/minetest/issues/8075 -- Hudbars displays twice in 5.0
19:59 VanessaE   dude, it segfaults.  lack of parameter checking in the engine would be a bug.
19:59 VanessaE   that's like saying a kernel panic isn't really a bug if it's triggered by some userland program that just happens to tickle the kernel the wrong way.
19:59 Wuzzy      I think we are talking about entirely different bugs?!
19:59 rubenwardy yes
19:59 paramat    hudbars works now as you know
20:00 VanessaE   Wuzzy: paramat is also talking about a bug I filed earlier.
20:00 Wuzzy      no, hudbars does ***not*** work, or otherwise it wouldnt happen in MCL2. MCL2 uses hudbars :-/
20:00 paramat    "lack of parameter checking in the engine would be a bug." yes as i just said
20:01 Wuzzy      must be a regression because it did work in 0.4.17
20:01 VanessaE   paramat: ergo, "unconfirmed bug" tag would be incorrect.
20:01 Wuzzy      rubenwardy, please help me out here... :(
20:01 rubenwardy reopening if the duplicated thing still happens
20:02 rubenwardy this is probably an ID issue, if removing and readding causes issues
20:02 paramat    lack of parameter checking is a bug, but lighting not working when set to emin/emax isn't
20:02 rubenwardy what happens if you add then remove?
20:02 rubenwardy rather than remove then add
20:03 Wuzzy      rubenwardy: the weird thing is. when i add hudbars in MTG it works but not in MCL2?
20:03 rubenwardy oh interesting
20:03 VanessaE   paramat: you know, the ironic thing about THAT ("not working") is that...er... the lighting DOES work.  I have never seen broken lighting on the server that produced the backtrace.
20:03 paramat    i mean: the hudbars mod works on it's own
20:03 VanessaE   paramat: so apparently the server process is somehow managing to get the lighting correct before it actually barfs.
20:03 VanessaE   (or at least it looks right :P )
20:04 Wuzzy      yes hudbars is its own mod but if hud elements randomly fail after updating to 5.0.0, thats a regression
20:04 paramat    yes i'll change it ..
20:04 Wuzzy      rubenwardy: I am not sure what you want me to do
20:04 rubenwardy I don't know really, this is a pain
20:05 rubenwardy paramat, it's a sign of a race condition or some weirdness, which is imo a worse bug than just not working at all
20:06 rubenwardy paramat: segfaults can be the sign of a security issue, if it's out of bounds writing etc
20:06 rubenwardy so even with an unchecked API like LVM, it's a good idea to avoid that
20:06 rubenwardy the question is whether it's a regression or not
20:07 VanessaE   I don't recall seeing segfaults before 0.4.17 (or ...1).
20:12 paramat    duplicated hudbars does not still happen, wuzzy confirms that's fixed, we need a new issue for MCL2 black hearts not appearing
20:13 paramat    i'll leave it up to other devs to decide if that should be a blocker or not
20:13 Wuzzy      maybe that's also duplicated hearts? red hearts drawn over bcakground hearts?
20:13 Wuzzy      hmmmm
20:13 Wuzzy      i could maybe move some HUD elements around to see if they are visible or not
20:17 paramat    well, please open a new issue for the MCL2 hudbar bug :)
20:22 paramat    with the duplicated hudbar bug,which bar was on top, normal or custom?
20:22 Wuzzy      no idea
20:23 paramat    rubenwardy ^
20:24 paramat    wuzzy you confirmed the bug so i though t you'd know
20:25 paramat    issue lacks detail or screenshots
20:27 paramat    anyway, remaining hudbar bugs are attended to by #7360
20:27 ShadowBot  https://github.com/minetest/minetest/issues/7360 -- Statbars: Reduce initial update interval by SmallJoker
20:33 paramat    game#2311 Shara
20:33 ShadowBot  https://github.com/minetest/minetest_game/issues/2311 -- Kelp, corals: Add missing footstep and dug sounds by paramat
20:38 Shara      Now I need to listen to some leaves.
20:40 paramat    hm maybe i should lower the gains as it's usually underwater ...
20:42 Shara      well, it should always be underwater really
20:43 Shara      I'm not really sure if they should have sounds to be honest
20:47 paramat    yeah, however all nodes underwater have sounds so, consistency
20:49 fwhcat     does the player makes bubbles sound while underwater ?
20:49 fwhcat     i forgot sorry
20:49 paramat    swimming sounds
20:49 paramat    slosh slosh
20:50 paramat    ok, i'll reduce the gains for a more underwater feel
20:51 fwhcat     audacity ?
20:51 paramat    nah in mtg sound parameters :)
21:10 paramat    Wuzzy could you test #7360 to see if it fixes the remaining hudbar bugs?
21:10 ShadowBot  https://github.com/minetest/minetest/issues/7360 -- Statbars: Reduce initial update interval by SmallJoker
21:10 Wuzzy      ooooh nice!
21:10 paramat    i'm going to test rubywarden's proposed lowercase fix
21:22 Wuzzy      oh, it was a race condition? that explains things
21:54 paramat    ruben's lowercase fix works for me, will make a PR
21:56 rubenwardy #8221
21:56 ShadowBot  https://github.com/minetest/minetest/issues/8221 -- Fix extract zip writing lowercase files by rubenwardy
22:04 paramat    oops
22:07 paramat    you beat me by 6 mins :)
22:07 rubenwardy techinically 2 hours
22:09 paramat    ^ needs testing on windows
22:26 paramat    nice #7360 fixes remaining hudbar bugs, made it a blocker
22:26 ShadowBot  https://github.com/minetest/minetest/issues/7360 -- Statbars: Reduce initial update interval by SmallJoker