Minetest logo

IRC log for #minetest-dev, 2014-09-06

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

All times shown according to UTC.

Time Nick Message
00:13 OldCoder joined #minetest-dev
00:19 VanessaE damn it.
00:19 * VanessaE kicks HLuaBot
00:21 * VanessaE looks around and discretely sweeps the scattered bolts under the nearest table
00:28 ShadowNinja https://github.com/minetest/minetest/pull/1606
00:30 VanessaE "Use require() or anything in the package library (require() and some package functions can load C modules, the rest are only useful if you have access to the other functions)."   <--- how will this affect the IRC mod?
00:36 VanessaE "This isn't quite ready yet, you can still load bytecode if init.lua is a bytecode file"  <---  Is there any legit use case for bytecode files in Minetest?
00:37 HLuaBot joined #minetest-dev
00:37 ShadowNinja VanessaE: Hiding the source code.
00:37 VanessaE besides that.
00:37 VanessaE mauvebic is not a legit use case :P
00:38 ShadowNinja VanessaE: Super-optimizing things (SoniEx does that).
00:38 VanessaE ehm
00:39 ShadowNinja You can do things like foo[1]:bar() in bytecode.
00:39 VanessaE interesting.
00:40 VanessaE frankly I'm not one to suppose closed-source at all but that's not what your pull is about, so better to approach that some other time.
00:41 VanessaE s/suppose/support/
01:00 Cylus_ joined #minetest-dev
01:05 SoniEx2 joined #minetest-dev
01:12 Miner_48er joined #minetest-dev
02:02 paramat left #minetest-dev
02:35 AnotherBrick joined #minetest-dev
02:36 CraigyDavi joined #minetest-dev
03:06 OldCoder joined #minetest-dev
03:28 sol_invictus joined #minetest-dev
03:37 zat Simple deco_types DON'T respect biomes option!!!!!!!!!!!!!!!!!!!!
03:44 gentoobro with mapgen v7?
03:45 gentoobro they work for me, but there was some strange behavior i found dealing with decorations that were buildable-to combined with 'falling' mg7 dust nodes
03:45 gentoobro basically, a dusting of snow or sand would "squash" my grass
03:46 zat yes they work you are right lol
03:46 zat uh
03:46 zat schematics are the real buggy ones
03:47 gentoobro i haven't experimented much with those; my plan is to use special placeholder nodes and a one-time abm
03:47 gentoobro kind of like moretrees
03:51 zat gentoobro: I am doing that right now
03:57 zat joined #minetest-dev
04:07 gentoobro how's it working out for you?
04:16 kaeza joined #minetest-dev
04:21 zat gentoobro: horrible
04:40 gentoobro :(
04:51 zat really, its a ugly hack
04:55 gentoobro schematics in general, or this application of them?
05:03 zat using normal decos with abms
05:06 Zeno` joined #minetest-dev
05:26 nore joined #minetest-dev
07:10 Hunterz joined #minetest-dev
07:11 CraigyDavi joined #minetest-dev
07:44 PenguinDad joined #minetest-dev
07:48 psedlak joined #minetest-dev
07:56 Krock joined #minetest-dev
08:11 Megaf joined #minetest-dev
08:26 Zeno` what is legacy_mineral? In minetest.register_node()
08:26 Zeno` e.g. https://github.com/minetest/minetest_g​ame/blob/master/mods/default/nodes.lua
08:31 Krock legacy is to support older worlds
08:38 Zeno` So, I guess it doesn't harm if it's left there (will just be ignored)
08:38 Zeno` Thanks
08:48 Taoki[mobile] joined #minetest-dev
09:12 CraigyDavi` joined #minetest-dev
09:18 casimir joined #minetest-dev
09:25 Zeno` so much fun
09:39 Zeno` geez
09:39 Zeno` Why didn't he just implement the SIMPLE solution where on_generated callbacks can be ordered?
