Minetest logo

IRC log for #minetest-dev, 2015-07-27

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

All times shown according to UTC.

Time Nick Message
00:06 est hmmmm, how do you think should areastore do feature detection?
00:06 est e.g. when you have a mod and you want to utilize areastore, but use a slower implementation if its not available
00:14 hmmmm you mean in the case a mod is running on an older server?
00:15 est yes
00:15 est I have done it with if AreaStore which checks for the existence of the constructor I think
00:16 est but if it doesnt exist, it creates an "undeclared global" warning
00:16 hmmmm i think that's acceptable
00:20 est ok, can you +1 it, or is more work needed?
00:21 hmmmm uhhh
00:21 hmmmm what's the point of forEach?
00:22 est mostly serialisation.
00:22 hmmmm but i mean what does that function do
00:22 est it calls a passed function for every object of the areastore
00:23 hmmmm why do you need a separate function to do that?
00:24 hmmmm i have a feeling that you just want an excuse to use function pointers for some reason
00:24 est its the most general way
00:24 est e.g. one day that can be exported to lua
00:25 hmmmm well if you say so
00:25 est but I should comment it out
00:26 hmmmm I guess so
00:26 est and then I do what sn suggested, flattingout the area partly for the calls
00:26 est not for the returns though
00:26 hmmmm btw you never actually wrote the reason why you disabled serialization capabilities
00:27 est first, I haven't implemented forEach for the spatial store yet
00:27 hmmmm well write about it in the comments
00:27 est second, perhaps we should use libspatial version
00:27 hmmmm tell everybody why, not just me
00:27 est they have serialisation too
00:28 est ok I place a comment somewhere where they have to change stuff if they enable
00:33 paramat hmmmm do you approve of game#592 ? i'm not sure i need approval from the mtgame team but perhaps best i'm cautious and wait
00:33 ShadowBot https://github.com/minetes​t/minetest_game/issues/592 -- Default/mapgen: Disable nyancats in singlenode mapgen by paramat
00:33 hmmmm meh fine for me
00:33 hmmmm it makes sense
00:33 est can people enable them?
00:34 paramat i think so by calling the function in a mod
00:34 est also, I think we need a way of lua based unit tests too
00:34 est ok
00:35 est (unrelated to nyan cats :P)
00:40 bobomb est, so what, the textures folder is pre-scanned?
00:40 est yes bobomb.
00:42 bobomb ahhh.
00:42 bobomb thanks for clarifying
00:42 bobomb rubenwardy was saying that this is slated to change though?
00:43 est one day it will
00:50 bobomb i'd be just as happy with "draw" abilities in the formspec, not loading an image but being able to define primitives or a raster that never exists as a file
00:59 paramat left #minetest-dev
01:10 Zeno` joined #minetest-dev
01:10 est yaay zeno
01:11 Zeno` yaay? :D
01:12 est refactor more stuff, make more regressions :)
01:12 est #2959
01:12 ShadowBot https://github.com/minetest/minetest/issues/2959 -- Ability to launch singleplayer game from command line removed by 'refactor main.cpp'
01:13 Zeno` lol, it's being refactored again?
01:13 Zeno` oh wait
01:14 Zeno` sheesh, that took a long time to surface
01:15 Zeno` est, is that confirmed?
01:15 est I usually believe rubenwardy
01:15 est but no it isnt
01:15 est lemme see
01:16 Zeno` I've been launching singleplayer games from command line forever
01:17 est whats the command
01:17 Zeno` but by using an existing world/game. I wonder if it means launch and create
01:17 est minetest --worldname name doesnt work
01:17 est havent tested 0.4.10 though
01:18 Zeno` hmm ok
01:18 Zeno` https://github.com/Zeno-/mtutil​s/blob/master/massif_client.sh
01:18 Zeno` I see that
01:18 Zeno` I see that's multiplayer
01:19 Zeno` I'll investigate
01:20 Zeno` He must mean in conjunction with skip main menu?
01:20 est yes
01:20 est at least thats what I think he means
01:23 Zeno` ok, I
01:24 Zeno` ok, I need to reboot. Will bb in 5
01:25 Zeno` joined #minetest-dev
01:58 Zeno` what was the bug report link again?
01:59 est #2959
01:59 ShadowBot https://github.com/minetest/minetest/issues/2959 -- Ability to launch singleplayer game from command line removed by 'refactor main.cpp'
01:59 est top on the lis
01:59 est t
02:00 Zeno` Oh, I wasn't logged in... no wonder I couldn't see it :)
02:18 est31 joined #minetest-dev
02:33 * Zeno` confused now
02:36 est31 ah I see the problem
02:36 est31 or at least I think
02:36 Zeno` yeah I've seen the problem at least 3 times now
02:37 est31 heh
02:37 Zeno` actually I dunno if it's my commit
02:37 Zeno` I've just compiled cb2d46766582f3b3696d907efd6a93f80e6a53ef and it seems the same ?????
02:38 est31 also bin/minetest --worldname name --address '' --go works
02:40 Zeno` yeah I think that's how it might be meant to work
02:40 Zeno` those command line options confuse me to now end
02:40 Zeno` no*
02:40 Zeno` I can't see a regression anyway
02:41 est31 then I think it can be labeled unconfirmed
02:41 Zeno` I can't say I understand what the command line is supposed to be either, though :D
02:41 Zeno` probably
02:41 est31 https://github.com/rubenwardy/NodeBoxEd​itor/blob/master/src/minetest.cpp#L150
02:42 est31 thats his usecase
02:42 Zeno` I vaguely recall this being discussed about 6 months ago
02:42 Zeno` what's the markup for strikeout?
02:42 Zeno` I wanna strikeout my "confirmed" comment hehe
02:43 est31 so its not even a regression if --address '' isnt specified?
02:43 Zeno` I'm not sure
02:44 est31 does it work back then without --address ??
02:44 Zeno` if I checkout cbdd467 then minetest --go  doesn't launch a singleplayer game
02:44 Zeno` it tries to connect to VE_Vanilla which I assume is the last server I connected to
02:45 Zeno` cb2d467*
02:45 Zeno` But, as mentioned, I am confused as to how it's *supposed* to work
02:46 est31 well then its really the bug creator's fault. I guess its ok to label it unconfirmed and close it with the note to the correct way to use it
02:46 est31 and if rubenwardy gives a command line or a point where it worked
02:46 Zeno` maybe yeah...
