Minetest logo

IRC log for #minetest-dev, 2015-05-29

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

All times shown according to UTC.

Time Nick Message
00:00 zat mapgen_sand* I mean
00:05 zat paramat: http://uppix.com/f-screenshot_263355567acfb0019089e.png
00:05 zat without an alias for mapgen_sand there is nothing below y +3
00:05 * paramat looks
00:06 paramat seems to be placing air instead
00:06 paramat oh you removed the alias
00:06 zat there are tons of 20:59:59: ERROR[main]: Resolver: Failed to resolve node name 'mapgen_sand'.
00:07 zat its lke no biomes allowed below y+3
00:07 paramat no idea, but make sure to use latest MTGame also
00:08 zat remember that mod earth is actually a game
00:09 zat I am running it without most of the stuff, its minimal
00:09 zat so I confirm that no biomes are placed below y+3
00:09 paramat ah i see
00:10 paramat so your beach and underwater biomes are failing
00:11 zat right
00:11 troller joined #minetest-dev
00:12 zat if I make my beach biome have ymin -31000 and ymax 10, it is like ymin 3 and ymax 10
00:12 zat ymin 4 I meant
00:17 paramat this is probably due to your code not the engine, good luck
00:17 paramat left #minetest-dev
00:42 nolsen joined #minetest-dev
00:43 nolsen joined #minetest-dev
01:02 zat not a very motivating attitude for a dev
01:12 Hijiri joined #minetest-dev
01:15 VanessaE is someone going to address https://github.com/minetest/minetest/issues/2612 soon?
02:09 cheapie joined #minetest-dev
02:27 Wayward_Tab joined #minetest-dev
03:43 paramat joined #minetest-dev
03:45 paramat zat talk to me in channel
03:46 paramat hmmmm 4D fractal mapgen http://i.imgur.com/eIIbvcQ.png
03:46 zat paramat: I could not configure a buddy pounce in channel :P
03:48 paramat i saw your comment earlier, i was not being rude i was helping you by stating my best guess at what the bug was
03:48 paramat i had to go do stuff, i#m busy
03:48 paramat (i'm)
03:49 paramat you're lucky i helped you at all, i work voluntarily and have no commitments to help people
03:51 paramat everyone else's biome systems work, so it's likely your code or incompatibility due to my changes
03:53 zat paramat: I should use more emoticons or people take me wrong
03:53 zat I was not calling you rude lol
03:53 zat whatever, it is that change that introduces shores
03:54 paramat you just bitched at me in private message :)
03:54 zat paramat: no, really
03:54 zat IMO the idea of noisy average altitude for biomes would make better shores
03:55 paramat yes
03:55 zat in real life beaches are not everywhere and are not the same height
03:55 paramat some vertical borders need to be level though, so a 'variation' parameter would be good
03:56 paramat if this is possibe..
03:57 zat strength maybe
03:57 zat or a factor
03:58 zat so a big variation would make better snowed mountains and little variation for beaches so they are not grossly different.
04:00 zat this makes me remember a good idea that I had and that I somehow worked around to implement it
04:01 zat increase the density of ores by a certain factor, depending on the depth
04:06 paramat a smooth increase?
04:07 zat right, and lineal
04:07 paramat problem is any noticeable increase over 500 nodes would become ridiculous at the world base
04:08 paramat we can already increase it in steps with current code, as is already done
04:08 zat depends on the numbers I guess
04:09 zat in my world, I registered lava as an ore, with a 3d noise that gets bigger and bigger with the depth
04:09 zat sure the minerals are a lot under -15000, but go try mine down there without finding huge lava pits.
04:15 Miner_48er joined #minetest-dev
04:15 paramat requests for features for one person's use often get nowhere, it's best to join github and/or the forum document your ideas and gain support, github is the most effective way to do this
04:17 zat I will soon I expect, meanwhile dropping ideas still help :P
04:45 paramat yes, ideas are still welcome here of course
04:52 paramat left #minetest-dev
05:46 hmmmm i wonder if it'd be useful at all to make a worldedit command to heal map irregularities using a heightmap blur
05:46 hmmmm for example if there's a hard cut in the map due to a mapgen problem
05:50 Hunterz joined #minetest-dev
06:01 VanessaE good idea, hmmmm
06:09 RealBadAngel joined #minetest-dev
06:11 leat joined #minetest-dev
06:20 cib0 joined #minetest-dev
06:21 leat joined #minetest-dev
06:31 leat joined #minetest-dev
06:49 Darcidride joined #minetest-dev
06:50 err404 joined #minetest-dev
06:52 leat joined #minetest-dev
07:04 leat joined #minetest-dev
07:19 SopaXT joined #minetest-dev
07:31 Darcidride_ joined #minetest-dev
08:00 Yepoleb_ joined #minetest-dev
08:13 SopaXT joined #minetest-dev
08:26 twoelk joined #minetest-dev
08:27 cib0 joined #minetest-dev
08:28 twoelk hmmmm: that could be a way to achieve what I described here: https://forum.minetest.net/viewtopic.php?p=176870#p176870
08:44 leat joined #minetest-dev
08:45 twoelk a tool for border blur would make it easier to combine map content of different origin. A user could blend terrain made by different mapgens or even insert prebuild areas with a more pleasing result. Such maps could become unique in a way that cannot be recreated from a seed as the player decission how to create which terrain cannot be predicted by  any code.
