Minetest logo

IRC log for #minetest-dev, 2016-04-27

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

All times shown according to UTC.

Time Nick Message
01:04 iangp joined #minetest-dev
01:26 Void7 joined #minetest-dev
02:10 electrodude512 joined #minetest-dev
02:30 STHGOM joined #minetest-dev
02:37 everamzah https://youtu.be/h2TthRVU3Jk?t=1m56s <-- tnt nodes explosion infinite loop
02:40 everamzah game#1059
02:40 ShadowBot https://github.com/minetest​/minetest_game/issues/1059 -- Possible to trigger TNT explosions infinite loop
02:52 gregorycu joined #minetest-dev
02:54 gregorycu !tell Fixer Think I found the bug. It's very very strange.
02:54 ShadowBot gregorycu: O.K.
03:45 hmmmm paramat:  a big reason why I wanted to move generateBiomes to BiomeGen is because that function is manually creating the biomemap
03:45 hmmmm it's a separation of concerns issue that set off an alarm bell in my head making me wonder if what's currently there is proper organization or not
04:13 everamzah i updated game#1059 with dual videos of client machine minetest and server machine minetestserver+htop showing how a client can hang a server machine by igniting too many tnt nodes
04:13 ShadowBot https://github.com/minetest​/minetest_game/issues/1059 -- Possible to trigger TNT explosions infinite loop
04:44 everamzah #4041 squishy textures using john smith TP from https://forum.minetest.net/viewtopic.php?id=5383
04:44 ShadowBot https://github.com/minetest/minetest/issues/4041 -- Squishy textures using john smith TP
04:59 everamzah #4041 updated.  looks like 31c1fca texture overlay scaling breaks at least one texture pack
04:59 ShadowBot https://github.com/minetest/minetest/issues/4041 -- Squishy textures using john smith TP
05:00 VanessaE "squishy"?  heh
05:00 everamzah :D  /me realizes he's been up for 24hours now ---g'night
05:01 VanessaE paramat ^^^
05:10 Hunterz joined #minetest-dev
05:18 jin_xi joined #minetest-dev
05:58 sofar !tell paramat d6c33da3554051b3e326d3b78ef3012f3746fb3a breaks drops :(
05:58 ShadowBot sofar: O.K.
06:06 Zeno` joined #minetest-dev
06:20 Hunterz joined #minetest-dev
06:22 hcorion joined #minetest-dev
06:35 MoNTE48 joined #minetest-dev
06:35 nore sofar: isn't it that explosions destroy drops of previous ones? If so, isn't it actually a feature? ;)
06:36 sofar unintentional, it removes ALL of them
06:36 sofar but yeah, I could see that
06:37 sofar all this setting of 20+ tnt is just nonsense, I want to see if I can make it more robust somehow
06:37 sofar trying out my vmanip combine right now
06:37 sofar that'll make it one huge boom instead of ~25 small ones
06:38 nore oh, good idea
06:57 Icedream joined #minetest-dev
07:04 sofar you know, it SUCKS I have to use radius as a measure of TNT strength
07:04 sofar I now have to do oddball calculations to recalculate the new radius
07:05 sofar because it's DUMB to use radius as a measure of TNT power
07:05 sofar just... DUMB
07:07 * Zeno` removes TNT
07:07 Zeno` it's causing too much stress
07:07 sofar lol
07:07 sofar where? haha
07:08 MoNTE48 joined #minetest-dev
07:10 Zeno` adding code to check for TNT-like mods and automatically disable and delete them
07:14 sofar whoahhh I got it to work
07:23 Zeno` got it to work? Sounds like a bug
07:28 sofar what's causing your stress though? too many entities?
07:28 sofar https://youtu.be/Aj53S8dKjl4
07:28 Zeno` I was referring to your stress lol
07:28 sofar ah, sure :)
08:02 nrzkt joined #minetest-dev
08:02 Megal joined #minetest-dev
08:46 loggingbot_ joined #minetest-dev
08:46 Topic for #minetest-dev is now Minetest core development and maintenance. Feature freeze (2 weeks): sometime later than 2016-04-27. Last release: 0.4.13, Aug 20 2015. Chit-chat goes to #minetest. Consider this instead of /msg celeron55. http://irc.minetest.net/minetest-dev/ http://dev.minetest.net/
09:24 nrzkt joined #minetest-dev
