Time  Nick          Message
00:12 iqualfragile  kahrl: while this is not a bugfix it would make sense in the context of a release: https://github.com/minetest/minetest/pull/743
00:13 PilzAdam      what is the compression rate of this?
00:14 PilzAdam      and how much absoulte memory is saved?
00:42 iqualfragile  should be around 30 %
00:43 iqualfragile  and there is no memory saved as it get unpacked before it can be displayed
00:47 kahrl         this was meant for here: https://github.com/minetest/minetest/pull/739 I can confirm the bug and the fix works, any objections?
02:27 VanessaE      what's the story on clearobjects?
02:28 VanessaE      my server's starting to fail under the overload of entities I can't delete because clearobjects takes too much RAM to run
02:29 kaeza         VanessaE, try PilzAdam's "clean" mod for now
02:30 kaeza         (until that is fixed in core I mean)
02:30 VanessaE      no good, that requires knowing exactly the names of the objects to delete.
02:30 Exio          do clearobjects, take a nap
02:31 Exio          and then maybe some day it will finish
02:31 ShadowNinja   VanessaE: Modify it to delete all of them.
02:31 Exio          the solution for now ^
02:31 VanessaE      Exio: 400-some MB map, needs some 13GB of RAM to do that, which sends my machine off into lala-swap-land (as I have "only" 16GB, and a few GB is used normally by other stuff)
02:32 Exio          exactly VanessaE
02:32 Exio          swap = slow as hell, so clearobjects will be even slower
02:32 VanessaE      yup
02:32 VanessaE      people are starting to complain about it "raining gravel" on my server, this is an old bug that should have been fixed months ago.
02:33 Exio          what is that bug?
02:33 VanessaE      duplication and relocation of entities
02:33 Exio          ah
02:33 VanessaE      if there are enough in an area, they get teleported to some other, random location
02:34 VanessaE      if it's been fixed already, it obvious wasn't fixed enough :)
02:34 Exio          s/enough // i guess
02:34 VanessaE      heh
02:34 Exio          ;P
02:39 VanessaE      jeez my typing sucks.
02:42 sweetbomber   where does /clearobjects get stuck? in m_map->listAllLoadableBlocks?
02:42 VanessaE      dunno
02:42 VanessaE      it's not that it gets stuck, afaik, it works fine other than using enormous amounts of RAM
02:42 sweetbomber   i could do a quick (ugly) fix by only clearing a fixed (small) amount of objects
02:43 sweetbomber   but that would only fix it if the problem was in the clearing itself...
02:43 VanessaE      what I don't understand is why it doesn't just load and clear one chunk/mapblock at a time, and delete that mapblock from memory if it wasn't already in use.
02:43 sweetbomber   "fix it"
02:44 VanessaE      you already have a table/array/something describing which blocks are loaded, why not just consult that data structure when deciding whether to throw away a loaded-and-cleared block?
02:47 VanessaE      imho this is an example of a critical bug that should be fixed before the feature freeze ends.
02:47 VanessaE      as clearobjects simply cannot be used on anything but the smallest maps without rendering the server totally dead for perhaps hours at a time.
02:49 Exio          for hours unless you have free ram
02:49 Exio          it was quite fast here with a 200~220 mb map
02:49 VanessaE      that's if you have free RAM - I have 16GB, which is WAY above what most machines have, and it's not enough.
02:50 Exio          exactly
03:07 sweetbomber   vanessa, if it wasnt so late here, and i handt to catch a train by tomorrow, i would try to fix that now...but that needs to be carefully done, as you would "deploy" it afterwards in your server.
03:07 VanessaE      *nod*
03:07 VanessaE      I guess there's some plan to fix it, but when and how...who knows
03:08 sweetbomber   if no one realeases a quick fix by tomorrow, ill try do do something
03:08 VanessaE      thanks
03:09 sweetbomber   the simplest solution that comes to my mind is just clearing a few objects from the already loaded blocks
03:09 sweetbomber   and forget about the others
03:09 VanessaE      that would help to some degree, and I think hmmmm had an idea to that end
03:09 VanessaE      I think the idea was to have several modes - all blocks, only the loaded ones, and some others.