09:40 Zeno` That SOLVED the problem
09:40 Zeno` Now there is some problem that nobody seems to even be aware of (well, very few people)
09:41 Zeno` I suppose the modern trend is to do it the hard way
09:41 Zeno` no K.I.S.S
09:42 Zeno` just add crap until it doesn't work hehe
09:42 Zeno` obvious solution is obviously no good
09:49 Zeno` who peer-reviewed and approved f3eefeb7948b8b8d1a98f2f89baa39abc807f72d, 9e4e7072da8f565eef37da7558053a436b9cbba3 and 3fa4f782d90dac0d800251a9ab0f0afb9d32560c
09:49 Zeno` I can't see a pull request
09:49 Zeno` I thought 2 core devs had to approve pull requests
09:52 T4im joined #minetest-dev
09:52 CWz joined #minetest-dev
10:26 Benja joined #minetest-dev
10:31 Amaz joined #minetest-dev
11:21 ImQ009 joined #minetest-dev
11:38 OldCoder joined #minetest-dev
11:41 Zeno` who peer-reviewed and approved f3eefeb7948b8b8d1a98f2f89baa39abc807f72d, 9e4e7072da8f565eef37da7558053a436b9cbba3 and 3fa4f782d90dac0d800251a9ab0f0afb9d32560c <--- Can these commits be taken out of master?
11:45 Krock *reverted
12:27 PilzAdam joined #minetest-dev
12:40 sol_invictus can anyone give me a hint on why /clearobjects causing pos over limit exception when generation limit is lower than 31000?
12:41 sol_invictus I was thinking the cause is the *blocks* over the limit being processed, but it's not
13:20 AnotherBrick joined #minetest-dev
14:00 RealBadAngel Zeno`, why do you want those commits to be reverted?
14:10 proller joined #minetest-dev
14:36 NakedFury joined #minetest-dev
14:38 VanessaE RealBadAngel: because they broke something about the way vmanips behave, in that they are now causing server crashes
14:39 VanessaE ERROR: An unhandled exception occurred: caught (...)
14:39 VanessaE [C]: in function 'get_data'
14:39 VanessaE stuff like that, unless he's found a new bug, that is.
14:41 casimir It's #1607
14:43 VanessaE ok then it's two bugs.
14:44 proller joined #minetest-dev
14:44 Krock RealBadAngel, one (or more) (or all together) caused a new bug which destroys by-lua-generated terrain
14:45 Krock nyan cats are not affected. only VM stuff
14:54 VanessaE nice.
14:55 VanessaE zeno and I discussed this earlier (in /msg) as well
14:55 VanessaE so the crash is something new, related or not idk..
15:02 khonkhortisan joined #minetest-dev
15:03 Anchakor_ joined #minetest-dev
15:33 sapier joined #minetest-dev
15:34 sapier Hello do we have more information about that malformed packet crash? because according to the dump crash happens in send thread not in receive thread?
15:35 VanessaE sapier: look closer - there were two crashes - one in the send thread and one in the receive thread.
15:35 VanessaE at separate times, on separate servers.
15:35 sapier ah
15:36 VanessaE I can't confirm if the malformed packets were the cause; all I got from it was that dmesg dump after the fact.
15:36 sapier I've already found something wrong if a packet with size 0 is received but I'm not sure how this is supposed to cause a crash
15:44 sapier I've pushed the fix, yet I don't really believe this was reason for it
15:44 Krock joined #minetest-dev
15:51 sapier hmm I don't see any way a empty packet could cause a crash :-( VanessaE if it's still happening can you try to find out what has to happen and how to reproduce?
15:51 VanessaE sapier: it's totally random
15:51 VanessaE seems like every time that happens, I'm asleep or afk
15:52 VanessaE and it may not be network related at all - only reason I highlighted you was that both crashes were preceded by those malformed packets.
15:53 VanessaE I had one crash the other day out of the blue that was a sudden, extremely high memory usage (as in, tried to allocate tens of GB all at once)
15:53 VanessaE OOM killer caught it
15:54 sapier what's size of textures on that server?