02:46 est31 we can reopen
02:46 Zeno` yep, let's do that
02:47 Zeno` hmm wait
02:47 Zeno` leave me to investigate more
02:47 Zeno` actually close it
02:48 Zeno` minetest --address '' --go
02:48 Zeno` 12:47:09: ERROR[main]: No world selected and no address provided. Nothing to do.
02:48 Zeno` that's what cb2d467 says
02:49 Zeno` what have you got built atm?
02:49 Zeno` (save me rebuilding please :))
02:49 est31 I have built something close to master (my areastore pr changes applied)
02:49 Zeno` Either way I think the bug is not a bug but the behaviour *may* have changed slightly
02:49 est31 it says ERROR[main]: Angegebener Weltpfad existiert nicht: /home/user/.minetest/worlds/world
02:50 Zeno` yeah, so same thing? (essentially)?
02:50 Zeno` just a different message?
02:50 est31 it seems to look for the world in there.
02:50 Zeno` yeah
02:50 est31 but yes its a different thing
02:51 Zeno` Can't really say that's a bug or a regression; it might have been added as a feature since
02:53 est31 yea
02:54 Zeno` Are you release maintainer now?
02:54 est31 yes
02:54 Zeno` Maybe for the release after this one a goal should be to clarify the command line options
02:55 Zeno` I've always found those "implies" and the combinations of implications confusing
02:55 est31 isnt release maintainer a temporary job you get assigned per release?
02:55 Zeno` will piss a lot of people off though (those with scripts hehe)
02:55 Zeno` I dunno... maybe it is
02:56 est31 at least I think that
02:56 Zeno` you're probably correct
02:56 Zeno` release before last was handled by hmmm I think
02:56 Zeno` and last by ??
02:57 est31 0.4.12?
02:57 est31 I think hmmmm too
02:59 Zeno` let's assign #2959 to kahrl
02:59 ShadowBot https://github.com/minetest/minetest/issues/2959 -- Ability to launch singleplayer game from command line removed by 'refactor main.cpp'
02:59 Zeno` lol
03:02 Zeno` #2958
03:02 ShadowBot https://github.com/minetest/minetest/issues/2958 -- inventory.cpp:629: InventoryList::getItem: Assertion 'i < m_size' failed.
03:02 Zeno` ^-- backtrace in 2913
03:03 * Zeno` assigns it to est31
03:03 Zeno` I am the assignation manager now
03:03 Zeno` haha
03:14 bobomb when a texture is not found a default solid color is used, is there any way to control that color?
03:16 Zeno` I don't think there is
03:17 Zeno` I could be wrong but I think it's random and part of irrlicht
03:23 leat joined #minetest-dev
03:31 Zeno` valgrind: ==25494== ERROR SUMMARY: 433467 errors from 55 contexts (suppressed: 0 from 0)
03:31 Zeno` that's encouraging
03:33 est31 hah
03:33 Zeno` at last release it was 2
03:33 Zeno` (2 false positives)
03:38 paramat joined #minetest-dev
03:38 est31 hmmmm, updated
03:40 paramat will push game#593 and game#592 soon
03:40 ShadowBot https://github.com/minetes​t/minetest_game/issues/593 -- Default: New appletree schematic with tuned apple chance by paramat
03:40 ShadowBot https://github.com/minetes​t/minetest_game/issues/592 -- Default/mapgen: Disable nyancats in singlenode mapgen by paramat
03:48 Zeno` what is onMapEditEvent?
03:49 Zeno` addNodeWithEvent?
03:49 Zeno` hmm
04:01 paramat now pushing to game
04:09 paramat complete!
04:11 est31 ok what about #2827
04:11 ShadowBot https://github.com/minetest/minetest/issues/2827 -- Allow overriding the hand by TeTpaAka
04:12 est31 and #2852
04:12 ShadowBot https://github.com/minetest/minetest/issues/2852 -- Make players respect makes_footstep_sound in the object properties by TeTpaAka
04:22 paramat i certainly want the ability to change the hand texture in-game and per-player
04:24 est31 ok opinions on https://github.com/est31/mi​netest/tree/ncurses-console
04:26 VanessaE est31: aalib for rendering? ;)   seriously though, what's it look like in practice?
04:32 Zeno` is there any way to suspend mapgen?
04:32 Zeno` (for debugging purposes)
04:32 paramat not sure how
04:33 Zeno` ok, so I guess atm the answer is no
04:33 Zeno` might be something useful (but wayyyyyyyyyyyyyyyyyyyyy down on the list of priorities)
04:34 jin_xi joined #minetest-dev
04:34 est31 VanessaE, http://pasteboard.co/2glpdhqJ.png
04:34 est31 it shows off some of its features
04:35 est31 tab complete, ability to do commands, fact that you "see whats happening"
04:35 Zeno` est31, you may know. Does the protocol give up trying to send the map to the client after a certain time?
04:35 est31 dunno sorry
04:35 est31 why
04:35 VanessaE est31: looks good, but you should echo the given command to the console
04:36 Zeno` #2852 looks ok to me (the code)
04:36 ShadowBot https://github.com/minetest/minetest/issues/2852 -- Make players respect makes_footstep_sound in the object properties by TeTpaAka
04:36 est31 the "issued command" thing?
04:36 VanessaE e.g. where you kicked yourself, "/kick est31" should have been echoed to the console ahead of the success message
04:36 Zeno` also #2898 seems fine to me as well
04:36 ShadowBot https://github.com/minetest/minetest/issues/2898 -- Add AreaStore data structure by est31
04:36 Zeno` as long as hmmmm's comments are addressed
04:36 hmmmm ??
04:36 Zeno` (not necessarily in the PR)
04:37 hmmmm I think he got everything
04:37 VanessaE est31: yeah.
