Time  Nick       Message
02:10 hmmmm      too many questions, not enough answers :(
02:11 VanessaE   sounds about normal for here.
02:26 hmmmm      this is such a load of horseshit
02:26 hmmmm      i can't even begin to debug anything because LIBGL_DEBUG is broken
03:08 hmmmm      I GOT IT
03:08 hmmmm      omg
03:08 hmmmm      moral of the story is never link against GTK libraries before linking against libGL
06:48 paramat    game#523
06:49 est31      the bot is down :((((
06:49 paramat    wake up bot!
06:50 paramat    https://github.com/minetest/minetest_game/pull/523
06:51 paramat    hmmmm and all, as you might remember i had a hard time adding river water nodes to mtgame, so i would appreciate support for this =)
06:52 hmmmm      is it just rivers you want to add the turquoise hue to?
06:53 paramat    yes, just default:river_water..
06:53 hmmmm      hrm
06:54 paramat    details and reasons in description
06:55 hmmmm      no i know
06:55 hmmmm      i was just thinking of something
06:55 hmmmm      I approve of it, +1
06:55 est31      ok next pr
06:55 est31      https://github.com/minetest/minetest_game/pull/522
06:55 est31      general attitude to this
06:55 est31      I'm neutral
06:55 hmmmm      obviously yes
06:55 paramat    thanks
06:55 hmmmm      trailing whitespace is literally hitler
06:56 hmmmm      +1 to 522
06:56 est31      he has a similar PR for the engine too
06:56 hmmmm      hrmm
06:56 hmmmm      actually you know what
06:56 hmmmm      lets' not merge it yet
06:56 * paramat  pounces on nore
06:57 * est31    looks up "pounce, to" in the dictionary
06:57 hmmmm      omnibus commits, i would argue, are frowned upon when not necessary because they either just sit there and rot, or they create merge conflicts
06:58 est31      that's been why I am neutral
06:58 hmmmm      they should be done one file at a time or not at all..
06:58 hmmmm      and they shouldn't interfere with other PRs
06:58 hmmmm      but there shouldn't be this many PRs to begin with
06:58 hmmmm      tomorrow's a saturday, do you guys want to have a PR grooming session?
06:58 paramat    hi nore https://github.com/minetest/minetest_game/pull/523
06:58 est31      hehe for me its 9 AM saturday...
06:59 hmmmm      ahh
06:59 hmmmm      well I'll be around all day tomorrow
07:00 hmmmm      in my opinion, we should start at the very back of the queue
07:01 hmmmm      not only will they be easier to decide on, but it's... much more fair
07:01 est31      good idea
07:01 nore       hi paramat
07:01 paramat    O/
07:02 nore       hmmmm: ok
07:02 hmmmm      with any luck we'll be able to get through a page of PRs.  which is roughly 25 or so
07:02 hmmmm      nore: ??
07:02 nore       for pr session
07:02 hmmmm      yea
07:02 nore       (just what time precisely?)
07:02 paramat    nore also this https://github.com/minetest/minetest_game/pull/520 can push it for you if you want
07:02 hmmmm      i have no idea
07:02 hmmmm      everybody here is in a different time zone
07:02 paramat    i'll join channel when i wake up later
07:02 hmmmm      it's a good time for est, for example, but it's 3 AM for me
07:03 hmmmm      and paramat is like 8 AM iirc
07:03 paramat    yes 8am sat but it's still fri night for me =)
07:03 nore       for me it is right now 9 AM, and I won't have time until 17 PM today :(
07:03 Calinou    those are my PRs: https://github.com/minetest/minetest_game/pulls/Calinou
07:03 nore       just 12 hour now
07:03 hmmmm      ahh
07:04 hmmmm      yea no problem
07:04 Calinou    it looks like only https://github.com/minetest/minetest_game/pull/518 is worth merging, judging by comments
07:04 hmmmm      minetest_game PRs are more difficult because they're mostly preferences
07:04 hmmmm      i think we need a handful of people in charge of minetest_game who have an actual direction in mind
07:05 hmmmm      and care about it :)
07:06 hmmmm      hmmmmmmm... Calinou, would you like to be a minetest_game developer?
07:06 OldCoder   Isn't Carbone a _game fork?
07:07 hmmmm      also, I thought Nore was a committer, guess not
07:07 hmmmm      no idea what Carbone is
07:07 nore       hmmmm: I am
07:07 hmmmm      you're novatux?
07:07 nore       yep
07:07 hmmmm      ahh okay
07:07 OldCoder   He already is a _game developer, hmmmm; he forked minetest_game.