08:54 leat joined #minetest-dev
09:16 leat joined #minetest-dev
09:27 Darcidride joined #minetest-dev
09:31 Megaf_ joined #minetest-dev
09:34 Puma_rc joined #minetest-dev
09:46 Darcidride joined #minetest-dev
09:48 rubenwardy joined #minetest-dev
09:54 Calinou joined #minetest-dev
09:55 leat joined #minetest-dev
10:05 leat joined #minetest-dev
10:17 proller joined #minetest-dev
10:19 Amaz joined #minetest-dev
10:24 cib0 joined #minetest-dev
10:24 rubenwardy joined #minetest-dev
10:43 Darcidride joined #minetest-dev
11:06 leat joined #minetest-dev
11:17 leat joined #minetest-dev
11:22 Darcidride joined #minetest-dev
11:36 leat joined #minetest-dev
11:38 Wayward_Tab joined #minetest-dev
11:57 leat joined #minetest-dev
11:58 Megaf_ Hi all
11:58 Megaf_ can we use this instead of LuaJIT? https://code.google.com/p/llvm-lua/
11:59 Megaf_ AOT instead of JIT
12:00 Calinou “JIT compile”
12:00 celeron55 why would you
12:01 Megaf_ Smaller memory footprint, faster initialization time and lower CPU overhead.
12:01 Calinou people should not have to compile their mods
12:01 Megaf_ The engine could compile mods on startup
12:01 celeron55 Megaf_: they haven't even published any benchmarks; it's likely not any faster
12:02 celeron55 and llvm is huge; nobody wants to add tens of megabytes to their downloads and builds for that
12:02 Megaf_ celeron55: so just as big as the bitmap fonts...
12:03 Calinou you could make Minetest use Clang by default :3
12:03 Megaf_ I rather have faster performance than use bitmap fonts, what about you?
12:03 Calinou AssaultCube switched to Clang as default compiler
12:03 celeron55 and that's not properly documented either; i have no idea how you interface with that from code that normally calls actual lua or luajit
12:03 Megaf_ Calinou: I see no runtime gain in using cland, it might compile faster, but that's about it
12:04 Megaf_ celeron55: I agree with you on the lack of documentation.
12:05 celeron55 also it's not maintained either
12:05 celeron55 latest stuff is from 2012
12:05 Megaf_ I have just noticed that.
12:05 celeron55 probably doesn't even compile with latest llvm
12:06 Calinou LuaJIT is barely maintained too :/
12:06 Calinou IIRC there's no 5.3 support
12:06 Megaf_ Calinou: I get updated on my fork of LuaJIT at least once a month
12:07 Zeno` joined #minetest-dev
12:07 Megaf_ I'm on LuaJIT v2.1 - commit 5de95edf4c03296f65ff435c587719a7dff5e6f5
12:07 celeron55 Calinou: but luajit works basically perfectly and 5.3 isn't that useful anyway
12:08 Megaf_ That's the latest one AFAIK
12:08 Calinou we don't even use 5.2, I think :P
12:17 leat joined #minetest-dev
12:23 cib0 joined #minetest-dev
12:28 Zeno` E/AndroidRuntime(32447): java.lang.UnsatisfiedLinkError: dlopen failed: could not load library "libgmp.so.10" needed by "libminetest.so"; caused by library "libgmp.so.10" not found
12:28 Zeno` ^--- I thought this was fixed
12:29 leat joined #minetest-dev
12:33 Zeno` stupid internet
12:33 Zeno` I still can't push/pull/reach github 75% of the time :(
12:35 Zeno` Review please? (Trivial change)   https://github.com/Zeno-/minetest/commit/e01787d36f0a32d06200c83bf6e9df60a4bf0030
12:36 Zeno` In response to #2734
12:37 Megaf_ How is the CMakeLists.txt file generated/updated?
