Minetest logo

IRC log for #minetest-dev, 2015-06-21

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

All times shown according to UTC.

Time Nick Message
00:00 ekem joined #minetest-dev
00:03 RealBadAngel ok
00:03 RealBadAngel how that feels? https://imgrush.com/Qh7JBrh-8dvp.png
00:04 Kray joined #minetest-dev
00:07 VanessaE joined #minetest-dev
00:07 eeew joined #minetest-dev
00:19 Taoki RealBadAngel: Looks very nice :)
00:19 Taoki On a different note: https://github.com/minetest/minetest/issues/2811
00:20 Taoki If anyone wishes to look into that sometime
00:37 chchjesus_ joined #minetest-dev
00:41 Kray joined #minetest-dev
00:41 rom1504 joined #minetest-dev
00:41 CraigyDavi joined #minetest-dev
00:42 ShadowBot` joined #minetest-dev
00:44 technoma` what's the word for when you have keys that act as modifiers or non-modifiers depending on whether they're held or tapped?
00:44 technoma` oops; wrong channel
00:44 kahrl_ joined #minetest-dev
00:45 RealBadAngel *** Error in `./bin/minetest': corrupted double-linked list: 0x00007fd0a000ecd0 ***
00:45 werwerwer joined #minetest-dev
00:45 RealBadAngel who feels responsible for that?
00:48 Eivel_ joined #minetest-dev
00:55 sofar_ joined #minetest-dev
00:55 OldCoder joined #minetest-dev
00:58 johnnyjoy joined #minetest-dev
01:06 Taoki OldCoder: Poke
01:06 Taoki (sorry, in the other channel #minetest , this isn't dev related)
01:11 RealBadAngel Taoki, when you are checkin you should rather PEEK instead of POKE
01:11 Taoki That works :)
01:12 RealBadAngel but that way it looks like youre bashin some1 in the head with a hammer asking: alive or not?
01:13 RealBadAngel segfault will means hes not alive ;)
01:27 technomancy left #minetest-dev
01:42 Wayward_Tab joined #minetest-dev
01:43 VanessaE joined #minetest-dev
01:52 VanessaE joined #minetest-dev
02:06 OldCoder RealBadAngel, the peek is to poke
02:07 VanessaE joined #minetest-dev
02:07 RealBadAngel OldCoder, im refering to old days and ZX Spectrum Basic
02:08 RealBadAngel peek is to check, poke is to write
02:09 RealBadAngel a memory cell, ofc
02:11 * OldCoder knows
02:11 OldCoder Those were the new daya
02:11 OldCoder Those were the new days *
02:12 RealBadAngel so more corect should be:
02:12 RealBadAngel if PEEK OldCoder > 0 then write "Hello" ;)
02:12 OldCoder peek = poke(1)
02:12 OldCoder Yes
02:12 OldCoder No
02:13 OldCoder If peek(OldCodger) > 0 then poke(OldCodger)
02:13 OldCoder If peek(OldCodger) > 0 then poke(OldCodger,"0x01")
02:13 RealBadAngel "codger" rotfl ;)
02:28 paramat joined #minetest-dev
02:30 paramat hmmmmm pushing soon https://github.com/minetest/minetest/pull/2789 'Mgv7: Edit noise parameters'
02:37 Miner_48er joined #minetest-dev
03:13 sofar joined #minetest-dev
03:24 Zeno` joined #minetest-dev
03:25 Zeno` Does anyone understand the requirement for this do-while: https://github.com/minetest/minetest/pull/2795/f​iles#diff-65f34680878a6bd86f3a59ebc0c06c6dR3373
03:25 zat joined #minetest-dev
03:26 VanessaE zeno! *pounce*
03:26 Zeno` oh oh
03:26 * Zeno` hides
03:26 deltib joined #minetest-dev
03:27 Zeno` maybe it's a strange way to avoid using goto
03:28 VanessaE I'd have done "while true; do ... done" (or whatever the exact C++ syntax is)
03:28 Zeno` this is do { } while (false); though
03:28 Zeno` it must be to avoid goto
03:28 deltib joined #minetest-dev
03:30 VanessaE good point
03:30 Wayward_Tab joined #minetest-dev
03:32 paramat now pushing 2789
03:33 Zeno` lol (ironically) it would have been better and clearer to use goto
03:33 Zeno` break; is effectively a C-type goto (i.e. local jump) anyway
03:35 deltib joined #minetest-dev
03:35 Wayward_Tab joined #minetest-dev
03:36 RealBadAngel Zeno`, have you saw report about lists crashes?