07:07 hmmmm      I have no idea man
07:07 OldCoder   Carbone is a minetest_game variation; I use it myself. It is fine; just filling in the gap.
07:08 hmmmm      do the people in charge of minetest_game want minetest_game to go in the Carbone direction?
07:08 OldCoder   I doubt it
07:08 OldCoder   Some optimizations, yes
07:08 OldCoder   But he is developing a somewhat different _game
07:08 OldCoder   Faster, which is positive, but different style
07:09 hmmmm      i personally feel that minetest_game should stay somewhat conservative because it is the default bundled game
07:09 OldCoder   I would agree
07:09 hmmmm      i'm somewhat disappointed nobody ever packaged different games with minetest distributions though
07:09 OldCoder   I use Carbone because it is said to be somewhat server friendly
07:09 OldCoder   I have
07:09 OldCoder   I rather like
07:09 hmmmm      well
07:09 OldCoder   the idea of a basic default package
07:09 est31      some people do package
07:10 OldCoder   and a comprehensive collection
07:10 OldCoder   Then people can choose
07:10 hmmmm      i think there needs to be a variety of game types
07:10 est31      I think that would be a good idea
07:10 hmmmm      minetest_game is the basic minecraft-like
07:10 OldCoder   Vanessa's Dreambuilder for example is the upscale version
07:10 est31      without extras like mobs
07:10 OldCoder   Yes; basic
07:10 OldCoder   Then larger collections with different flavors
07:10 OldCoder   Or even different goals
07:11 OldCoder   Hungry Games comes to mind as a candidate for the PVP _game
07:11 nore       well, what we need it really at least more default games
07:11 hmmmm      are there any totally different genres
07:11 hmmmm      like any RPGs?
07:11 hmmmm      or how about platformers?
07:11 OldCoder   Yes
07:11 OldCoder   One is very different
07:11 hmmmm      some of those too
07:11 OldCoder   Totally unique approach
07:11 hmmmm      celeron had a dungeon game
07:11 hmmmm      throw that on
07:11 OldCoder   Taoki and I will make Ghostworld
07:11 hmmmm      s/on/in/
07:11 hmmmm      strategy games even
07:11 OldCoder   Entire play is different
07:12 hmmmm      we want to show people that this really isn't just some minecraft clone
07:12 OldCoder   There is no distinction between players and mobs; players possess animals or people
07:12 OldCoder   Until death
07:12 hmmmm      need to display the flexibility we claim minetest has
07:12 hmmmm      making it better
07:12 OldCoder   I will ask Taoki to stop by sometime. His ghostworld, Vanessa's Dreambuilder, Hungry Games, basic, Moontest
07:12 OldCoder   These are quite a range
07:13 hmmmm      I thought moontest was an RBA game
07:13 OldCoder   And all can be packaged easily as one download
07:13 OldCoder   Hm? No...
07:13 hmmmm      hmm
07:13 OldCoder   It was started by Amaz
07:13 OldCoder   and a kid; we hope to make it Solartest
07:13 OldCoder   Where you can actually fly to all the planets
07:13 OldCoder   With a special mapgen
07:13 OldCoder   That puts Venus, Mars, Earth at different levels
07:13 nore       maybe Moontest would benefit from VAE (adding some kind of ship :))
07:13 OldCoder   We have UFOs
07:14 OldCoder   Not ready for general use; I am waiting for Windhero (sp?)
07:14 OldCoder   to finish classes and work on the mapgen
07:14 OldCoder   But this will be quite unique. You will be able to physically fly
07:14 paramat    i think moontest was partially forked from my moonrealm mod
07:14 OldCoder   I believe it was
07:14 nore       anyway, some diversity would be good
07:14 OldCoder   More than partially; you would need to ask Amaz
07:15 OldCoder   Proposal: default _game (basic), Vanessa Dreamworld, Hungry Games (my variation or Gunship Penguin's), Moontest (with other worlds added)
07:15 paramat    it's been developed a lot though
07:15 OldCoder   And Taoki Ghostworld
07:15 OldCoder   paramat, moonrealm?
07:15 paramat    yes
07:15 OldCoder   If you can add provisions for other worlds, I will be delighted to rest
07:15 OldCoder   Imagine the fun
07:16 OldCoder   Different skybox for each world, physical travel between them
07:16 OldCoder   ^ Not a bad concept
07:16 paramat    i'd love to, i have several lifetimes of ideas for MT
07:16 OldCoder   Let us PM sometime
07:16 OldCoder   Taoki's idea is the most original by far; he will explain it in due course
07:16 paramat    sure
07:17 OldCoder   Everybody is a phantom and you must possess NPCs
07:18 paramat    oh sorry, i don't mean moonrealm has been devved a lot, i mean moontest has. moonrealm needs my attention, so much potential..