12:37 Zeno` Megaf_, by hand
12:37 Megaf_ Ok, that's what I thought
12:38 Zeno` (cmake may make a cached copy though)
12:40 RealBadAngel joined #minetest-dev
12:41 Zeno` hi RBA
12:42 Megaf_ Will this work? ...
12:42 Megaf_ test.txt < EOF
12:42 Megaf_ This is a test
12:42 Megaf_ EOF
12:42 Zeno` huh
12:42 Megaf_ hm, its <<
12:43 Megaf_ cat test.txt << EOF \ This is a test \ EOF
12:43 Megaf_ Ah, you know what I mean
12:44 Zeno` err shouldn't it be >> ? If you're talking about bash/shell?
12:44 Zeno` anyway take that to #minetest :)
12:44 Hunterz joined #minetest-dev
12:44 Zeno` << redirects input
12:44 Zeno` but I have NFI what you're doing :)
12:45 Megaf_ EOF Topic > #Minetest
12:47 est31 joined #minetest-dev
12:47 Zeno` Zeno`> Review please? (Trivial change)   https://github.com/Zeno-/minetest/commit/e01787d36f0a32d06200c83bf6e9df60a4bf0030
12:47 Zeno` <Zeno`> In response to #2734
12:48 est31 m_cache_hold_aux1 ?
12:48 Zeno` yeah... clumsy name :/
12:48 Zeno` better than m_cache_hold_aux1_key though lol
12:49 est31 yea
12:53 Zeno` yay or nay (to push)?
12:53 est31 yay
12:54 Zeno` before I do... what's the difference between __ANROID__ and ANDROID
12:54 Zeno` they both seem to be used :/
12:54 est31 ah
12:54 Megaf_ (insert joke here)
12:54 Zeno` I'll fix that in another commit though
12:54 est31 I think I had problems with one of them when I compiled in strict mode
12:54 Zeno` pushing that simple one now
12:54 est31 so with C++99 mode
12:55 est31 (not with ANDROID but with LINUX)
12:55 Zeno` because I know it won't cause extra problems related to __ANDROID__ vs ANDROID that we don't already have
12:55 Zeno` hmm
12:56 Zeno` does c99 even apply?
12:56 Zeno` err c++99?
12:56 Zeno` I guess __ are for reserved things (I'd have to check though)
12:56 est31 Zeno`, this is the commit https://github.com/minetest/minetest/commit/4abb96fb546b5975c74ff2cc295255fe17126287
12:57 Zeno` well nothing is currently broken so I guess the inconsistency can stay until we decide what to do
12:57 est31 yea
12:58 leat joined #minetest-dev
13:07 rubenwardy joined #minetest-dev
13:08 leat joined #minetest-dev
13:09 RealBadAngel hi Zeno`
13:09 Zeno` ;-)
13:09 Zeno` est31, can you see the logs?
13:09 Zeno` <Zeno`> E/AndroidRuntime(32447): java.lang.UnsatisfiedLinkError: dlopen failed: could not load library "libgmp.so.10" needed by "libminetest.so"; caused by library "libgmp.so.10" not found
13:09 Zeno` <Zeno`> ^--- I thought this was fixed
13:10 Zeno` Or maybe I'm "installing" via adb incorrectly... no idea
13:10 est31 this should have been fixed yes
13:11 Zeno` will it be installed by a debug build or do I have to go through the insane process of making a release apk?
13:11 est31 ah was after the lua discussion and before you complaining bout internet
13:12 est31 it should be installed.
13:12 est31 I'll try to reproduce
13:12 Zeno` ok
13:13 est31 do you have more logcat available?
13:13 Zeno` yeah
13:13 Zeno` hold on, I'll run it again and get a fresh copy
13:15 Zeno` pm'd
13:15 Zeno` brb
13:18 Megaf_ look at the size difference after removing a couple of files and directories from Minetest (git clone minetest.git)
13:18 Megaf_ 58MMinetestOriginal
13:18 Megaf_ 12MMinetestStripped
13:19 leat joined #minetest-dev
13:22 rubenwardy joined #minetest-dev
13:22 Wayward_One joined #minetest-dev
13:23 Zeno` est31, do you think I am installing it incorrectly?
13:23 Zeno` grr
13:23 Zeno` peer is annoying
13:23 Wayward_One Zeno`, you install it through adb?