03:36 Zeno` no..
03:37 paramat complete
03:37 Zeno` RealBadAngel, link?
03:38 RealBadAngel Zeno`, http://pastebin.com/retyxwFX
03:38 RealBadAngel thats all
03:39 paramat please could the HP bug be fixed soon? drives me nuts =)
03:41 Zeno` RBA, will look into it
03:42 Zeno` for everybody's info: I've been away because first my modem broke and I had to wait for a new one; and then it turned out it wasn't my modem after all and I've had no internet (basically)
03:42 Zeno` but I'm back now :)
03:42 Zeno` I'll do the HP fix as well.. pretty sure I have a patch already
03:43 Zeno` RealBadAngel, hoow do you know that's from lists?
03:43 Zeno` and by lists do you mean these inventory list things?
03:43 RealBadAngel ask the one who reported it
03:44 RealBadAngel im just passing it by
03:44 Zeno` yeah, just noticed that hehe
03:44 Zeno` I wish people would supply test cases
03:44 paramat cool =)
03:44 Zeno` I hate writing Lua (hehe)
03:44 Zeno` I suppose I could use world edit
03:45 RealBadAngel btw, https://imgrush.com/tuH6QNkKtYAO.png
03:45 paramat good to have you back
03:45 RealBadAngel its almost complete
03:45 Zeno` looks awesome, RBA! :)
03:45 RealBadAngel paramt, i think thats the reason for you to start using shaders
03:45 Zeno` paramat, the break did me good anyway
03:45 RealBadAngel you will be able to see your mapgen in action way better
03:46 Zeno` shaders do not work at all on Android for some reason :(
03:46 Zeno` not on any device or emulator I've tested anyway
03:46 RealBadAngel shaders  need GPU, thats why :)
03:46 paramat RBA is it possible for the minimap lighting to always come from top or top-left even when it rotates?
03:47 paramat looks really good
03:47 RealBadAngel yeah, you can just set its position
03:47 paramat okay, from above is best
03:47 paramat erm from top-left would be best
03:48 RealBadAngel atm its 0, 0.5, 1.0
03:48 Zeno` RealBadAngel, my android has a GPU :p
03:48 RealBadAngel Zeno`, or you were just told that ;)
03:49 Zeno` lol
03:49 RealBadAngel unless we are talking bout x86 android and real box
03:49 Zeno` I have nfi what it is
03:49 Zeno` Samsung Galaxy S6
03:49 Zeno` pretty sure it's ARM
03:50 RealBadAngel i will take a closer look on this devices (i do have galaxy tab 3)
03:50 RealBadAngel but atm i do have more important things to code
03:51 Zeno` true enough
03:59 Zeno` Ok, I was really just dropping in to tell everyone why I'd be away. I will be back later (properly) tonight, and tomorrow onwards :)
03:59 Zeno` O/
04:04 hmmmmm paramat:  Looks okay to me I guess
04:04 hmmmmm really never even thought about the periodicity of mgv7
04:05 hmmmmm Zeno`:  whenever you see do { ... } while (0), it should probably be a separate function :p
04:06 paramat cool
04:06 Zeno` hmmmmm, I would agree ;-)
04:06 Zeno` but even goto would have been better than the do while construct hehe
04:06 hmmmmm perhaps yeah
04:07 hmmmmm I think keeping large blocks of code like that out of a separate function is a Carmack-ism
04:07 hmmmmm this is something he strongly advocates that I can't really get my head around
04:08 hmmmmm it's not advantageous from an optimization standpoint and atrocious from an organizational one.