03:10 sweetbomber   yeah, i think that idea is good
03:10 sweetbomber   well, but ive to go
03:10 VanessaE      night
03:10 sweetbomber   4:10 am here
03:12 * sweetbomber goes to bed
03:19 * sweetbomber couldn't go to bed
03:19 sweetbomber   VanessaE, are you still there?
03:19 VanessaE      yeh
03:19 sweetbomber   couldn't resist to fix it
03:20 sweetbomber   i made a quick fix, i *think* it will work, and only clear objects on loaded blocks
03:20 sweetbomber   please, test it first on a dummy server
03:21 sweetbomber   ill push it to /sweetbomber/minetest
03:21 VanessaE      I'll test tomorrow
03:23 sweetbomber   ok
03:24 sweetbomber   for reference: https://github.com/sweetbomber/minetest/tree/clearobjects
03:26 sweetbomber   if it works, please let me know
09:27 kahrl         good morning
09:28 kahrl         I will push https://github.com/minetest/minetest/pull/739 in half an hour if no one objects; brb
09:42 RealBadAngel  its a fix, go on with it
09:42 RealBadAngel  we shall not ask each other for obvious bug fixes
09:43 RealBadAngel  new features, changes, of course
09:43 RealBadAngel  but not killing the bugs
09:45 RealBadAngel  we do have ways to revert the changes if somebody will dare to abuse it
09:45 RealBadAngel  bugfixing shall have shortest path
09:47 RealBadAngel  ImQ009, czesc. co cie sprowadza w to dziwne miejsce?
09:48 ImQ009        RealBadAngel, Dość spóźnione pytanie biorąc pod uwagę fakt, iż uczęszczam na ten kanał już od dłuższego czasu :P Po prostu śledzę jak ten projekt się ma i jak się rozwija
09:49 RealBadAngel  haha, wyszlo na to zem dupa i nie patrze kto podglada ;)
09:49 Jordach       english?
09:50 ImQ009        Jordach, Poop
09:50 ImQ009        Highly elaborated version of that word
09:50 RealBadAngel  Jordach, allow a few other words, hes polish as me
09:50 * Jordach     googles
09:50 RealBadAngel  hehe
09:50 ImQ009        Google Translate is evil
09:50 ImQ009        Everytime you use it a kitten dies
09:51 RealBadAngel  lol
09:51 Jordach       still better than bing
09:51 RealBadAngel  ImQ009, jakas ocena tego co sie tu dzieje?
09:52 ImQ009        Jeśli pytasz się co się dzieje na TYM kanale to Ci powiem, że jest to jeden z bardziej ogarniętych 'dev-kanałów' na jakich byłem :P
09:52 RealBadAngel  czemu tak sadzisz?
09:53 ImQ009        Ludzie ściśle mówią o konkretach związanych z samym projektem, dużo offtopicu tu nie widziałem
09:53 ImQ009        Ale może po prostu chodzi o to, że NIE widziałem :P
09:54 RealBadAngel  ja sie "patrze" od prawie roku
09:54 RealBadAngel  offtopicu tu raczej nie ma
09:55 RealBadAngel  ostatnio jest tylko mala wojna pomiedzy PA a mna
09:55 RealBadAngel  co tylko zaczniemy dyskutowac to idzie na noze
09:56 ImQ009        heh
09:56 RealBadAngel  PA chce zeby to byl minecraft
09:56 RealBadAngel  ja tego nie chce
09:56 ImQ009        A w jakÄ… stronÄ™ idziecie?
09:56 ImQ009        Jak chcecie, żeby projekt wyglądał
09:56 RealBadAngel  a cholera wie :)
09:57 ImQ009        Coś jak FTB/Tekkit dla Minecrafta może? :3
09:57 RealBadAngel  niektorzy chca 16px jako swietosc
09:57 RealBadAngel  ja chce shadery i textury minimum 128x
09:57 RealBadAngel  i oczywiscie technic + mesecons
09:58 ImQ009        I FILTROWANIE BILINEARNE, GOD DAMNIT
09:58 RealBadAngel  musi byc
09:58 Jordach       kto polak -> #minetest
09:58 RealBadAngel  bumpmapping juz dziala
09:58 RealBadAngel  widziales?