04:37 hmmmm like at this point it's just a bunch of minor style preferences
04:37 hmmmm which doesn't matter
04:37 Zeno` yeah, so I'm ok with it
04:38 * Zeno` is currently reviewing btw (may as well do something while waiting for valgrind)
04:38 Zeno` I dunno about 2852 so I'll skip that one
04:38 hmmmm #2898 approved
04:38 ShadowBot https://github.com/minetest/minetest/issues/2898 -- Add AreaStore data structure by est31
04:38 hmmmm let's get it added
04:39 Zeno` yep
04:39 est31 yaay
04:39 Zeno` maybe it'll make my valgrind tests faster so the sooner the better :D
04:39 hmmmm hhhhhh
04:39 hmmmm before you valgrind test we should add the minimap leak fix
04:40 Zeno` takes a long time to trace origins of 48K errors
04:40 hmmmm it's probably the minimap
04:40 Zeno` a lot seem related to that and tiles
04:40 hmmmm #2946
04:40 ShadowBot https://github.com/minetest/minetest/issues/2946 -- Fixed minimap memory leak by t0suj4
04:40 hmmmm tiles?!
04:40 Zeno` but I'm going to fully check just the server for now.
04:40 hmmmm oh god I thought I caught that leak before RBA merged it
04:40 hmmmm fucking RBA
04:41 hmmmm "it works!  just rubber stamp it!"
04:41 hmmmm "here look at this screenshot!"
04:41 hmmmm or maybe there's another different one i wasn't aware of
04:41 hmmmm est31:  do you concur?
04:41 hmmmm with 2946
04:42 est31 well there is this move / give thingy. I think I dont like how c++11 calls it move, like give much better.
04:42 est31 but guess its ok to merge.
04:43 hmmmm oh who cares
04:43 hmmmm i don't give you crap for your style preferences
04:44 hmmmm i say push the button on 2946
04:46 est31 well ok then.
04:46 est31 push it
04:46 Zeno` yeah, push it!
04:46 Zeno` the big red button with "do not push" written on it
04:47 Zeno` I think maybe my valgrind session has crashed
04:47 Zeno` I hate my computer since the components got wrecked :(
04:48 Zeno` it's using 100% CPU but...
04:48 Zeno` hmm
04:49 Zeno` I think it's stuck trying to do mapgen... I wish I could see if it's a deadlock or something
04:55 est31 its ok if c++11 calls it move, but this has nothing to do with lvalues or rvalues
04:55 est31 this is about pointers
04:55 Zeno` what's libnspr4?
04:56 paramat left #minetest-dev
04:56 est31 https://packages.debian.org/en/squeeze/libnspr4-0d
04:56 Zeno` netscape
04:56 Zeno` wtf is using that? lol
04:57 Zeno` oh... curl
04:59 sofar a bunch of things can use it.... can compile curl without it too
05:00 Zeno` hmmmm, are the onMapEditEvent things new?
05:00 hmmmm hum
05:00 hmmmm no
05:00 hmmmm i don't believe so
05:01 Zeno` http://codepad.org/r0GqwEAq
05:01 hmmmm est31:  if it's any consolation to you, that code is being changed quite a big by RBA if he's truthfully fixing up the minimap code
05:02 hmmmm do you understand now why I want high quality code merged, and not half-assed stuff with a half-assed promise that it will be worked on later?
05:02 hmmmm because that promise usually is not true
05:02 est31 heh, no big deal. you can push the pr if you want.
05:02 Zeno` what are map edit events?
05:02 hmmmm shrug
05:03 est31 hmmmm, I see
05:03 Zeno` I didn't edit the map, I just dug a piece of tree trunk heh
05:03 hmmmm i'm working on something
05:03 hmmmm can somebody else merge it
05:03 est31 hmmmm, the problem is however that some times criticism only comes if its pushed to master
05:04 est31 I dont think TBC_x has leak-checked all PRs in the queue
05:04 hmmmm TBC_x hasn't been pushing new features though
05:05 est31 well I say that people like TBC_x only check master, and not the whole list of open prs
05:05 hmmmm oh my god i need to stop talking on IRC otherwise nothing will ever get done
05:05 hmmmm i've been working on this same exact thing for like 4 days now
05:05 est31 have fun
05:06 hmmmm it's my fault
05:06 hmmmm i must have ADD or something
05:06 Zeno` but you're our mental
05:06 Zeno` mentor*
05:08 Zeno` sofar, I guess Fedora compiles curl to use it
05:08 Zeno` *shrug*
05:11 VanessaE hmmmm: nah, you don't have ADD, you just need to ta---SQUIRREL!
05:11 VanessaE ;)
05:12 hmmmm hmm
05:13 hmmmm i'm wondering how much minetest would crash if I add assertions about floating point range in serialization
05:18 est31 argh
05:18 est31 dammit somebody has removed IS_UTF8_MULTB_START
05:18 est31 and I think it was me
05:19 est31 because I've followed somebody asking me
05:19 est31 now I have to re-add it
05:19 est31 because I need it
05:32 Zeno` est31, srp.cpp line 815  (malloc not free'd)
05:34 est31 Zeno`, client or server
05:35 Zeno` seems to be client
05:35 est31 yeah
05:35 Zeno` I have to re-do that test because I didn't do it in my "controlled" way
05:35 Zeno` but I'm pretty sure that one is not a false positive
05:37 Zeno` (I'll probably forget to go back to that because I'm sidetracked on something else atm so I thought I'd better let you know ;))
05:38 est31 ahhh
05:38 est31 I see now
05:38 est31 two liner
05:38 Zeno` yep, it would be :)
05:38 est31 4 liner
05:39 est31 its not freed at two places
05:39 est31 handleCommand_AcceptSudoMode and handleCommand_AuthAccept
05:39 est31 at both it should
05:40 est31 and at both its a 2 liner to fi
05:40 est31 x
05:40 Zeno` don't forget to make a PR so it can be reviewed for a few weeks
05:40 Zeno` </sarcasm> hehe
05:40 est31 hehe
05:41 est31 can you make a pr for all the fixes, or do you want me to make a separate pr for this?
05:43 Zeno` ?
05:43 Zeno` can't you just push as trivial?
05:43 est31 srp.cpp line 815
05:47 est31 pushing in 15 minutes https://github.com/est31/minetest/commit/​8a3edfee4e8af61af946347407ca6a7d26815a05
06:01 Zeno` +1
06:03 hmmmm maybe fix the commit message?
06:04 est31 note: it also causes a double free which im investigating
06:04 est31 to which?