04:10 Zeno` I dunno what's causing it
04:10 Zeno` there seems to be an aversion to using functions; I raised it once with Warr and he said he didn't make it a function because it would only be used once (like the only reason for functions is reusability :/)
04:11 Zeno` I like what the Linux Coding Guidelines say about functions
04:11 Warr1024 I don't have an aversion to functions, I have an aversion to large, difficult-to-review diffs.
04:11 Zeno` It wasn't you anyway, I was about to correct myself
04:12 Zeno` :)
04:12 Zeno` I think maybe it was est31... don't really know no
04:12 hmmmmm celeron was the one who said that I thought
04:12 Zeno` now*
04:12 Zeno` maybe it was celeron... it was somebody
04:12 Wayward_Tab joined #minetest-dev
04:12 Zeno` doesn't matter who it was really :)
04:12 hmmmmm in any case, not everybody has code folding capabilities in their editor, nor is it already named as such
04:12 Zeno` But there does seem to be a general aversion to using functions :/
04:13 Warr1024 I won't make code a function if it's not reusable AND the function it lives in is small enough even with it included.
04:13 hmmmmm erm, I mean nor is the code block named
04:13 Zeno` https://www.kernel.org/doc​/Documentation/CodingStyle <--- chapter 6
04:13 Zeno` nice summary (honestly)
04:13 hmmmmm when you start having multiple bail-outs in a do { ... } while (0), that's a pretty damn good candidate for a function imho
04:14 Warr1024 vi does not do code folding, last I checked.
04:14 hmmmmm there's always a plugin
04:14 Warr1024 I'm fair sure vi doesn't do plugins either.
04:14 Zeno` vim does
04:14 Zeno` or maybe that's elvis
04:14 Warr1024 vim is not vi
04:14 Zeno` correct
04:15 Zeno` vi does only vi stuff :)
04:15 Zeno` vi as in class vi :)
04:15 kahrl_ IIRC, ed also has no code folding
04:15 Zeno` classic*
04:15 Zeno` ed doesn't last I looked
04:15 Zeno` LOL
04:15 Warr1024 vim cannot be used to fix a busted fstab on a system that failed to mount /usr, and thus I still have a reason to hone my plain-vi skills ;-)
04:16 Zeno` that's why all real hackers use nano
04:16 Warr1024 haha
04:16 Zeno` :-D
04:19 Zeno` I hate code folding anyway
04:19 Zeno` I alway forget that I've got stuff hidden hehe
04:20 Zeno` the only place I think it is for the file headers (as in the big (C) messages at the top of files)
04:20 Warr1024 I'm nota fan of big (C) mesages either ;-)
04:20 Zeno` anyway, I'm rushing now... got an appointment. Must go. BBL. O/ :)
04:32 deltib joined #minetest-dev
05:05 paramat left #minetest-dev
05:37 deltib joined #minetest-dev
05:39 Megaf joined #minetest-dev
06:38 NakedFury joined #minetest-dev
07:06 AnotherBrick joined #minetest-dev
07:09 nore joined #minetest-dev
07:25 Hunterz joined #minetest-dev
08:00 Yepoleb_ joined #minetest-dev
08:04 Calinou joined #minetest-dev
09:35 Jbb joined #minetest-dev
10:05 Krock joined #minetest-dev
11:05 selat joined #minetest-dev
12:31 SopaXT joined #minetest-dev
12:36 Warr1024 joined #minetest-dev
12:45 Hunterz joined #minetest-dev
12:53 Amaz joined #minetest-dev
13:01 Kray_ joined #minetest-dev
13:03 Naked joined #minetest-dev
13:04 rom1504_ joined #minetest-dev
13:07 Eivel_ joined #minetest-dev
13:13 CraigyDavi joined #minetest-dev
13:14 Calinou joined #minetest-dev
13:17 cvtsx1 joined #minetest-dev
13:21 RealBadAngel joined #minetest-dev
14:24 rubenwardy joined #minetest-dev
14:43 MinetestForFun joined #minetest-dev
14:57 RealBadAngel hi guys
14:57 RealBadAngel kahrl, hmmm around
14:57 RealBadAngel ?