15:54 VanessaE they're just normal defaults.
15:54 VanessaE most are 16px
15:54 VanessaE a couple dozen 64px inventory images, that's about it
15:55 sapier ok so our known attack most likely wont work
15:55 VanessaE (used for stuff where the client can't render the model correctly, that sorta thing)
15:56 VanessaE heh, no
15:56 VanessaE this was just...a couple guys were digging some stone, probably deep mining or something (I didn't note the coords) and then BAM, the server mem usage shot right off the top of the chart, ran off into swap for a few hundred megs, and the OOM killer nailed it.
15:57 sapier that's fit, some memory corruption happening prior allocating memory
16:02 sapier question is who does it. Happening in send/receive thread is actually quite logic as those do frequent memory allocations.
16:06 hmmmm joined #minetest-dev
16:09 sol_invictus I confirm these crashes, had another one today
16:09 sol_invictus should I provide the log?
16:13 sapier of course, that bug most likely will be hard to track down any piece of information could provide the hint we need
16:15 VanessaE hmmmm: you need to address #1607 right away.
16:15 VanessaE major regression
16:15 VanessaE sol_invictus: I don't suppose you got a gdp backtrace of it? :)
16:15 VanessaE gdb*
16:16 sol_invictus VanessaE: am I supposed to run the server with gdb? xD
16:16 VanessaE sapier: see? ;)
16:17 sapier if you can run your server within gdb and reproduce it that could help yes
16:17 sapier but if you can reproduce it explaining how to reproduce it would be even better ;-) as there are plenty of ways to use gdb which are quite hard to explain
16:19 sol_invictus I wish I knew how to reproduce it
16:19 sapier me too ;-)
16:19 sol_invictus would higher debug level help tracking it down?
16:20 sapier possible but it's still pure luck
16:20 sol_invictus maybe I should compile debug version?
16:20 sapier right now I guess chances someone finds this bug by accident are as good as us looking for it
16:20 sapier yes if you intend to run it in gdb you should do that
16:21 sol_invictus okay, will try that when I'm done with some other stuff
16:23 sol_invictus sapier: what exactly logstream is?
16:23 sapier right now a stream you don't get
16:23 sapier I just changed it to errorstream and pushed
16:23 sapier uncought exceptions are errors for sure
16:26 Sokomine joined #minetest-dev
16:31 sol_invictus ok, I hope next time I will get more detiled output
16:35 sol_invictus q
16:35 sol_invictus left #minetest-dev
16:39 sapier left #minetest-dev
16:43 sol_invictus joined #minetest-dev
16:44 sol_invictus so, anyone can help me on this clearobjects problem?
16:49 VanessaE [09-06 08:41] <sol_invictus> can anyone give me a hint on why /clearobjects causing pos over limit exception when generation limit is lower than 31000?
16:49 VanessaE wat
16:50 VanessaE ok that's just wrong.
16:50 VanessaE that should NEVER happen
16:50 VanessaE wtf is causing the mapgen to create blocks outside the normal limits?
16:59 sol_invictus maybe I explained wrong
16:59 sol_invictus I have set #define MAP_GENERATION_LIMIT (15000)
17:00 sol_invictus tried to run /clearobjects
17:00 sol_invictus got pos over limit
17:00 sol_invictus uncaught exception
17:01 sol_invictus VanessaE: maybe there is a way to run clearobjects on a particular area, not all loadable blocks?
17:01 VanessaE well you can use worldedit to clear selected areas
17:01 VanessaE but otherwise that's a bug :)
17:02 sol_invictus I tried, but it fails when you select a huge area
17:02 sol_invictus even crashes the server iirc
17:02 sol_invictus okay, if it's a bug, I shoulld run it with gdb and reproduce it?
17:03 zat joined #minetest-dev
17:03 sol_invictus because output is not clear in that case either
17:03 Krock sol_invictus, where's that mapgen limit defined?