06:04 hmmmm 8a3edfee
06:04 est31 rrp
06:04 Zeno` -1
06:04 est31 i seeeee
06:05 Zeno` I'll paste the full valgrind output if you like
06:06 est31 I have enough information to fix the leak
06:06 est31 right now im investigating :)
06:06 Zeno` ok
06:07 Zeno` as said I'm not really concentrating... looking at the mapblockmesh issue
06:13 Zeno` pushing #2946
06:13 ShadowBot https://github.com/minetest/minetest/issues/2946 -- Fixed minimap memory leak by t0suj4
06:14 Zeno` or not
06:15 est31 ok new version https://github.com/est31/minetest/commit/​c34054cea3fc45d6bd5e385315c01f5677cefce9
06:15 est31 found the free that wasn't good
06:16 est31 now with fixed commit msg https://github.com/est31/minetest/commit/​f336d3ffb657d2fcc72d06c28642c1b418e8dec4
06:16 est31 can I push?
06:16 Hunterz joined #minetest-dev
06:20 Zeno` yes, much better. Isn't that what I said to do 30 minutes ago? :P
06:20 nore joined #minetest-dev
06:20 est31 pushed
06:20 Zeno` but yeah, push
06:20 Zeno` cool
06:35 H-H-H joined #minetest-dev
07:40 nrzkt joined #minetest-dev
07:46 julienrat joined #minetest-dev
07:49 Krock joined #minetest-dev
08:00 Yepoleb_ joined #minetest-dev
08:01 blaise joined #minetest-dev
08:20 kilbith joined #minetest-dev
08:28 Darcidride joined #minetest-dev
08:29 julienrat left #minetest-dev
08:50 nrzkt est31: gg !
08:58 Krock Ehm. What's the difference between spatialindex-32.dll and spatialindex_c-32.dll ? Which one do I need for minetest?
09:09 Krock nvm
09:12 TBC_x joined #minetest-dev
09:12 est31 nrzkt, gg??
09:12 nrzkt est31: areastore
09:13 est31 :D
09:13 est31 I've found yet another usecase: custom gravity
09:16 Calinou joined #minetest-dev
09:31 Krock est31, could you extend this https://github.com/minetest/mine​test/blob/master/README.txt#L202 with "spatialindex-32.lib" for win32 please?
09:41 est31 Krock, nice idea, can you make a pr?
09:41 Krock ...k
09:43 est31 man why is our render speed so slow?
09:43 est31 http://pasteboard.co/2gFFjTvU.png
09:43 est31 look at that game
09:44 Calinou est31, Craft is an extremely basic game
09:44 est31 it has double of the fps than minetest, and the area it renders is at least 4x as large
09:44 Calinou also, only 21 fps
09:44 Calinou I'm not sure if world height is large in Craft too
09:44 Krock blame irrlicht
09:44 Calinou Irrlicht isn't slow in itself, it's the way we use it
09:45 Calinou we should stop blaming Irrlicht for everything :/
09:45 est31 it is basic agreed
09:45 est31 but its renderer is still faster
09:45 est31 it even has fancy leaves
09:46 Calinou programming your own renderer by hand will always be faster
09:46 Calinou but who in this community has the skills to do it?
09:47 kilbith i tested Craft, it uses advanced GL shaders apparently
09:48 kilbith but i'm more interested by that : https://github.com/fogleman/​Craft/blob/master/src/sign.c
09:48 kilbith you can write on every node
09:50 nrzkt there are two optimisations to do on clients, as i said in PM to est31
09:51 nrzkt 1. for particles: have a particlemanage server side which send particle manager to clients depending on their distance. This manager tells clients to add particle mgr where in range and remove it when not in range if he have it.
09:51 nrzkt Atm client receive all particle mgr loaded on the map and do something with it
09:51 nrzkt also this particlemgr should remove its entries based on particlespawner timeout and notify clients
09:51 nrzkt 2. Same thing but for SAO
09:52 nrzkt we don't need to have all map SAO shown to client
09:52 est31 this is possible already now
09:52 nrzkt i think 100 blocs is sufficient
09:52 nrzkt not really
09:52 Krock est31, #2961
09:52 ShadowBot https://github.com/minetest/minetest/issues/2961 -- Extend CMake variable description by SmallJoker
09:52 est31 and particlespawners is no that much of an improvement
09:52 est31 #
09:53 nrzkt it's a little improvement for maps with many many particlespawner and that permit to fix bug to particlespawner send to clients only one time
09:53 nrzkt if we connect 10 sec after particle spawner spawn we don't see it if it lasts 10 minutes.
09:53 est31 Krock, can you remove the sorting? its nice, but these things are "listed" at other places too and it would be cool to have the same order
09:53 nrzkt for sao, effectively we don't have this problem
09:54 nrzkt but i'm not sure of the distance
09:54 nrzkt also clients doesn't need to know about too far blocks, they should be removed , regarding client distance to mapblocks
09:54 est31 yes and no
09:54 est31 yes because its otherwise a leak
09:54 nrzkt if i can't see the mapblock i don't need to render it
09:55 est31 no because e.g. players can teleport between locations
09:55 nrzkt and to keep it in memory
09:55 est31 its not rendered if its far away
09:55 nrzkt if they teleport it's like teleport to unknown area
09:55 est31 that would kill rendering speed
09:55 est31 many servers have a travelnet mod
09:56 est31 even vanessae's vanilla server had it as it still was called vanilla
09:56 est31 erm sorry wont kill rendering speed but it will make teleporting back and forth tedious
09:56 est31 we need a better mechanism for this
09:56 nrzkt yes
09:57 nrzkt a timeout + distance could be good
09:57 nrzkt if distance > MAX && timeout done => unload
09:57 est31 e.g. record the last 10 128x128x128 blocks a client was in.
09:57 est31 for those we store stuff
09:57 est31 for the old ones, we delete
09:57 est31 the limit can be configured
09:59 nrzkt yes
09:59 nrzkt it's a good thing for modest devices like android
09:59 Krock est31, updated.
10:01 est31 +1
10:01 est31 pushing #2961 in 30 minutes
10:01 ShadowBot https://github.com/minetest/minetest/issues/2961 -- Extend CMake variable description by SmallJoker
10:02 TBC_x I think mapblocks could use slab allocation
10:02 est31 ??