07:20 OldCoder   paramat, either way. We will see what works. If Windhero does not return, I can attempt to learn mapgen and will ask you about this.
07:20 paramat    stacked planet realms, each with differing gravity, skyboxes and atmosphere
07:20 OldCoder   Exactly ^
07:21 OldCoder   And travel by flight as well as travelnet between them!
07:21 paramat    yes
07:21 OldCoder   This is important; it will open up all sorts of possibilities
07:21 * OldCoder imagines little starships, Tie Fighters, etc. Over Time. Need not to bite off too much to chew.
07:21 OldCoder   As Lego grew
07:21 OldCoder   Lego, the *original* voxel world
07:22 OldCoder   So can this one
07:22 paramat    i get overwhelmed by the potential of MT
07:22 * paramat  loves LEGO
07:22 OldCoder   Then it is simple
07:22 OldCoder   One step at a time
07:22 OldCoder   We need lego support BTW
07:22 OldCoder   Though we must call it something else
07:23 OldCoder   hmmmm, excuse any interruption. Wanted to say that a collection of novel _games makes perfect sense.
07:25 OldCoder   paramat, I haven't Githubbed Moontest yet. It is rather different now than Amaz's copy. I'd like to post or email you the current sources sometime.
07:26 paramat    okay, let's arrange this later
07:27 paramat    nore, thanks for the comments, i'll push the sandstone brick texture later, assuming that's okay to do
07:28 paramat    (if no-one else does it first)
07:28 nore       ok :)
07:49 Calinou    <+hmmmm> hmmmmmmm... Calinou, would you like to be a minetest_game developer?
07:49 Calinou    well, why not
07:49 Calinou    <+hmmmm> no idea what Carbone is
07:50 Calinou    it's a (no longer actively maintained) subgame by me
07:50 Calinou    it tweaked a minetest_game (based on 0.4.9) a lot, back when it was really lacking
07:50 Calinou    I now consider minetest_game to be decent, so I don't maintain it much
07:51 Calinou    <+est31> without extras like mobs
07:51 Calinou    we should have stock mobs somewhen.
07:51 Calinou    something basic, like we had in 0.3.1
07:52 Calinou    it was enough to scare people and have something to make food of
07:52 Calinou    we don't need more than 4-5 mobs for now
13:02 Krock      No longer red screen caused by the environment when damage is disabled: https://github.com/minetest/minetest/pull/2740
13:55 rubenwardy http://irc.minetest.ru/minetest-dev/2015-05-30#i_4270443
13:55 rubenwardy hmmmm
13:55 rubenwardy Jump is probably the closest to a platformer
13:55 rubenwardy https://forum.minetest.net/viewtopic.php?t=9326
13:55 rubenwardy Well, not really
15:28 daswort    Was there a change in dependencies needed to compile MT? I installed the dependencies required, but after the recent git sync it failes.
15:28 daswort    http://paste.ubuntu.com/11454597/
15:40 sfan5      damnit est
15:40 sfan5      daswort: install gmp in the meantimer
15:40 sfan5      --r
15:40 sfan5      keyboard pls
15:46 sfan5      wait
15:47 sfan5      daswort: that shouldn't happen, minetest should only try to use gmp when it's installed
15:48 daswort    Which one? libgmp10-dev or libgmp3-dev
15:48 sfan5      no idea
15:49 daswort    libgmp10-dev  is a virtual package anyway.
15:49 jin_xi     or you could edit src/util/srp.cpp
15:53 hmmmm      instead of working around the symptom, maybe we should figure out why USE_SYSTEM_GMP is set when it shouldn't be
15:53 hmmmm      daswort, you're not compiling for android, are you?
15:55 daswort    Nope amd64. hmmmm
15:58 daswort    So with gmp it does that. But the is more fun to have with failing builds! http://paste.ubuntu.com/11455040/
15:59 hmmmm      session manager?
15:59 hmmmm      do you not have libsm6?
16:00 daswort    ls /usr/lib/x86_64-linux-gnu/libSM.so.6*
16:01 daswort    d
16:01 daswort    /usr/lib/x86_64-linux-gnu/libSM.so.6@  /usr/lib/x86_64-linux-gnu/libSM.so.6.0.1
16:01 daswort    So it exists
16:01 daswort    Session Manager is GDM. hmmmm
16:02 hmmmm      i have no idea why this is required for minetest, and why it's failing
16:02 hmmmm      how did you call cmake?