09:58 ImQ009        Nope
09:59 RealBadAngel  sek
09:59 ImQ009        A będzie implemetacja deferred shading?
09:59 ImQ009        Czy oldschool?
09:59 RealBadAngel  http://imgur.com/a/Eq09o#0
09:59 ImQ009        Z czystej ciekawości
09:59 RealBadAngel  na razie jest to czesciowa implementacja
10:00 ImQ009        Z deferred shading fajnie może pojechać z dynamicznym oświetleniem
10:00 RealBadAngel  bo nie mamy hardware lights
10:00 RealBadAngel  kuzwa nawet slonca nie mamy
10:01 RealBadAngel  a bez tego i hercules dupa
10:02 ImQ009        Why don't we speak english? Just to be less annoying for other people :P
10:02 kaeza         translate.google.com pls
10:02 kaeza         :P
10:08 RealBadAngel  ok, lets stop
10:09 RealBadAngel  it was just nice for me to talk in polish for a moment here
10:09 kahrl         RealBadAngel: I wouldn't say that bugfixes are obvious
10:09 RealBadAngel  kahrl, why?
10:09 kahrl         for example the original version of that pull request had a badly placed NO_MAP_LOCK_REQUIRED that sapier saw but I probably would have missed
10:10 RealBadAngel  same happened for me with a drawtype
10:10 RealBadAngel  but this assumes another one should compile and check the pull
10:10 RealBadAngel  not only just read it
10:10 kahrl         at least with a drawtype you can sometimes see what's wrong
10:11 kahrl         actually reading is often enough
10:11 RealBadAngel  funny was that i didnt
10:11 RealBadAngel  it popped up on a developed world
10:11 RealBadAngel  on a brand new, testing one everything was ok
10:12 RealBadAngel  i put new glass drawtype in the middle of enum list
10:12 kahrl         ah that one
10:12 RealBadAngel  so half of the drawtypes were shifted
10:13 kahrl         yeah things like that you don't see if you're only testing singleplayer
10:13 kahrl         but you can't test everything on a big multiplayer server
10:18 RealBadAngel  neither you can test everythin in singleplayer mode
11:11 PilzAdam      thexyz, when do we merge the translations for 0.4.7?
11:15 thexyz        PilzAdam: any time
11:15 thexyz        do you think I should do that now?
11:17 PilzAdam      yea
11:17 PilzAdam      so we dont have to merge everything right before releasing
11:18 thexyz        sure
11:19 thexyz        done
11:19 PilzAdam      /home/adam/Minetest/minetest/po/fr/minetest.po:470: `msgid' and `msgstr' entries do not both end with '\n'
11:20 Jordach       PilzAdam, its an obvious fix
11:20 PilzAdam      yes, but it makes built fail
11:21 thexyz        it does?
11:21 PilzAdam      "found 1 fatal error"
11:22 thexyz        what a stupid msgfmt!
11:22 thexyz        okay, I'll fix it
11:24 thexyz        done
11:24 thexyz        that's strange
11:24 thexyz        there definitely is a check for this in weblate
11:25 thexyz        someone probably ignored it
12:02 proller       why gettext not found by default?
12:06 proller       on bsd need to enable manually
12:11 kaeza         proller, that could be enabled for post-0.4.7
12:12 * kaeza       wonders why it isn't enabled by default too
12:45 kahrl         https://github.com/minetest/minetest/pull/740 objections?
12:51 kaeza         seems reasonable
12:54 kaeza         I've seen that README.txt and minetest.conf.example point to c55.me/minetest/. shouldn't they point to minetest.net?