15:03 Krock kahrl_ ^
15:20 hmmmm joined #minetest-dev
15:30 CraigyDavi Anyone going to merge https://github.com/minetest/minetest_game/pull/541 ?
15:37 zat joined #minetest-dev
16:38 MinetestForFun joined #minetest-dev
16:39 MinetestForFun_ joined #minetest-dev
17:08 ElectronLibre joined #minetest-dev
18:00 Taoki Weird. If I set_attach() a Lua entity to a bone on the player, it seems to lag and only move slowly as the player moves back or forth. Is this a bug?
18:00 Taoki It doesn't happen if I set_attach to the player without specifying a bone
18:01 Taoki Why the smooth movement?
18:07 Taoki It only seems to happen randomly as well... bizarre
18:09 VanessaE wield3d mod has the same problem
18:13 rubenwardy joined #minetest-dev
18:15 Taoki It's clearly a bug in the code, yeah x_x
18:15 Taoki Either Minetest's or Irrlicht
18:16 Taoki Since local attachment processing uses an Irrlicht function, the second is pretty probable as well
18:16 Taoki But it only happens sometimes. I wonder why
18:36 CraigyDavi joined #minetest-dev
18:45 CraigyDavi joined #minetest-dev
18:45 VanessaE celeron55: *poke*
18:46 CraigyDavi joined #minetest-dev
18:51 VanessaE celeron55: you're needed for a forum bug.
18:56 Krock Yay, create more bugs, that's fine!
19:09 * Sokomine suffers from a concert in the neighbourhood
19:09 Sokomine any clues as to protect the ears from too loud bass music? the windows are already closed...
19:09 Sokomine watching a film louder than i normally would doesn't help either
19:11 Naked joined #minetest-dev
19:46 MinetestForFun joined #minetest-dev
19:49 proller joined #minetest-dev
19:50 CraigyDavi joined #minetest-dev
19:55 CraigyDavi joined #minetest-dev
20:04 CraigyDavi My server has a memory problem, it keeps increasing until it crashes when it runs out of memory. Is there any way to find out what mod/function is causing the problem? Would valgrind help or should I try removing mods one by one?
20:28 sofar all of lua runs in the same context I think
20:29 VanessaE CraigyDavi: remove all mods and start *adding* them one by one
20:29 VanessaE start with the ones you suspect, first... if there's a memory leak in a mod, you'll find it faster that way
20:30 CraigyDavi Thanks, I'll try that!
20:35 proller joined #minetest-dev
20:37 troller joined #minetest-dev
20:37 proller joined #minetest-dev
20:49 sofar I may have something that will help
20:51 sofar CraigyDavi: https://gist.github.com/sofar/41d1e70893f16d457ff1
20:51 sofar CraigyDavi: throw that in an init.lua mod, enable it and type /dump in game
20:51 ElectronLibre left #minetest-dev
20:51 sofar CraigyDavi: it outputs a dump of the _G global table and estimates the size of subtables
20:52 sofar CraigyDavi: e.g. G[moreblocks]=(table)=26
20:52 sofar meaning, the "moreblocks" table is in the order of 26 bytes large
20:52 sofar the size is a total hack, but it's a good proportional estimate
20:52 CraigyDavi Thank you!