13:23 Zeno` Wayward_One, yes
13:23 Zeno` adb install -r bin/Minetest-debug.apk
13:24 Wayward_One oh. i just copy the apk to my device and install it there
13:24 Zeno` hmm
13:24 est31 joined #minetest-dev
13:24 Zeno` I dunno how to do that :)
13:25 Zeno` I just have a nagging feeling that it's something *I* am doing incorrectly
13:25 Zeno` But I have no idea what
13:26 est31 no the adb method should work too
13:26 Zeno` hmm
13:26 Wayward_One the apk is in /minetest/build/android/bin. just copy it to your device, and open it in a file manager (if you're still having problems with adb)
13:27 Zeno` est31, did you get my PM?
13:27 est31 yes
13:27 Zeno` ok, cool
13:27 est31 its weird
13:27 Zeno` Wayward_One, the adb command says it succeeds
13:27 Wayward_One Ah, ok
13:27 est31 the manual call works the non manual one sais it doesnt work
13:28 TenPlus1 joined #minetest-dev
13:28 TenPlus1 Hi folks...
13:29 Zeno` let me uninstall everything and see if it persists
13:29 TenPlus1 Q. what kind of error correction does minetest have when loading maps ? does it check if it's valid or corrupt ?
13:30 TenPlus1 *error handling
13:31 Zeno` I'll clean all as well
13:33 est31 I think there is no error correction
13:33 est31 more like error detection
13:33 TenPlus1 so if minetest loads part of a map and an area/node is corrupt, it can handle that without crashing server ?
13:34 Zeno` yes it should
13:34 est31 mapgen_v5.cpp:524: error: undefined reference to 'CaveV5::CaveV5(MapgenV5*, PseudoRandom*)'
13:34 est31 I can't compile for android
13:34 Zeno` you may have to add a line to minetest.conf though
13:34 TenPlus1 am trying to figure this std::bad_alloc error that server keeps on getting
13:34 Zeno` est31, I am doing a clean make now... will see if it happens for me as well
13:36 est31 Its easiest to help you TenPlus1 with a stacktrace
13:37 TenPlus1 how do I go about doing that ?
13:37 Puma_rc joined #minetest-dev
13:38 est31 various ways, I always open the server in a gdb instance
13:38 est31 and then do bt
13:39 TenPlus1 so:   gdb minetest --server      and see if it crashes with that ?
13:39 est31 better is to use gdb minetestserver
13:39 est31 because gdb doesnt pass on arguments
13:39 Zeno` minetestserver is no longer build by default :/
13:40 est31 so --server is an argument for gdb
13:40 TenPlus1 minetestserver doesnt run on 0.4.12 now... had to use --server
13:40 Zeno` s/build/built
13:40 est31 also use a debug buld
13:40 est31 that will give you actual lines
13:40 est31 its possible to give params to the called program too
13:40 nore joined #minetest-dev
13:40 est31 lemme see how
13:40 Zeno` I really wish ShadowNinja hadn't removed minetestserver from the default build
13:41 Zeno` it doesn't add to compile time... (except maybe on travis)
13:41 Zeno` because the files (for the most part) are already compiled! All it does is link
13:42 est31 ok here: gdb --args minetest --server
13:44 TenPlus1 thanks... when owner turns up we'll try and do that for a single error instance
13:45 Megaf_ Zeno`: I had a discussion with him already and I believe that he said that it adds a lot of time to compiling time
13:46 Megaf_ Hi TenPlus1, finally I meet you, Still having that problem?
13:46 Zeno` est31, it builds for me
13:46 Zeno` installing now
13:46 TenPlus1 hi megaf, yep... std::bad_alloc errors constant now...   12 crashes in last 30 mins
13:46 RealBadAngel i wonder what is the average number of blocks loaded into memory?
13:47 Megaf_ Zeno`: TenPlus1: couldn't that be fixed with world edit/deleteblock commands? deleteblock would regenerate the corruped block
13:47 Zeno` est31, ok false alarm. After uninstalling minetest from my android, rebuilding and installing again it now works
13:47 est31 good
13:47 Zeno` sorry about that :)
13:47 est31 np
13:47 RealBadAngel anybody have an idea?