12:54 kahrl         it makes some API functions return nil in a number of new cases, but the kinds of mods that don't check for nil would have caused server crashes with the old system too I think
12:54 kaeza         indeed
12:54 kaeza         at least is crahes in a nice way :P
12:54 kaeza         +s
12:55 kahrl         agreed on changing the c55.me links
12:57 kaeza         there's also a link in lua_api.txt:3
12:57 kahrl         I don't see the link in README.txt, but I see ones in the manpages
12:58 kaeza         hmm... maybe not in readme
12:59 kaeza         also, lines 16-18 of lua_api.txt
12:59 kaeza         that should point to github issue tracker
13:01 kahrl         I think there have been contributions in the past via email by folks who don't like using git
13:01 kahrl         so I'd say point out that pull requests are the preferred way but keep the email address
13:01 kaeza         agreed
13:03 kaeza         but I guess it's celeron who has to decide whether or not he wants to keep his address there
13:04 kahrl         yeah
13:05 kahrl         https://github.com/minetest/minetest/pull/741 a simple lua stack indexing fix, objections?
15:29 kahrl         any other bugs that need to be fixed before the 0.4.7 release?
15:29 VanessaE      yeah, clearobjects :)
15:32 kahrl         it would have to be a hacky fix though... I don't see any proper revamp of clearobjects being coded in the next 32 hours
15:34 VanessaE      joking aside, the most irritating thing I can think that needs fixed is that glitch where closing a formspec or menu doesn't properly reset your focus back onto the world
15:35 VanessaE      (where you have to jiggle the mouse first to get the focus to reset, so the keyboard can work again)
15:35 Zeg9          Some mods depend on this behaviour currently
15:36 VanessaE      they can be fixed then.
15:36 kahrl         seriously? what mods are those?
15:36 VanessaE      this bug even happens while chatting via the command input.
15:36 Zeg9          Then we need a new callback when formspec is closed
15:36 VanessaE      t blah blah blah enter t blah bl---wait, the dialog didn't open---jiggle mouse t blah blah enter...
15:37 Zeg9          True that
15:37 Zeg9          Also when you right click a door with a bucket it just pours the liquid from the bucket which is quite irritating
15:37 Zeg9          --> https://github.com/minetest/minetest_game/pull/166
15:37 Zeg9          :P
15:38 VanessaE      Zeg9: a similar bug occurs with homedecor - right click a door while wielding a sign and the door doesn't always open/close
15:38 Zeg9          Actually, there are two ways to solve it. Per-mod, as I did in my pull request, or globally, but the second one would mess up a few things
15:39 VanessaE      kahrl: there's also some question about the true randomness of the node selection routine for abms
15:39 kahrl         VanessaE: most likely outside the scope of 0.4.7
15:39 VanessaE      (where the specific nodes being operated on seem to be heavily biased toward the end of the list rather than being evenly distributed)
15:40 VanessaE      kahrl: this results in flowers heavily favoring dandelions, which is a bug.
15:41 VanessaE      (the list being that which is passed via the nodenames param, particularly if the list is a group
15:43 kahrl         from the code I don't see why that would happen
15:43 kahrl         are the nodes that it is biased towards in multiple groups?
15:43 VanessaE      nope, a single group in this case
15:43 VanessaE      (group:flora=1)
15:44 kahrl         can I see the ABM definition?
15:44 VanessaE      PA and I went over the Lua part pretty thoroughly..
15:44 VanessaE      https://github.com/minetest/minetest_game/blob/master/mods/flowers/init.lua#L123
15:45 VanessaE      when my flowers mod generated them, they were distributed evenly, but I used discrete node names rather than groups.
15:48 kahrl         so does it happen with the discrete names and/or with group:flora?
15:49 VanessaE      with discrete names, e.g. my flowers via plants_lib, it didn't happen.  With that ^^^^ code, it heavily favors the end of the list of nodenames that are counted as part of that group
15:49 VanessaE      for example,
15:49 VanessaE      if I define a new flower and add it to the flora group, that new flower will tend to show up more than anything else does, even though it shouldn't.
15:51 kahrl         so is it based off node ID?
15:51 VanessaE      no idea.
15:52 VanessaE      flowers:cotton_plant was the one from that example (I've since discontinued it)
15:52 kahrl         are any of the involved nodes redefined at any point?
15:53 VanessaE      nope.