17:04 sol_invictus constants.h
17:04 Krock kty
17:13 Krock This leads me to the following feature request: https://github.com/minetest/minetest/issues/1608
17:19 sol_invictus yeah, that's a surprise it's still not there
17:20 Krock possibly because it's a define, so minetest-wide
17:20 sol_invictus btw Krock, remote media takes longer than expected, we need to harden our servers first
17:21 sol_invictus there is no conf setting as well
17:22 sol_invictus which is super easy
17:30 sfan5 remote_media takes longer?
17:31 sol_invictus adding remote_media takes longer*
17:36 davexunit joined #minetest-dev
17:37 davexunit hello all.  I'm currently working on packaging minetest for the GNU system distribution.
17:38 davexunit in order to make minetest's modular subgames work well in the system, I have created a patch that adds support for a MINETEST_SUBGAME_PATH environment variable: https://github.com/minetest/minetest/pull/1609
18:02 SoniEx2 left #minetest-dev
18:17 RealBadAngel https://www.youtube.com/watch?v=​znvySoFWum8&amp;feature=youtu.be
18:17 RealBadAngel preview of node highlighting
18:19 davexunit RealBadAngel: neat :)
18:27 f-a joined #minetest-dev
18:27 f-a just downloaded compiled minetest, good job guys
18:36 garfonzo joined #minetest-dev
18:40 sfan5 uh, thanks
18:46 sol_invictus RealBadAngel: that's awesome, but in my opinion it could blink a bit slower
18:46 VanessaE agreed.
18:46 RealBadAngel im trying now to fine tune it
18:47 RealBadAngel it should also be not as bright at night
18:50 Krock RealBadAngel, looks good. I hope it won't depend on shaders
18:51 RealBadAngel Krock, its not shaders feature
18:51 Krock god
18:51 Krock +o
18:51 Krock well, at night it looks to bright
18:52 VanessaE [09-06 14:48] <RealBadAngel> it should also be not as bright at night
18:52 Krock k.
18:52 Krock VanessaE ninja'd me once more again
18:53 sol_invictus someone on forum said me that kaeza was working on head animations, anyone has details on that?
18:54 Krock I've read it on the IRC
18:54 sol_invictus is it ever going to be implemented? (:
18:54 sol_invictus that feels so frustrating, you can't even nod to someone
18:54 * VanessaE pokes Taoki[laptop]
18:54 Taoki[laptop] hi
18:55 VanessaE Taoki[laptop]: I think that ^^^ is your department? :)
18:55 Krock ^^^^
18:55 Taoki[laptop] sol_invictus: I wanted to do head animations too, but there's a problem: Irrlicht breaks models when you try to animate bones directly
18:56 Taoki[laptop] The Lua function in Minetest to allow that already exists. But if you use it, the model will simply break and stop animating and get flipped and stuff. It's an Irr problem, and I'm hoping it will get fixed
18:56 VanessaE sol_invictus: what I find frustrating is simply others not being able to see that you're looking up or down at something
18:56 Taoki[laptop] i don't know if it happens with both b3d and x models. I only tried with x
18:56 VanessaE Taoki[laptop]: have you looked into it recently?  surely by now this has been fixed.
18:56 VanessaE and if it doesn't happen with b3d then ditch that crusty .x format finally
18:56 RealBadAngel i remember seeing animation with player head movement?
18:57 Taoki[laptop] VanessaE: I need to check again against latest irrlicht (1.8). I don't have a b3d exporter howeveer
18:57 VanessaE RealBadAngel: kaeza did it by sending tons of events over the network from the server --> client
18:57 sol_invictus Taoki: ah, I'm using enchanced model with bending limbs and sometimes I see players disassembled :D
18:57 Taoki[laptop] sol_invictus: Yeah, that's the bug
18:57 sol_invictus it doesn't happen often
18:57 VanessaE sol_invictus, Taoki[laptop] I use that knees/elbows model and I don't see that bug.