16:02 hmmmm      rather, execute cmake
16:03 daswort    cd ~/.minetest-dev &&  time cmake . -DRUN_IN_PLACE=1 -DENABLE_GETTEXT=1 && time make -j2
16:04 hmmmm      try make VERBOSE=1
16:04 hmmmm      and then paste the output
16:08 daswort    nothing enlightening, or did i pass the option in a wrong way? http://paste.ubuntu.com/11455231/ hmmmm
16:09 hmmmm      no no, without any -
16:09 hmmmm      just VERBOSE=1
16:09 hmmmm      and add that to make's arguments, not cmake
16:13 hmmmm      actually it seems like libSM gets included along with irrlicht
16:13 hmmmm      makes sense, I suppose...
16:13 daswort    still in the dark. hmmmm  http://paste.ubuntu.com/11455359/
16:15 daswort    But i didn't need it a few days ago with the same Debian Version but XFCE instead of GNOME.
16:15 hmmmm      well the reason why it's saying there's no rule to make the target is because it really doesn't have a rule for it
16:15 daswort    And Debian does not have any fancy wayland stuff and unpriviledged Xorg yet.
16:15 hmmmm      it only attempts to make the target if the target did not exist
16:16 hmmmm      check the permissions on those directories
16:16 hmmmm      both the symlink and the actual file
16:17 daswort    On libSM? the link: lrwxrwxrwx | the file: -rw-r--r--
16:17 hmmmm      oh waita  minute
16:17 hmmmm      do you have libSM.so
16:17 hmmmm      not libSM.so.6
16:18 daswort    Oh, no please not again. Indeed there is no file without number suffix.
16:18 hmmmm      well that's the immediate problem
16:18 hmmmm      you can fix that and build minetest by creating a symlink to .so.6, but that's not really solving anything
16:19 hmmmm      is it possible that you're missing a couple -dev packages?
16:20 daswort    Yeah, linking works, but then comes libICE over the hill and then the next.
16:20 daswort    hmmmm, it is a new system so it is possible. Will check.
16:21 hmmmm      try apt-get build-dep mana
16:21 hmmmm      erm
16:21 hmmmm      mesa
16:21 hmmmm      rather no, libgl1-mesa-dev
16:22 rubenwardy build-essentials ?
16:25 daswort    libgl1-mesa-dev is installed. Everything else from the wiki. Except for libjpeg, but that was renamed some time ago (in jessie). jpeg turbo dev is installed.
16:26 daswort    build-essential is therefore also installed.
16:31 daswort    So i installed the dependencies for mesa and linked also libICE and it appears to compile without errors.
16:31 hmmmm      sweet
16:31 hmmmm      now just hopefully it won't crash when you start it up
16:32 daswort    Compile from last week did often, directly after starting. But never while gameplay. :)
16:33 hmmmm      i'm joking btw, it shouldn't crash
16:34 daswort    But it did.
16:36 daswort    But it were the fastest crashes i ever experienced, so there is that. :P
16:41 daswort    Did make clean && make and it works great. A wall of wonderful green text. Thank you very much hmmmm
16:41 hmmmm      np
16:41 hmmmm      sorry that minetest is like a broken car
16:42 hmmmm      i'm not in charge of the debian packaging and i have no idea who is, so it's missing a dependency and i can't really do anything about it
16:47 daswort    I didn't bother to try the debian Package.
16:48 daswort    But i really like that there are no compiling warnings like in MOST games.
16:52 Calinou    -Wnone fixes all
16:56 daswort    Oh :(
16:59 TenPlus1   hi folks
16:59 TenPlus1   ran a gdb session on Xanadu server and it crashed, we have an output
17:03 hmmmm      the out of memory crash?
17:04 TenPlus1   std::bad_alloc error crash, we got this:  https://github.com/minetest/minetest/issues/2661  (LAST POST)
17:04 hmmmm      maybe it would be better if you were able to run massif
17:04 TenPlus1   massif ?
17:05 hmmmm      yes, check it out: valgrind --tool=massif
17:07 TenPlus1   gotta find new time to bring server down and run that, we just finished gdb instance
17:07 TenPlus1   server is back up and running normally again
17:51 TenPlus1   hmmmm, valgrind runs far too slowly on server to be usable for players to crash server
18:04 hmmmm      TenPlus1, how slowly valgrind runs highly depends on which tool you are using
18:04 hmmmm      are you sure you're using massif?
18:06 TenPlus1   yes hmmm... typed the command u showed me and server runs sooo slow I cant do anything on it let alone have players join
18:14 hmmmm      hmm
18:15 TenPlus1   least gdb caught that one error and filename it was in: https://github.com/minetest/minetest/issues/2661
18:15 hmmmm      yeah
18:15 hmmmm      that output is literally useless though
18:16 TenPlus1   was all it gave me...