10:03 TBC_x slab allocation allocates several pages of memory up front
10:03 TBC_x and it should be cache friendly
10:04 TBC_x I doubt that calling mmap 50 times a second is good for performance
10:05 leat joined #minetest-dev
10:06 est31 mmap is something for communicating which is not cpu or ram like hdd or graphics card no?
10:06 TBC_x mmap is syscall that is allocating memory
10:07 TBC_x need a mapblock pool or something
10:07 TBC_x mmap is called from stdlib
10:08 TBC_x every allocation
10:08 TBC_x and mmap implies a context switch
10:08 est31 well yeah, malloc and free are expensive
10:09 est31 " mmap, munmap - map or unmap files or devices into memory"
10:09 est31 I dont see how this is used by every malloc
10:09 est31 this is more about "read file"
10:09 est31 perhaps its called 50 tiems every second on plan9
10:09 est31 but thats only because everything is a file over there, including the registers of the cpu
10:10 est31 (ok last was sarcasm)
10:10 est31 do you really mean this call http://man7.org/linux/man-pages/man2/mmap.2.html
10:12 TBC_x i should not mention mmap
10:12 TBC_x but malloc
10:15 TBC_x oh it is brk syscall
10:16 VanessaE est31: um, problem
10:16 TBC_x but whatever
10:16 TBC_x need less calls to malloc
10:16 VanessaE est31: areastore seems to have... um... broken:  http://pastebin.ubuntu.com/11947686/
10:17 VanessaE all 8 of my servers failed to restart after the backup cycle
10:17 est31 VanessaE, did you update the areas mod too?
10:17 VanessaE no?
10:17 est31 perhaps i should have mentioned that...
10:17 VanessaE ok.
10:17 est31 it was needed, because api broke
10:18 * est31 feels sorry for the downtime.
10:18 VanessaE no worries, got a link to the areas patch?
10:18 bobomb joined #minetest-dev
10:18 est31 https://github.com/ShadowNinja/areas/pull/13
10:19 VanessaE ok
10:20 VanessaE ok, deployed it.
10:21 kilbith_ joined #minetest-dev
10:22 Taoki[laptop] joined #minetest-dev
10:24 msantana joined #minetest-dev
10:35 * Krock punches est31
10:45 Routh joined #minetest-dev
11:16 Amaz joined #minetest-dev
11:17 enesbil joined #minetest-dev
11:30 FR^2 joined #minetest-dev
11:39 kahrl est31, TBC_x, yeah, mmap is used internally by malloc
11:40 kahrl for example the malloc in glibc has a simple dynamic algorithm to choose between (s)brk for small allocations and mmap for larger allocations
11:43 est31 kahrl, but mmap is a "get me this file into memory" call, no?
11:44 kahrl not if you pass MAP_ANONYMOUS
12:30 Wayward_One joined #minetest-dev
12:32 deltib joined #minetest-dev
12:35 alket joined #minetest-dev
12:42 proller joined #minetest-dev
12:49 proller joined #minetest-dev
12:53 est31 joined #minetest-dev
12:54 Megaf joined #minetest-dev
13:13 FR^2 left #minetest-dev
13:28 hmmmm joined #minetest-dev
14:00 everamzah joined #minetest-dev
14:25 jin_xi joined #minetest-dev
14:36 SopaXT joined #minetest-dev
14:37 H-H-H joined #minetest-dev
14:37 est31 nrzkt, what do you mean with "getting stun"?
14:37 est31 perhaps its a wow thing, i dont know it
14:38 Zeno` joined #minetest-dev
14:46 nrzkt to be stun = ... stun ?
14:46 nrzkt like if i hit you with a hammer on your head, you are stun :p
14:47 unkempt` nrzkt, go west
14:47 nrzkt i prefer north
14:47 unkempt` no, no
14:48 unkempt` west is better
14:48 nrzkt but maybe it's an english expression and i didn't understand it ?
14:48 unkempt` not really
14:48 unkempt` it's https://www.youtube.com/watch?v=D3X1aExRYqs
14:48 unkempt` i.e. drunk talk
14:48 unkempt` I just cut my own hair and look like an idiot
14:48 unkempt` and LOVE it
14:49 unkempt` it's the new me
14:49 est31 man android is BROKEN
14:50 unkempt` est31 it is
14:50 est31 and RBA's attitude to it OMG ITS OGL ES WHO THE FUCK WANTS TO USE IT
14:50 unkempt` shaders are broken
14:50 est31 its decades old
14:50 est31 not just shaders
14:50 est31 somebody made all nodes white
14:50 est31 shaders disabled
14:51 nrzkt Zeno you are crazy man
14:51 unkempt` hmm, it was ok when I last compiled it
14:51 unkempt` which, I admit, was about 3 months ago
14:51 ElectronLibre joined #minetest-dev
14:51 nrzkt est31: you didn't like a white world ? :p
14:51 unkempt` west end... girls...
14:52 est31 and textures are broken too
14:52 unkempt` textures are broken in master
14:52 unkempt` (using unititialised values)
14:52 unkempt` there is a lot to fix
14:53 unkempt` I should not have left you ragamuffins alone for so long
14:53 unkempt` :P :D :)
14:53 nrzkt Zeno, are you drunk ?
14:53 unkempt` nrzkt, affirmative
14:53 nrzkt or did you punched by a kangaroo ? :p
14:53 unkempt` nah, just drunk
14:54 sloantothebone joined #minetest-dev
14:54 unkempt` but there is something wrong with tiles and textures (honestly)
14:54 unkempt` I checked that before I got drunk
14:55 unkempt` it's partially the reason I GOT drunk. The other part was because I wanted to
14:56 est31 http://pasteboard.co/2h0bJusd.png
14:57 est31 this becomes a high priority bug, blocker, and blocker for 0.4.13.
14:57 nrzkt est31: approved
14:57 hmmmm i don't understand why 0.4.13 is being rushed anyway
14:57 hmmmm this is ridiculous
14:57 hmmmm there are regressions and memory leaks galore
14:57 nrzkt hmmmm: approved
14:58 hmmmm nevermind approval, nobody actually approved the 0.4.13 release date to begin with
14:58 est31 nobody spoke against
14:58 hmmmm let's get minetest in a stable state first
14:58 hmmmm then we'll see about releasing
14:58 est31 well yes
14:58 est31 thats what feature freeze is for
14:58 est31 no more breakage, now we fix bugs
14:58 nrzkt 2 weeks of feature freeze is too light i think.