18:58 VanessaE however wield3d doesn't work right - the item doesn't stay attached to the player's hand
18:58 VanessaE (it floats around, the faster you move, the further away it floats until you stop moving)
18:58 sol_invictus lol
18:58 Krock VanessaE, that's a movement speed fail
18:59 Taoki[laptop] VanessaE: The attachment system also uses Irrlicht's builtin functions for object-to-object attachment. So the problem is again Irrlicht most likely
18:59 Taoki[laptop] I'm not an Irr dev myself, so I can't do much there
18:59 sol_invictus btw, when I got my lag ~.2s I stopped seeing movement lags on carts
18:59 VanessaE Taoki[laptop]: this needs addressed.  I'm sure irrlicht is not the problem here, otherwise it basically couldn't be used with anything.
19:00 Taoki[laptop] possibly
19:00 VanessaE if kaeza could do it over the network from server-->client, then it can be done entirely in-client too
19:01 VanessaE idk how he did it though, probably by modifying the model in realtime
19:02 OldCoder joined #minetest-dev
19:04 sol_invictus I'm wondering, why is irrlicht causing so many problems everywhere?
19:04 sol_invictus was it the only choice?
19:04 Krock yeah. why not making our own 3d graphics processing library?
19:04 * Krock headdesks
19:05 sol_invictus I didn't say that
19:05 sol_invictus just curious if it's the only available
19:05 VanessaE in some cases, irrlicht is simply being underutilized
19:05 VanessaE e.g. minetest code being used to do something irrlicht could do better with just a little bit of preprocessing
19:06 VanessaE in some cases, features not being used at all
19:07 sol_invictus hmm, quite likely
19:09 VanessaE Taoki[laptop]: what splitting the head from the body and attaching it as a separate mesh?
19:10 VanessaE do some magic with the texture file to map it properly, etc.
19:10 Taoki[laptop] VanessaE: Slightly more doable. But then it would also have that lag problem, and overall be more complex than needed
19:10 VanessaE sure, it'd have that movement lag problem, but THAT problem needs addressed too by the sound of it
19:11 VanessaE s/what/what about/
19:13 VanessaE oh well, just trying to spur development a bit :)
19:17 sol_invictus joined #minetest-dev
19:25 sol_invictus screenshot from lua - is it possible?
19:26 Krock afaik no
19:27 Krock use F12 instead
19:28 sol_invictus ik about F12, it not what I want ):
19:28 sol_invictus it's not*
19:29 mrtux joined #minetest-dev
19:29 hintss joined #minetest-dev
19:31 kahrl sol_invictus: lua is only run on the server
19:33 sol_invictus I see. that means there is no way you can get a screenshot from client to server (e.g. /screenshot command) without patching the client, right?
19:33 RealBadAngel kahrl, https://www.youtube.com/watch?v=​znvySoFWum8&amp;feature=youtu.be
19:34 kahrl RealBadAngel: saw it, good work!
19:34 kahrl IMHO it shouldn't blink though
19:34 T4im sol: other engines only can do that usually because the client runs the game code virtualised clientside as well, unless the client can do it nativly
19:35 RealBadAngel kahrl, blinkin makes it work for all kind of textures, bright or not
19:35 kahrl hmm
19:36 sol_invictus T4im: got it
19:37 kahrl RealBadAngel: you still have the selection box so that shouldn't be a big deal
19:38 RealBadAngel im still playing with the effect, trying to make it look even more nice
19:40 sol_invictus I think bliking just needs to be smoother, so it's not that eye-dragging and it would be fine
19:46 sol_invictus joined #minetest-dev
19:54 mrtux joined #minetest-dev
20:03 Megaf joined #minetest-dev
20:06 f-a left #minetest-dev
20:09 Megaf joined #minetest-dev
20:18 sfan5 joined #minetest-dev
20:30 sfan5 joined #minetest-dev
21:00 Krock how about merging https://github.com/minetest/minetest/pull/1605 ?
21:53 paramat joined #minetest-dev
22:06 mrtux joined #minetest-dev
22:10 ImQ009 joined #minetest-dev
23:35 paramat left #minetest-dev

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