15:53 VanessaE      not in the case of cotton anyway - now, there was an alias for it pointing back to a very old node name
15:53 VanessaE      in fact I think all of the flowers have those aliases.
15:54 kahrl         I don't think aliases would matter
15:54 VanessaE      I wouldn't have thought so either.
16:01 kahrl         VanessaE: can you run this and paste the results? https://gist.github.com/kahrl/5690845
16:03 kaeza         lol barney
16:04 VanessaE      kahrl: actually not right now, I'm about to head out
16:06 VanessaE      bbl
17:29 rubenwardy    where is the register_node defined in https://github.com/minetest/minetest/tree/master/src/script/cpp_api ?
17:30 rubenwardy    or https://github.com/minetest/minetest/tree/master/src/script/cpp_api
17:30 rubenwardy    or https://github.com/minetest/minetest/tree/master/src/script I mean
17:30 kahrl         rubenwardy: l_register_item_raw
17:31 rubenwardy    ah thanks
17:32 rubenwardy    Where is it added to the minetest table?
17:34 kahrl         retval &= API_FCT(register_item_raw);
17:36 rubenwardy    Ah, nice
17:37 kahrl         oh my, I may have found a one-liner fix for the menu focus bug
17:37 kahrl         VanessaE: ^
17:38 kahrl         https://gist.github.com/kahrl/5691120
18:05 rubenwardy    What is the point of NO_MAP_LOCK_REQUIRED?
18:06 rubenwardy    it is only #define NO_MAP_LOCK_REQUIRED
18:06 rubenwardy    Is it meant as a comment?
18:07 Exio          no idea what it does, but afaik it should be meant for, "later", having special locks (for a long-future, multithreaded lua mods, being able to do more than one stuff at the same time)
18:59 whirm1        hi
19:00 whirm1        I'm getting this segfault when exiting to menu: http://paste.debian.net/8014/
19:06 kahrl         whirm1: stable or git?
19:06 whirm1        git master's HEAD
19:09 kahrl         can't really see what is causing this :|
19:10 kahrl         do you happen to have a debug build handy? so gdb would show which line caused the segfault
19:12 whirm1        let me rebuild it
19:18 whirm1        kahrl: is 1 -DCMAKE_BUILD_TYPE=debug the proper way to do it?
19:20 whirm1        kahrl: http://paste.debian.net/8016/
19:23 kahrl         yep, that's the proper way
19:27 whirm1        kahrl: if you need me to do anything else just let me know (this is a Debian SID amd64 using the privative AMD drivers)
19:28 kahrl         i'm looking through the code atm
19:28 kahrl         can't find anything suspicious though
19:30 whirm1        kahrl: it crashes either going out to menu or to OS
19:31 whirm1        and I have a modified minetest.conf, texture pack and two mods
19:31 kahrl         does it crash in a fresh world with no addon mods enabled?
19:38 whirm1        let me check
19:40 whirm1        kahrl: http://paste.debian.net/8020/ <-- playing with minimal in a fresh world with no mods enabled (modified config file and texture pack)
19:40 whirm1        Ill try with no config file
19:43 whirm1        it crashes too
19:44 sweetbomber   lol, its notch's birthday?
19:45 kahrl         whirm1: well I don't think I can help with this, sorry (and I also need to go)
19:45 whirm1        kahrl: np, thanks for trying :)
19:48 kahrl         in any case, if no one comes around who knows the solution leave an issue on github
19:48 kahrl         it may hang in there indefinitely if noone is able to debug it but it will at least show others that this has been seen before
19:49 kahrl         (and hopefully allow to find the bug with the combined debug info)
19:51 kahrl         bye everyone
22:43 BlockMen      i have reworked the windows section -> http://minetest.net/download because i noticed a few times something like that -> http://forum.minetest.net/viewtopic.php?id=6156
22:44 BlockMen      is everyone fine with that?
22:46 ShadowNinja   BlockMen: It doesn't look like that will help clear up that issue. Maybe it wouldbe better to just call the installer stable, it seems to work.
22:46 ShadowNinja   Although I don't think we need to list download links for such old versions there, maybe under an old versions link.