13:47 est31 my build failure is most likely false alarm too
13:48 Megaf_ TenPlus1: can you paste your minetest.conf please? (on some pastebin like http://paste.debian.net/
13:48 est31 no RealBadAngel
13:48 TenPlus1 megaf: I try to do that just now for areas we KNOW have errors, but there's too many and some are hard to pinpoint... especially if in the middle of a large build
13:48 Megaf_ TenPlus1: It looks like your map keep getting more and more corrupted every time
13:48 Megaf_ keeps*
13:49 Megaf_ over time*
13:49 RealBadAngel im thinkin about holding data for minimap (created on mapblock mesh updates - 16*16 grid of 4 bytes per mapblock) for each block actually loaded into memory
13:50 RealBadAngel and update it (or delete) when block changes
13:50 TenPlus1 http://pastebin.com/NQfs0CzF
13:50 rubenwardy Just wondering, what is the Minetest code style way of doing this: https://gist.github.com/rubenwardy/43ba16e0d46aba1c38da
13:51 Megaf_ RealBadAngel: I believe it depends on the number of players online no? For example, the max_block_send_distance, if set to 3, will load a radious of 3 blocks per client. Or not?
13:51 TenPlus1 server memory handling is really good, never goes above 600mb even after hours...
13:51 Megaf_ I know if I set that number too high it will crash my server because of out of memory
13:51 Zeno` TenPlus1, who hosts your server again? At this stage I think we *really* need a full copy of it to nail this nasty thing.
13:52 Megaf_ the irony is strong with this one > TenPlus1
13:52 TenPlus1 Titanius has the server in Wisconsin... it's well over 60gb in size and he's not that tech savvy
13:52 Megaf_ TenPlus1: on my server it avarages 260 MB of ram
13:52 Zeno` 60gb... *sigh*
13:53 Wayward_One joined #minetest-dev
13:53 TenPlus1 we tried local map saving but that doesnt nab the errors, just crashes server when it comes across them...
13:53 TenPlus1 and I cant get a map large enough on standalone server to crash it anymore
13:53 Zeno` I had database errors on my server and I set something to ignore them
13:53 TenPlus1 but I can try and have Tianius do a gdb session
13:53 Zeno` wish I could remember what
13:53 TenPlus1 that's enabled...  ignore_world_loads_errors = true
13:54 TenPlus1 doesnt work for these one's
13:54 Megaf_ TenPlus1: have you tried to start your server with it false?
13:54 Zeno` a gdb session would be brilliant. a gdb session on a debug build would be perfect
13:54 LittleJoe joined #minetest-dev
13:54 TenPlus1 so:   gdb --args minetest --server --debug    ?
13:55 TenPlus1 oops... --verbose  even
13:55 Megaf_ theres the --trace
13:55 Zeno` valgrind would be the next step and if that doesn't narrow it down massif (but both of those will probably slow the server down so much as to make it unusable while they're running :()
13:56 Zeno` TenPlus1, does Titanius ever come online?
13:56 Zeno` on IRC I mean
13:56 Wuzzy joined #minetest-dev
13:57 est31 rubenwardy, I think thats the right way
13:57 Wuzzy I wonder if this is a bug or intentional: set_properties does not seem to change anything for players, yet the lua_api.txt claims it works for *all* objects
13:57 rubenwardy Okay.
13:57 rubenwardy I guess it's best to avoid m_name, etc
13:57 Zeno` 60gb is not much for my server to handle. If you can get him to agree to transfer the whole server to my server I will give him an ssh login to perform that
13:57 est31 see here for an example https://github.com/minetest/minetest/blob/master/src/server.h#L144
13:58 rubenwardy no...
13:58 Wuzzy i.e. if me is the player object, and I call me:set_properties({makes_footstep_sound=false}), me does still do make a footstep sound
13:58 Zeno` but I suppose he can't use scp if he's not tech savvy :(
13:58 rubenwardy I meant when a parameter has the same name as a member
13:58 rubenwardy Ie: I want initialise this->x with some x parameter
13:58 est31 ah
13:58 est31 think of sth nice
13:58 rubenwardy Lol
13:58 rubenwardy Then keep to it
13:59 est31 starting with _ isn't nice
13:59 Zeno` or just do this::x(x)
13:59 Wuzzy It seems that set_properties is a no-op for player objects. Is this intentional?