18:16 TenPlus1   was expecting a huge error report
18:16 hmmmm      maybe run it under gdb again, and then when it crashes, don't just copy the output and close it
18:16 hmmmm      instead leave it open and then I'll have you enter in some commands
18:17 TenPlus1   will do that next time owner is aroun
18:17 TenPlus1   d
18:17 hmmmm      great
18:24 celeron55  probably setting up a breakpoint to catch the segfault earlier could work, altough sometimes things still just become a horrible mess
18:24 TenPlus1   hmmmm,  which commands shoudl I enter, so I know
18:25 celeron55  running something like "bt __cxa_throw" in gdb before "run" could work
18:25 hmmmm      TenPlus1, I'm really not sure right now
18:25 celeron55  that would catch any exception before they are handled
18:25 hmmmm      celeron, the exception being thrown isn't the issue
18:25 celeron55  but you'll have to call "continue" every time an exception happens that you don't care about
18:25 celeron55  hmmmm: i know
18:26 hmmmm      so then why bother saying that
18:26 TenPlus1   server is overseas and owner isnt tech savvy so ti would be difficult to get an output
18:26 celeron55  hmmmm: by catching it earlier than the abort of the program, you might end up with a better backtrace
18:26 hmmmm      this can be done in real-time thanks to the internet
18:26 celeron55  actually i guess the backtrace might have been there
18:27 hmmmm      the backtrace is useless
18:27 celeron55  what would you be looking for?
18:27 hmmmm      i want to take a look at some random pages allocated in the application and try to guess what the contents might be
18:27 hmmmm      I've done this before, but it was a lot easier to find the source of the leak because the beginning of every page looked like a PE file :)
18:28 celeron55  the thing is, this isn't a leak, right?
18:28 hmmmm      this is definitely a leak
18:28 TenPlus1   at times the server will crash every few minutes and we think it's a map error causing problems...
18:28 celeron55  ok if it is, then your approach is the best one
18:28 TenPlus1   but during gdb test we had 8 players and it wouldnt vcrash for a while... 8 mins later it vrashed
18:29 hmmmm      what would be the best approach would be to use massif but it's too slow
18:29 celeron55  if it isn't, then the backtrace would be where to first look for things
18:29 hmmmm      so i have to do this instead
18:29 hmmmm      the first backtrace is literally anywhere memory attempts to be allocated
18:29 hmmmm      memory gets allocated constantly
18:29 hmmmm      if we pick random pages to look at, though, there's a high likelyhood it's going to contain leaked memory
18:30 TenPlus1   still think problem lies in the map... loading a specific area causes std::bad_alloc to happen... it's happened to often in specific areas
18:30 celeron55  (well if you get an std::bad_alloc while all memory isn't actually being consumed, it's usually some stupid bug causing one huge allocated size that nevertheless would be very hard to find otherwise)
18:31 TenPlus1   ah
18:31 celeron55  i've run into a bunch of those too; it's stupid to leave that out as being too stupid to happen
18:32 celeron55  i'm going to assume something like that is happening until i see a backtrace with the size being allocated shown that proves otherwise
18:33 hmmmm      i've seen the backtrace before
18:33 hmmmm      it's usually totally random
18:34 celeron55  if you have, then go ahead with assuming it's a leak
18:34 TenPlus1   the fact that many servers are running fine using 0.4.12, but when the map gets to a certain size the crashing begins... I dont get it
18:34 TenPlus1   was running great, then the map got over 50gb in size...
18:36 celeron55  or actually
18:37 celeron55  don't assume it's a leak in the valgrind memcheck meaning of the word
18:37 celeron55  it could simply be memory that isn't leaked but is not being freed at that moment because it's simply kept around for future use
18:37 hmmmm      yea that's why i suggested massif
18:37 hmmmm      massif is usually more useful than memcheck...
18:37 TenPlus1   and running the server using massif is so slow it's unusable
18:58 TenPlus1   bye all
20:30 est31      pushing in 5 minutes https://github.com/est31/minetest/commit/b4dfaa3a7a9251cdb0f0867fc29240c808268b1b
20:30 est31      (non functional part of Krock's PR)
20:30 est31      also not related
20:35 est31      pushed
21:06 jin_xi     so its probably not related, but here goes anyway. we're leaking an u32 for every expiring particle spawner. spawner ids are only removed for deleted ones, expiration is not handled.
21:07 est31      ok
21:08 est31      when you make a simple enough PR, I'll merge it right away
21:08 jin_xi     pr made long time ago, not sure if best/proper way to handle. #2602