20:52 sofar CraigyDavi: paste the output - I'd like to see what you get out of it :^)
20:54 sofar hmm
20:55 sofar I don't think its ready just yet
20:55 sofar hold on
20:55 CraigyDavi Ok
20:56 est31 joined #minetest-dev
21:00 sofar almost ready with a real working version
21:02 sofar nope, not yet - hangs :^)
21:06 Puma_rc joined #minetest-dev
21:09 proller joined #minetest-dev
21:10 sofar it's surprisingly hard to determine the memory size of a table object, lol
21:11 CraigyDavi I see
21:15 sofar well hmm, I can make it dump everything now
21:15 sofar but it's io.write'ing it all and I just want to measure the size
21:16 Amaz joined #minetest-dev
21:18 sofar ok, I think I have it done now
21:19 sofar lemme test one more time
21:20 sofar yay, it works - it's very very slow but shows me correct size estimates now
21:20 sofar G[package]=(table)=1002560
21:20 sofar G[core]=(table)=769851
21:20 sofar G[minetest]=(table)=769851
21:20 sofar G[circular_saw]=(table)=2927
21:21 sofar CraigyDavi: https://gist.github.com/sofar/88435ba1fad327e9b76a
21:22 sofar CraigyDavi: it may take minutes to run depending on the size of the objects
21:22 CraigyDavi I'll try it
21:22 sofar just let it finish. big tables like "minetest" take 5-10 seconds to parse/estimate
21:23 Puma_rc joined #minetest-dev
21:26 CraigyDavi uh haha, it crashed due to low memory - while running a command trying to find out why memory usage increases
21:26 CraigyDavi How ironic
21:29 sofar heh
21:30 sofar anything it dumped before that may be useful?
21:30 sofar oh hey
21:30 sofar what was the last line before it crashed?
21:31 CraigyDavi https://gist.github.com/Crai​gyDavi/2d7059613a148ee9fc3f
21:31 CraigyDavi Shows that mt_irc is quite large
21:31 CraigyDavi IsPlayerOwner is last line
21:32 CraigyDavi *IsPlayerNodeOwner
21:33 sofar so what you can do
21:33 sofar remove the print_r(n) part
21:33 sofar and run it again
21:34 sofar and compare the order and if it's the same, then obviously the one after isPlayerNodeOwner is a likely culprit
21:34 sofar you want to remove
21:34 sofar ..string.len(print_r(v))
21:36 sofar CraigyDavi: 5000 is not large
21:36 sofar CraigyDavi: minetest itself is like 800000 on my base server with only a few mods
21:39 CraigyDavi sofar, here's the output https://gist.github.com/Crai​gyDavi/8e70de6cd04bebc176fb
21:40 sofar G[travelnet] is the suspect
21:40 VanessaE Sokomine: ^^^^
21:40 CraigyDavi So that's using the highest memory...or has a memory leak?
21:40 sofar remove it, put the code back, run /dump again
21:40 Puma_rc joined #minetest-dev
21:41 VanessaE sofar: you oughta run your script on Dreambuilder so time :P
21:41 sofar it's large enough to cause dumping it to make your box go OOM
21:41 VanessaE I'm sure it has its share of mem leaks
21:41 sofar CraigyDavi: if your box is really low on memory it may just point at other high-memory using modules
21:41 sofar but it's suspicious that it can't serialize the travelnet table
21:43 CraigyDavi Here's it without the travelnet mod: https://gist.github.com/Crai​gyDavi/93e9a0f73834b6ef649c
21:43 CraigyDavi It has 1GB of ram and that's always been enough when I have run the server on higher spec servers
21:48 CraigyDavi What now? I can't do without the travelnet mod!