14:59 nrzkt do a feature freeze on 1st august and release date on 1st september is better
14:59 unkempt` 0.4.13 is a long way off unless all these things valgrind reports can be fixed quickly
14:59 hmmmm how about not making a release date
14:59 est31 and stalling the release indeterminately?
14:59 hmmmm there can be a feature freeze date but release should be tenative based on whenever things are in a stable state
14:59 unkempt` tbh I am pretty fucking annoyed about how much shit valgrind is reporting
14:59 hmmmm it's better to stall a release than release in this state
14:59 est31 say "ah yeah there are these issues, but we wont release"
15:00 unkempt` 0.4.11 and .12 had... nothing
15:00 est31 unkempt`, valgrind did report stuff there too
15:00 est31 e.g. global variables
15:00 est31 like g_settings
15:00 est31 or g_logger
15:01 est31 or all the unittest stuff
15:01 est31 ofc they are no real leaks
15:01 unkempt` est31, maybe I didn't check 0.4.12 but in the release I *was* involved in there was nothing
15:01 est31 perhaps it is better to have a full list of errors unkempt`.
15:01 unkempt` maybe that was 0.4.11
15:02 unkempt` 0.4.12 was a rushed (IMO) so, yeah, maybe I didn't check it
15:03 unkempt` but seriously est31, this many errors is not acceptable
15:03 * H-H-H has been saying bout problems with android build on master for ages lol
15:03 unkempt` well, maybe it is, but I personally find it embarrassing
15:03 unkempt` 43K errors from the client is just stupid
15:04 hmmmm i think the issues aren't being pressed because there are no developers currently active who use android
15:04 unkempt` hmmmm, the 43k errors are on Linux (PC)
15:04 est31 well android isnt in a releaseable state at all
15:04 unkempt` most are openAL and crap though
15:04 hmmmm i mean the texture issues
15:04 hmmmm unkempt`, 99% of those errors are caused by the minimap
15:04 hmmmm here, i'll fix them right now
15:05 unkempt` hmmmm, I know
15:05 unkempt` good... push it :D
15:05 * H-H-H waits for the git bot to say reverted minimap commit
15:06 unkempt` I thought we needed to have a VERY good reason to revert a commit
15:06 unkempt` i.e. it very, very, very rarely happens once something is merged
15:06 est31 ^
15:06 * H-H-H was making light of the situation :|
15:06 unkempt` H-H-H, I know... was just sayin' :)
15:08 H-H-H i seriously admire you guys for doing what you do :)
15:10 unkempt` since I brought up OpenAL... does its current license affect us at all?
15:10 est31 #2962
15:10 ShadowBot https://github.com/minetest/minetest/issues/2962 -- Android graphics broken
15:10 est31 ah forgot a label
15:10 est31 there added
15:11 unkempt` est31, I'll bisect Android problems tomorrow if you like
15:11 unkempt` oh
15:11 unkempt` those cubes, est31?
15:11 est31 yup
15:12 est31 I'm not running "everything is white" subgame
15:12 unkempt` yep, that's been around for a while. I'll gather together all the bugs I think and put them in one place
15:12 unkempt` est31, that's with no shaders at all?
15:12 est31 yes
15:12 est31 with shaders its all red
15:13 H-H-H yeah shaders are really messy on android
15:13 unkempt` I can bisect that pretty easily because I have a working android build here (in a different directory)
15:13 est31 if you want to do, do it
15:13 est31 I could do it too
15:14 unkempt` yep, I will. It's the only reason I bought an android :D
15:14 est31 we should somehow separate forces here lol
15:14 est31 wow
15:14 unkempt` but if you beat me to it I don't mind it
15:14 est31 can you put the list of valgrind errors online?
15:14 est31 then I can work on tracing them
15:14 est31 you do android
15:15 est31 perhaps with the ones that are fixed by the last commit filtered out, if they are indeed 99%
15:15 unkempt` est31, not at the moment because I have to re-do them (I prefer to run valgrind in a "controlled" environment and the current ones I have were not controlled). By control I mean a known environment
15:15 est31 ok
15:16 unkempt` gotta be scientific otherwise it's kinda hard
15:16 est31 then I'm looking at another bug that bugged me since some time
15:16 unkempt` but yes, I will post them once I re-run
15:16 est31 #2097
15:16 ShadowBot https://github.com/minetest/minetest/issues/2097 -- Android build doesn't use proper paths
15:17 est31 how we currently do it is horrible, to put it mildly
15:17 unkempt` yes
15:17 unkempt` I've looked at that (only briefly) as well
15:17 unkempt` also the controls suck
15:17 est31 well yeah
15:17 est31 but thats gameplay
15:17 unkempt` I have patches to make to the controls better
15:17 est31 who cares about that lol
15:17 unkempt` me
15:17 est31 jk
15:17 Calinou <+unkempt`> since I brought up OpenAL... does its current license affect us at all?
15:18 Calinou aren't we using ALSoft?
15:18 Calinou Package: libopenal-dev
15:18 Calinou Source: openal-soft
15:18 Calinou this is what Debian reports
15:18 unkempt` also, android shutdown is different from PC/MAC shutdown sequence (for no reason at all)
15:18 unkempt` Calinou, if we're using ALSoft there is no problem
15:18 est31 yes there is a reason
15:18 est31 unkempt`, we had this already, android is a shit os, not even allowing applications to exit
15:19 unkempt` is openal dynamically linked? If that's the case then it matters not
15:19 est31 but trying to push its views how to design applications down users throats.
15:19 unkempt` est31, yes but the shutdown sequence should be (and can be) the same
15:19 est31 ok
15:19 unkempt` I had a PR that I closed but only because it was broken
15:20 Calinou it's dynamically linked normally
15:20 Calinou except on Windows perhaps...