13:59 rubenwardy Didn't know that'd work
13:59 TenPlus1 he works during day and time difference means I see him early morning and evening
13:59 Zeno` neither do I
13:59 Zeno` lol
13:59 Zeno` give it a try though hehe
13:59 TenPlus1 but he isnt on long...  just enough for server update / fixes
14:00 Zeno` I know it would work in the function body
14:00 Zeno` just don't know if it's ok for init lists
14:00 rubenwardy error: expected identifier before ‘this’
14:00 rubenwardy this::id(id),
14:02 Zeno` :)
14:03 rubenwardy C++11 allows x(x), and I have it enabled
14:03 rubenwardy Because it's my own project
14:04 TenPlus1 does mt run any better on c++11
14:04 est31 no
14:04 est31 not yet
14:04 est31 I'm for typedefing custom maps that can be converted to unordered map on occasions
14:05 est31 s/on occasions/when compiling with c++11/
14:05 rubenwardy I'm using it because for (auto &i : vec) is so much less to type than for (std::vector<int>::const_iterator i = vec.begin(); i != vec.end(); v++)
14:05 Zeno` rubenwardy, if it was me I'd just change _x to x_
14:06 rubenwardy That wouldn't be bad. But the standard allows x(x) etc
14:07 est31 I like the auto stuff too
14:07 proller joined #minetest-dev
14:07 est31 but people want minetest to not use c++11 till 2017
14:07 est31 or so
14:08 TenPlus1 c++11 will mature by then and hopefully have many optimizations
14:08 est31 c++11 isn't developed further
14:08 est31 next is c++17
14:08 rubenwardy ~g c++17
14:08 est31 and c++1
14:08 est31 4
14:09 TenPlus1 erk, version by version...
14:09 rubenwardy No ShadowBot, Nooo....
14:09 rubenwardy She's gone...
14:09 TenPlus1 no true standards anymore...  it better not be cobbled together
14:09 Zeno` wait a sec...
14:09 Zeno` it's not c11 is it?
14:10 est31 ?
14:10 Zeno` x(x) is c98?
14:11 Zeno` c++98 grr
14:11 rubenwardy On the SO I read it says it works in C++11
14:11 rubenwardy and that's all I needed to know :)
14:14 sofar joined #minetest-dev
14:18 TenPlus1 if a server crashes during a map write, what happens ? will it finish or just stop altoghether /
14:19 Zeno` I think it's 03
14:21 Zeno` it is
14:21 Zeno` c++98/03
14:21 Zeno` section 12.6.2.7
14:21 leat joined #minetest-dev
14:22 Darcidride joined #minetest-dev
14:22 Zeno` TenPlus1, if it crashes it crashes. If the write finishes could depend on where it is in the write, or even on the OS
14:23 Zeno` any modern OS will finish the write though *if* it's all been written (i.e. the write function is minetest is finished at the time of the crash)
14:23 TenPlus1 just curious, somehow I thought you had a separate process handling map functions, passing blocks to and from server/client to keep it safe
14:24 init if you use sqlite, without "performance tweaks", nothing will (probably?) happen
14:24 Zeno` well there is. If the client does something before the server recognises the crash and writes it then it will just be as if it never happened
14:25 Zeno` I am not convinced these bad allocs are due to database corruption
14:25 TenPlus1 is    sqlite_synchronous = 0    a tweak ?
14:25 Zeno` although it is weird that you're experiencing them so often and nobody else is
14:26 TenPlus1 I dont get it myself Zeno... typically in certain areas of the map...  could too many active objects in an area cause it ??
14:26 crazyR joined #minetest-dev
14:26 crazyR joined #minetest-dev
14:29 Zeno` I don't know :(
14:29 Zeno` The stacktrace will help :)
14:29 TenPlus1 will try to do the gdb later tonight when owner is home...
14:29 TenPlus1 thanks again folks
14:29 Zeno` fingers crossed
14:29 TenPlus1 :)  back later
14:30 Zeno` I think I may sleep now. Night all
14:30 leat joined #minetest-dev
14:36 hmmmm joined #minetest-dev
14:36 zat joined #minetest-dev
14:39 leat joined #minetest-dev
14:41 proller joined #minetest-dev
14:42 cib0 joined #minetest-dev
14:50 leat joined #minetest-dev
15:09 celeron55 looks like the screenshot-on-cover-of-publication guys ended up wanting a signature of mine
15:09 celeron55 top lel
15:10 celeron55 no way, i don't have the rights anyway and even if i had, i wouldn't do it
15:10 celeron55 it's like the year 1800 or something
15:11 est31 there is a publication featuring minetest?
15:11 leat joined #minetest-dev
15:11 est31 at the cover?
15:11 celeron55 not really
15:29 proller joined #minetest-dev
15:32 Hunterz joined #minetest-dev
15:52 ElectronLibre joined #minetest-dev
15:54 proller joined #minetest-dev
16:12 rubenwardy joined #minetest-dev
16:16 proller joined #minetest-dev
16:24 twoelk joined #minetest-dev
16:30 Krock joined #minetest-dev
16:37 Zeno` joined #minetest-dev
16:44 Hijiri joined #minetest-dev
16:44 cib0 joined #minetest-dev
17:18 rubenwardy What do you think about my macro? https://gist.github.com/rubenwardy/56a8bacc26c09207172a
17:33 Player_2 joined #minetest-dev
17:46 ElectronLibre left #minetest-dev
17:51 bobomb joined #minetest-dev
17:53 Hijiri joined #minetest-dev
17:56 Calinou “you should work at X.org!” —hmmmm
17:56 Calinou why are failed unit tests yellow, and not red?