22:47 BlockMen      ShadowNinja, i guess that would be best. and yes, they are to much. but now they are a bit more away fro the last release (which helos IMO)
22:47 BlockMen      *-o +p
22:47 celeron55     make a page for those and link it there
22:48 BlockMen      celeron55, ok
22:48 celeron55     or whatever
22:48 celeron55     i don't think it's necessary at all
22:48 celeron55     also, 0.4.6 is so old it should really be under "older builds", but there's no newer release 8)
22:49 celeron55     (yet)
22:49 PilzAdam      I say remove the installer link
22:49 ShadowNinja   BlockMen: Underscores should be converted to spaces in the title on the right.
22:49 ShadowNinja   (Currently it is only done for one page)
22:50 BlockMen      new page "old builds" or "old versions"?
22:50 PilzAdam      versions
22:50 PilzAdam      or releases
22:51 celeron55     releases
22:53 BlockMen      ShadowNinja, ok. will change and send pull request (later)
23:00 BlockMen      done. http://minetest.net/download and http://minetest.net/old_releases
23:02 celeron55     the link is not good; it's too verbose and nobody will even read it
23:02 celeron55     (even if they are looking for it)
23:03 BlockMen      how about "old releases" with no more text then?
23:04 celeron55     hmm, it's kind of hard to make it look reasonable with it being the only green link on the whole page 8)
23:04 BlockMen      same problem at main page, and all internal links. should i change link color to same like external links?
23:05 celeron55     i think no; it's useful to know if a link links to an external site or not
23:06 celeron55     well, i changed it a bit now; maybe it's slightly better
23:06 BlockMen      celeron, i think the globe makes it kinda obvious which one is external :P
23:07 celeron55     oh, indeed
23:10 celeron55     hmm, i quickly tried it in firebug and i think the green link actually is better than an identically colored one
23:11 celeron55     for... some reason
23:13 BlockMen      maybe not an identical..but the green is crap
23:15 celeron55     feel free to try to think of something better :P
23:16 BlockMen      will do
23:16 sweetbomber   is it only for me, or creative inventory pages only turn after the turn-page sound has finished playing?
23:22 PilzAdam      sweetbomber, what turn page sound?
23:23 sweetbomber   sorry, lack of info
23:23 sweetbomber   vanessaE's game
23:23 VanessaE      unified inventory
23:23 sweetbomber   the creative inventory has a sound when each page is turned
23:24 sweetbomber   it seems (could be just slowness of my pc + coincidence) that the page only turns when the sound finishes playing
23:24 sweetbomber   that is why i asked if it only for me, what would confirm the first hypothesis
23:25 VanessaE      I've not noticed such behavior in mine, but then again the render-to-texture lag tends to hide anything else
23:25 sweetbomber   ok, so its only a coincidence
23:26 sweetbomber   btw, VanessaE, have you tried my quickfix?
23:26 VanessaE      not yet, I just got in
23:26 sweetbomber   please, backup your server if possible, im not sure it will not break anything...
23:27 VanessaE      <offtopic> if you haven't seen ST:Into Darkness, you're missing out.</offtopic>
23:27 VanessaE      sweetbomber: I'm inclined to wait, actually
23:28 sweetbomber   np
23:29 sweetbomber   btw, could you give me interact, if you please?
23:29 sapier        atm NO_MAP_LOCK_REQUIRED just traces how much time is used within that scope to have a rough number what amount of time we are locking map where we wouldn't need to lock it
23:29 Exio          any numbers? :P
23:30 VanessaE      done.
23:30 sweetbomber   ty
23:30 sweetbomber   ill build something when i'm able to manage so many new blocks for me
23:31 sweetbomber   minetest has a bunch of awesome blocks
23:31 Exio          minetest or VanessaE's server? :P
23:31 Exio          well, this goes to #mt
23:32 VanessaE      -->#minetest
23:34 Exio          #mt = #minetest, #mt-dev / #-dev = #minetest-dev for me ;P
23:39 BlockMen      celeron55, done. -> https://github.com/celeron55/minetest.net_dokutemplate/pull/8