15:20 unkempt` apart from the broken bits I learned that there is stuff that happens that is not necessary (on Android but not on PC)
15:21 unkempt` Calinou, I think on Windows it must be dynamic as well because all builds I've seen include an openal DLL
15:22 unkempt` est31 in case you hadn't noticed this is the first time in 4 months that I've been on IRC for >8 hrs
15:22 unkempt` I'm back
15:22 unkempt` and I'm annoyed (which equals motivated)
15:22 unkempt` I kind of lost interest in MT for most of those 4 months truth be told
15:23 unkempt` but now there is solid stuff to be done :D So I'm happy again
15:23 est31 heh
15:23 est31 rba did a quite good job at it, perhaps me too
15:24 unkempt` nah
15:24 unkempt` I got depressed when my computer blew up
15:24 unkempt` and kind of went off the rails
15:24 unkempt` shit happens
15:25 unkempt` BUT I must say these valgrind things must be fixed before release
15:25 unkempt` I'll get them all in one place so we can all look at them
15:30 bobomb i know this is terrible but: https://github.com/bobombolo/woolimage (attempt to draw map pixel by pixel in formspec)
15:31 Calinou unkempt`, isn't that static?
15:31 Calinou if you include the DLL
15:32 unkempt` Nah, DLL is dynamic
15:32 unkempt` Dynamic Link Library ;)
15:32 unkempt` (or dynamically linked lib)
15:34 est31 bobomb, you can hack it if you want
15:34 julienrat joined #minetest-dev
15:34 est31 just make an 1 pixel image for each color you want to provide
15:34 est31 then show the images
15:35 est31 ofc it wont be "real" pixels
15:35 ElectronLibre joined #minetest-dev
15:35 est31 so they will be scaled
15:35 est31 it will look messy
15:35 est31 and loading will be ultra slow
15:35 est31 really ultra ultra slow
15:35 est31 but will work!
15:35 est31 or you wait for somebody to add minimap for formspecs
15:36 est31 we already have a minimap
15:36 est31 we only need the "show in formspec" feature
15:40 julienrat left #minetest-dev
15:52 crazyR last_login in the auth file: is it used for anything internally?? what i mean is could i change the name if the key without risking breaking something in the engine?
15:52 Player_2 joined #minetest-dev
15:53 est31 well its used for /last-login username
15:53 est31 very handy
15:53 Robert_Zenz joined #minetest-dev
15:54 est31 I guess you can indeed change it
15:54 est31 nothing breaks as of now
15:55 est31 question is, why you want to change it
15:55 est31 and perhaps we might rely on it in future
15:55 est31 if you can live with the risk ,great :)
15:58 crazyR i can updater the  /last-login cmd, the reason is ive added  reg_date to my custom auth and its driving me mad, as id like it named login_date so it maches haha
16:00 jin_xi joined #minetest-dev
16:03 ElectronLibre joined #minetest-dev
16:04 johnnyjoy joined #minetest-dev
16:13 Hunterz joined #minetest-dev
16:34 TBC_x use my patch before running valgrind
16:39 TBC_x otherwise you're really going to have 40k errors
16:39 nrzkt joined #minetest-dev
16:42 rubenwardy joined #minetest-dev
17:10 kilbith joined #minetest-dev
17:11 SopaXT joined #minetest-dev
17:14 MinetestForFun joined #minetest-dev
17:15 rubenwardy joined #minetest-dev
17:26 bobomb est: i don't want minimap, i want to show different thematic maps, like slope, aspect, elevation, water, etc. yes it takes about 2 full seconds to load that 100x100 pixels
17:33 SopaXT joined #minetest-dev
17:36 est31 only 2 seconds?
17:36 est31 minetest is blazingly fast
17:37 bobomb i actually think that's usable, but that's obviously not over a network...
17:38 TBC_x wow
17:38 TBC_x if you have a supercomputer... maybe...
18:05 Zeno` lmfao
18:05 Zeno` minetest is painfully slow
18:05 Zeno` and that's *after* I optimised it
18:05 bobomb 80x80 is only half a sec
18:06 Zeno` 1/2s is an eternity
18:06 Zeno` we need at least 60 fps
18:07 Zeno` so what is the longest something can take if aiming for 60fps?
18:07 Calinou 16.667 ms
18:07 Zeno` ^--- quite a bit less than 1/2 a second :(
18:08 Zeno` and that's EVERYTHING
18:08 Zeno` not just one thing
18:11 est31 joined #minetest-dev
18:14 * est31 commented out code to be scheduled for removal in the commit
18:14 * est31 realized he would need the functionality
18:14 est31 therefore est31 modified the methods called by that code
18:15 * est31 then compiled
18:15 est31 error
18:15 est31 turned out i forgot to remove the comment out
18:24 FR^2 joined #minetest-dev
18:25 Wayward_One Someone asked an interesting question on my build thread: https://forum.minetest.net/vi​ewtopic.php?p=185532#p185532, which poses another question: Is Minetest even supported on, let alone playable, on Android 2.2.2?
18:27 rubenwardy I very much doubt it]
18:27 ElectronLibre joined #minetest-dev
18:28 Wayward_One Heh, same here. I figured I'd make sure though
18:33 sloantothebone joined #minetest-dev
18:35 est31 Wayward_One, no, we start support with SDK 9
18:35 est31 aka gingerbread
18:35 est31 froyo isnt supported
18:35 est31 run aapt dump badging bin/Minetest-debug.apk
18:35 est31 youll see sdkVersion:'9'
18:36 est31 then look at this table http://developer.android.com/guide/topics​/manifest/uses-sdk-element.html#ApiLevels
18:37 est31 and if you look at this table, there is not much sense in supporting it http://developer.android.com/abou​t/dashboards/index.html#Platform
18:37 est31 that android is very very old
18:37 est31 you get few users
18:37 est31 perhaps its easy however, only flipping a switch
18:37 est31 dunno
18:38 sloantothebone When is the new version of minetest going to be released?
18:38 sloantothebone Oh nvm
18:39 rubenwardy sloantothebone, probably not on the 10th.
18:39 est31 ARGGGGGH
18:39 rubenwardy when it reaches a stable enough state
18:39 sloantothebone Ok
18:39 est31 we depended on our filesystem being fucking CASE INSENSITIVE on android
18:39 est31 CASE
18:39 est31 INSENSITIVE
18:40 rubenwardy Android is case insensitive?