17:56 Calinou oh I read it wrong
17:59 Krock we need colors in Minetest's console
18:02 bobomb left #minetest-dev
18:10 leat joined #minetest-dev
18:28 Calinou Xonotic and Unvanquished both support colored terminal output
18:28 Calinou but first, we need colored text in the game
18:32 Wayward_Tab joined #minetest-dev
18:38 est31 joined #minetest-dev
19:01 proller joined #minetest-dev
19:03 Hijiri joined #minetest-dev
19:13 Wuzzy How efficient are particle effects? Is it safe to add short-lived particlespawners (ca. 0.1s) every globalstep?
19:14 Calinou they are not very efficient, which is why we wanted to implement Irrlicht particles for a while
19:14 Wuzzy so I should not add particlespawners eeach tick, right?
19:15 leat joined #minetest-dev
19:15 VanessaE so long as you delete the previous ones, it shouldn't be much different than leaving a single, equivalent spawner going all the time
19:16 est31 Generally its bad to do things every ticks
19:17 VanessaE est31: generally yeah
19:17 ElectronLibre joined #minetest-dev
19:17 VanessaE but if the code is fast enough, it won't be a problem
19:17 VanessaE the real question is why do that at all?
19:17 VanessaE aside from their performance, they have another odd issue - ID numbers get reset after a server restart, so if you start a spawner, store the ID in e.g. the node's meta, restart the server, then go to where the spawner was, it won't be running and you still have to turn it "off" before you can restart it
19:17 est31 thats very bad
19:18 Wuzzy my particlespawner basically needs to move ...
19:18 VanessaE est31: try it with homedecor's faucets
19:18 hmmmm i was going to add colored console output, but I wanted to get shadowninja's logging pull request in first
19:18 VanessaE if you have say three of them running, and restart the client/server, you can't restart all three of those spawners until you turn them all off
19:18 VanessaE (if you start one, one of the others will turn off at the same time)
19:19 est31 hmmmm, whats missing on the logging PR for you?
19:19 VanessaE (I know how to work around this for homedecor, just haven't done it yet)
19:19 hmmmm not so much missing as having stuff i totally don't approve of
19:20 Wuzzy particlespawners get deleted automatically  when their timer runs out, right?
19:20 hmmmm i like the fact that lua's print() got fixed, i like the warningstream, i like the string enumeration for log levels, but not so much other stuff
19:25 jin_xi joined #minetest-dev
19:25 SopaXT joined #minetest-dev
19:26 leat joined #minetest-dev
19:30 est31 so what is exactly your usecase for dstream?
19:33 hmmmm just print things out to stdout, stderr, or a logfile without timestamp crud
19:33 hmmmm and it has to be separate from the logging facilities
19:34 hmmmm my own use case is... debugging
19:34 est31 so you need it for the unittests?
19:35 est31 or is your usecase only the "I make a printf here to test it runs through this line"
19:35 leat joined #minetest-dev
19:36 hmmmm both unit tests and the latter
19:36 hmmmm preferably, I'd like dstream to be no-overhead if debug logging is disabled
19:37 est31 no-overhead?
19:37 hmmmm maybe read the configuration entry into a global variable and replace dstream with a macro with if (g_use_debugging) { dstream << "stuff here"; } which would completely avoid writing dstream stuff
19:37 hmmmm but it shouldn't be dependent upon the config
19:37 hmmmm hmm
19:38 hmmmm the whole point of dstream is to provide a very low-level logging facility for minimal interaction with other components
19:38 hmmmm because the whole point of it is to be reliable for debugging
19:38 est31 I can understand it for unit tests
19:39 est31 but if you debug, why not use cout then?
19:39 hmmmm because cout can't write to a file without taking all the input
19:40 hmmmm though I wonder how necessary this is
19:41 hmmmm because in theory, dstream is much more verbose than other logging levels
19:41 est31 so dstream is an extra logging level for you?
19:41 hmmmm you know what
19:42 hmmmm i think we can remove dstream
19:42 hmmmm it's not absolutely necessary to separate dstream output from other outputs
19:42 hmmmm but we can also do that by using std::cerr
19:43 hmmmm std::cerr should be used in place of dstream then
19:43 est31 so why do we log at all outside of our loglevels?