21:48 VanessaE add more memory? :P
21:49 CraigyDavi eh, maybe :)
21:50 CraigyDavi Rather get the memory leak fixed though
21:52 VanessaE fwiw I run 8 server instances on 24 GB, and the lot of them (two or three of which use travelnet) generally max out at 3.5 to 4 GB
21:52 sofar CraigyDavi: you re-rean it but you didn't put the print_r code back
21:53 RealBadAngel joined #minetest-dev
21:53 VanessaE aw shit
21:53 VanessaE everyone scatter
21:56 proller joined #minetest-dev
21:59 CraigyDavi Well it crashes again when I put the print_r back
21:59 CraigyDavi last line was IsPlayerNodeOwner
21:59 VanessaE areas mod?
22:00 CraigyDavi Then I removed print_r and the line after that was skins
22:00 CraigyDavi yeah areas is installed
22:00 CraigyDavi Although I see no reference to IsPlayerNodeOwner there
22:01 VanessaE then the old node ownership mod
22:01 VanessaE one of the two contains that function
22:01 VanessaE or glomie's protector mod
22:02 CraigyDavi But I don't have any of those installed
22:04 VanessaE *shrug* I'm out of ideas :)
22:07 CraigyDavi Ah, found reference to it in areas! https://github.com/CraigyDavi/Craig-Serv​er_game/blob/c2d45efb4173ed099eeb6bd2059​1f3da31412bcb/mods/areas/legacy.lua#L108
22:07 VanessaE :)
22:09 CraigyDavi So areas.isNodeOwner might cause the problem...I'm confused...
22:10 sofar CraigyDavi: do you have access to a system with more memory? you could duplicate your server setup and see if the table dump completes on there?!
22:11 CraigyDavi Good idea, although it will be in singleplayer rather than on a server
22:12 sofar as long as all the mods load and the map is the same
22:19 CraigyDavi huh? Not enough memory? I have over 3GB free...
22:20 CraigyDavi Well it got past travelnet this time
22:20 sofar how far did it get now?
22:20 sofar what was the size of travelnet? paste the dump?
22:21 Robby joined #minetest-dev
22:21 CraigyDavi https://gist.github.com/Crai​gyDavi/8f06d0d05daa8e1d7982
22:21 sofar so then skins is suspect
22:21 Aaron1011 joined #minetest-dev
22:21 Aaron1011 joined #minetest-dev
22:21 RealBadAngel VanessaE, no point. i can see you on the minimap ;)
22:24 CraigyDavi Ok, so simple_skins mod seems to be the problem then. Thanks sofar!
22:24 sofar CraigyDavi: well, it's "suspect"
22:26 CraigyDavi Well just looking at the code it looks like it's creating an array with 1000+ entries, that can't be good
22:29 * sofar fixes completely unrelated issues in one of his C code projects
22:38 VanessaE RealBadAngel: you're lagging ;P
22:38 RealBadAngel i turned pc on and went to make a coffee, thats why
22:39 RealBadAngel i was thinkin about minimap scaling, maybe depending only on y would work?
22:40 VanessaE it doesn't take THAT long to craft a coffee maker .. I mean you just need glass, plastic, a heating element, and a bucket of water.......
23:05 Robby joined #minetest-dev
23:39 zat joined #minetest-dev
23:41 Taoki So... can https://github.com/minetest/minetest/pull/2372 be merged in the end? I really don't want to fix more merge conflicts if another change occurs, plus it's something people have requested since I implemented the animation system.
23:42 est31 hmmmm, kahrl RealBadAngel sfan5 ShadowNinja can sb give a second +1?
23:46 VanessaE I have to agree with you, Taoki... that we need to leave irrlicht 1.7.x way behind
23:46 est31 there are ifdefs around there
23:46 est31 so on irrlicht 1.8, it will still work
23:46 VanessaE I'm still reading the comments :P
23:46 est31 err 1.7
23:47 VanessaE but we *still* need to leave it behind
23:47 VanessaE everyone has 1.8 by now, or can get it
23:49 RealBadAngel est31 ive already checked that, so go on

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