18:40 est31 no
18:40 sloantothebone I'm using a daily build and the right click and drag isn't working right
18:40 est31 it depends on the filesystem
18:40 est31 android itself isnt case insensitive
18:41 est31 but somehow when you store data on the sd card its formatted fat32 or whatever, and that seems to be not case sensitive
18:41 est31 because I see here data being written to a path with "Minetest", and data being read from a path with "minetest"
18:42 est31 and it worked before, so i conclude it was due to filesystem
18:42 est31 and it broke when I switched filesystems to app-private storage
18:42 sloantothebone Ok I take a stack of 12 and I right click and drag and only 1 is dropped on the slot I clicked and then I right click and drag again and it works. Its like it wont work unless I've placed an item
18:43 est31 sloantothebone, confirm
18:43 rubenwardy That's the drawing feature, where you can make say a pick axe by swiping up with sticks then left with stone
18:43 est31 can you create a github issue?
18:43 Wayward_One est31: ok, thanks for clarifying.
18:43 rubenwardy unless I misunderstand
18:44 sloantothebone Uh sure
18:44 est31 and best, can you bisect
18:44 sloantothebone But I should update minetest first, because I dont update every day
18:44 rubenwardy you're right
18:44 rubenwardy confirmed too
18:50 sloantothebone Upgrading minetest
18:51 sloantothebone Where do I post issues for the daily build
18:51 est31 if its a bug with the engine then at github
18:53 rubenwardy another feature added recently, then broken recently XD
18:53 rubenwardy well, the other was changed
18:54 sloantothebone What feature?
18:54 sloantothebone Btw are there any plans for implementing the ability to evenly spread an entire stack by left clicking and dragging in the inventory?
18:54 rubenwardy #2959
18:54 ShadowBot https://github.com/minetest/minetest/issues/2959 -- Ability to launch singleplayer game from command line removed
18:55 alket joined #minetest-dev
18:55 sloantothebone Where do I submit feature requests? and hi alket
18:56 alket hi
19:04 alket sloantothebone: i guess its github issues
19:09 sloantothebone Ok
19:10 sloantothebone I'm making an issue about the right click and drag inventory bug
19:11 sloantothebone #2964
19:11 ShadowBot https://github.com/minetest/minetest/issues/2964 -- right click and drag inventory bug
19:17 est31 Wayward_One, might be an irrlicht regression
19:17 est31 i have newest irrlicht
19:20 sloantothebone Oh
19:20 est31 not that one
19:20 est31 talking bout #2963
19:21 ShadowBot https://github.com/minetest/minetest/issues/2963 -- Feature request: Implement dividing entire stack evenly into inventory slots when you left click and drag
19:21 alket feature request, make Cart not lag :p
19:21 Calinou not possible unless the server lags less
19:22 alket its the only "bug" that makes me angry when playing
19:22 alket not that i know any other bugs anyway
19:23 alket but its a mining game and cart is like one of basic features
19:24 sloantothebone #2963
19:24 ShadowBot https://github.com/minetest/minetest/issues/2963 -- Feature request: Implement dividing entire stack evenly into inventory slots when you left click and drag
19:25 sloantothebone Not possible?
19:25 est31 it is possible
19:25 est31 but youlll have to find sb to code it
19:25 est31 or code it yourself
19:25 alket i hate this reply :p i get it all the time
19:26 alket in all games
19:26 sloantothebone Well where is the code for the right click and drag?
19:26 est31 guiformspecmenu.cpp i think
19:27 est31 alket, my comment was more targeted at sloantothebone. making carts lag less requires touching SAOs which is very likely to fail
19:27 est31 because its horrible code
19:27 sloantothebone I have taken code, figured out how it works and modified it
19:29 sloantothebone Ok I'm in guiformspecmenu.cpp
19:30 sloantothebone Ok I've found the code for inventory stuff, at least thats what the comment says
19:30 Wayward_One est31: could be, I still have the irrlicht version that comes with Mint 17
19:30 sloantothebone field while holding right mouse button
19:30 sloantothebone */
19:31 sloantothebone Mouse event other than movement, or crossing the border of inventory
19:31 sloantothebone field while holding right mouse button
19:32 est31 Wayward_One, if you compile for android you build irrlicht yourself
19:32 sloantothebone Um could somebody pm me and walk me through this a bit, I don't mess with c++ very often
19:34 Wayward_One est31: ...oh lol
19:37 MinetestForFun joined #minetest-dev
19:43 ElectronLibre joined #minetest-dev
19:58 sloantothebone How do I debug c++?
20:01 est31 you can log stuff using errorstream << "text" << std::endl;
20:02 est31 or use debuggers
20:02 est31 alket had a link i think to youtube videos
20:02 alket i didn't bookmark it :s
20:03 alket this https://www.youtube.com/watch?v=xTmAknUbpB0 , but it isn't straightforward, its like a talk
20:12 sloantothebone Ok any suggestions for debuggers?
20:13 sloantothebone Maybe I can mess with the minetest source code and learn a bit about c++
20:26 sloantothebone Ok fine I'll google it
20:28 est31 Wayward_One, does iconv build for you?
20:29 est31 e.g. for the android builds
20:41 sloantothebone est31 is anybody working on implementing luajit into minetest? or client side modding?
20:42 Krock there's already luajit support
20:42 Krock client side modding is a bad idea
20:42 sloantothebone Ok how do I use luajit and why is it a bad idea?
20:43 sloantothebone the only problem I can see is hacking but I leave that to the lua writers to decide what should be server side
20:44 Krock set the cmake variables LUA_INCLUDE_DIR and LUA_LIBRARY. Hacking would be way too easy, that's why
20:46 sloantothebone I'm not talking about completely client side modding I'm talking about the option to make parts of the code run on the client
20:50 sloantothebone And so I have to compile minetest to get luajit?
20:53 Krock no, get luajit and compile minetest
20:53 sloantothebone That's what I meant
20:54 Krock or just use a compiled version...
21:02 Player_2 joined #minetest-dev
21:19 ElectronLibre left #minetest-dev
21:29 wingless_ joined #minetest-dev
21:30 Player_2 joined #minetest-dev
22:08 paramat joined #minetest-dev
22:18 BrandonReese joined #minetest-dev
22:27 deltib joined #minetest-dev
22:28 kahrl joined #minetest-dev
22:34 Miner_48er joined #minetest-dev
23:00 kahrl joined #minetest-dev
23:31 paramat left #minetest-dev

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