19:44 est31 I mean the only reason why to log outside is 1. because its nicer to look at and 2. its more reliable
19:44 hmmmm well there are two reasons
19:45 est31 but 1. is only the case for unittests
19:45 hmmmm anywhere, really... none of the current dstream uses want a timestamp
19:46 est31 and 2. is a reason yes, but I think no reason for code thats inside git
19:46 est31 but only when you want to debug logging itself
19:46 est31 there you can use cout too
19:46 est31 or cerr
19:46 est31 with a nice prefix
19:48 est31 and SN has added g_log_outputs_stream, so that can be used
19:49 est31 also what are your thoughts bout #2559 hmmmm?
19:49 est31 settings aren't available yet when we start logging, are they?
19:50 est31 gtg
19:50 est31 by
19:50 est31 e
19:56 kahrl I like using dstream for number crunching purposes
19:57 kahrl say, when there's a slow function that is called often, I may do dstream << "whatever_function took " << timetaker.getTimerTime() << std::endl;
19:57 kahrl and then run grep + some other commands on debug.txt to get meaningful statistics from that
19:58 kahrl so this is nice because it's written to a file, and the lack of timestamps makes extracting the data easier
20:08 Wuzzy does Minetest have some profiling utilities for mods?
20:10 VanessaE yes actually
20:10 VanessaE there's a build-in profiler, not sure how to use it though
20:11 VanessaE use this:  https://github.com/t4im/profiler
20:11 VanessaE (it works a bit better than the in-built one)
20:19 VanessaE hmmmm: you given any further thought about colored light sources?
20:23 geheimnis joined #minetest-dev
20:26 leat1 joined #minetest-dev
20:38 nolsen joined #minetest-dev
20:38 Hijiri joined #minetest-dev
20:46 zat Is there a profiler for Lua like freeminer's?
20:57 leat1 joined #minetest-dev
21:04 hmmmm zat: yes
21:04 hmmmm vanessae: not recently, no
21:06 VanessaE ok.
21:07 hmmmm hmm
21:08 hmmmm I wonder if there's an open source version of jira
21:08 hmmmm that would be useful for organizing these things
21:10 leat1 joined #minetest-dev
21:13 rom1504 gitlab
21:16 hmmmm est31: yeah, I like 2559
21:17 leat1 joined #minetest-dev
21:25 hmmmm ever since i upgraded to freebsd 10 i've been having problems combining certain libraries with certain others
21:25 hmmmm can't link -lIrrlicht together with anything GTK-related or else i get an instant segfault in libGL
21:25 hmmmm dammit
21:28 err404 joined #minetest-dev
21:29 ElectronLibre left #minetest-dev
21:37 zat hmmmm: how do I get it and use it?
21:40 proller joined #minetest-dev
21:40 kahrl what happened to lists.minetest.net?
21:41 kahrl I get a domain squatter like page when I go there
21:41 sfan5 parked domain
21:41 sfan5 wat
21:41 hmmmm zat, builtin/profiler.lua iirc
21:42 sfan5 looks like someone forgot to renew something
21:42 hmmmm no idea how to use it.  i never used it myself
21:42 hmmmm how on earth could a subdomain get squatted
21:43 sfan5 hmmmm: maybe it used different dns servers for some reason
21:43 sfan5 lists.minetest.net.600INCNAMEsmtp.inchra.net.
21:43 sfan5 or that
21:44 sfan5 ShadowNinja let inchra.net expire
21:49 sofar he should not make it a CNAME then
21:49 sofar just an A record to the host
21:50 sfan5 but in that case ShadowNinja would have to bug c55 when the IP changes
21:50 sfan5 also mails require domain names
21:50 sfan5 you need an MX record
21:51 sfan5 and you probably want a TXT record with SPF
21:51 sofar where's mailman hosted then?
21:51 sofar or, whatever list daemon it used
21:51 sfan5 looking at where the CNAME points i'd guess it's hosted on ShadowNinja's server
21:52 sofar my bet is that the smtp still works as it redirects to p.nsm.ctmail.com
21:56 sofar unless he also unplugged the smtp :^)
21:57 sofar this is how I ended up hosting 25 mailinglists for various projects at one point in time... lol
22:11 Amaz joined #minetest-dev
22:17 VanessaE sfan5: indeed, inchra.net domain expired a few days ago
22:24 asl97 joined #minetest-dev
22:46 proller joined #minetest-dev
22:58 est31 joined #minetest-dev
23:02 asl97 left #minetest-dev
23:57 est joined #minetest-dev

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