09:48 Fixer joined #minetest-dev
10:04 Zeno` stupid Windows
10:04 Zeno` oops, wrong channel (it's becoming a habit lol)
10:51 gregorycu joined #minetest-dev
10:55 gregorycu I don't want to say it, but the milestone looks so close
10:56 VanessaE shah!  mach nicht.
10:57 Zeno` we can always move the milestones; it's not as if a "mile" is a valid unit of measure these days
10:58 VanessaE so should we call them kilometerstones?
10:58 Zeno` yes
10:58 Zeno` :D
11:40 gregorycu I could reproduce Fixer's issue
11:40 gregorycu With a while loop
11:42 Taoki[mobile] joined #minetest-dev
11:44 gregorycu Looks like it was visual studio that caused it
11:46 proller joined #minetest-dev
11:55 Fixer hi gregorycu
11:55 Fixer gregorycu, glad you did it
11:55 gregorycu Hello there
11:55 gregorycu Tell me Fixer, do you have some sort of shell plugin
11:55 gregorycu Like git for windows
11:56 Fixer shell plugin? i have md5 in properties one
11:56 Fixer 7-zip menu
11:57 Fixer antivirus entry
11:57 gregorycu You have AV ey?
11:57 gregorycu I could reproduce with visual studio querying the file in Debug build
11:57 Fixer avast crap
11:57 gregorycu I'm going to attempt with release without visual studio running
11:57 gregorycu To see if windows alone can cause problems
11:59 Fixer it happens randomly, can happen right away or after a day
12:00 gregorycu I had it in a tight loop
12:00 gregorycu Basically it would keep trying until the issue happened :)
12:00 gregorycu My poor disk drive
12:03 Fixer SSD?
12:06 gregorycu Are you able to do me a favour :)
12:06 gregorycu can you run proc mon, and just see who tries to look at that file
12:07 gregorycu You don't need to reproduce the issue
12:07 gregorycu Just want to see what exe files touch it
12:07 gregorycu Leave your AV running in the same state it was when the issue happened
12:09 gregorycu lol
12:09 gregorycu I could reproduce by clicking the file in explorer
12:09 gregorycu That caused the file delete to be delayed
12:09 gregorycu Man... what a bug
12:16 Zeno` that's a bit weird as well
12:17 Zeno` if(rename(tmp_file.c_str(), path.c_str())) {  is called twice when connecting to server
12:17 Fixer gregorycu, ok
12:17 gregorycu Only if you have time Fixer
12:17 Zeno` from insert and also delete
12:17 gregorycu What do you do for a living?
12:17 gregorycu @Fixer
12:17 Fixer unemployed, thats why I have time for now
12:18 gregorycu Make sure you stay unemployed
12:18 Fixer no, lol
12:18 gregorycu What skills do you posses?
12:18 Fixer computers (medium), accounting
12:19 Zeno` doesn't fail on Linux though (if I open the file manually before exiting the server)
12:19 Zeno` i.e. rename still succeeds
12:19 gregorycu Yeah, file gets unlinked
12:19 Zeno` gregorycu, can you unlink/stat on Windows?
12:19 Zeno` or is that impossible?
12:19 proller joined #minetest-dev
12:19 gregorycu I believe it's impossible
12:19 Zeno` bummer
12:20 gregorycu Though there is an _unlink function
12:22 gregorycu I'll find out what it does under the surface
12:23 Zeno` It's probably a blocker as well
12:23 Zeno` I wonder why it's never shown up before
12:24 Zeno` safeWriteToFile() does seem to be called excessively but even if that's a new thing the bug must have occurred *sometimes* before recent changes
12:25 gregorycu Unless it's called more execessively than usual
12:26 Zeno` yeah... I'm tracing now to see what's going on. BUT the recent changes to the main menu should not have changed that because from what I can see at the moment it's all in C++
12:28 gregorycu unlink does the same thing
12:30 Zeno` a riddle
12:33 Zeno` that's odd. If a server connect fails it's removed from the fav list
12:36 PilzAdam joined #minetest-dev
12:38 STHGOM joined #minetest-dev
12:44 Zeno` odd way to do things: http://dpaste.com/1MJFY8V.txt
12:45 Zeno` (to remove duplicates apparently)
12:45 gregorycu Yeah
12:45 gregorycu But it "works"
12:45 Zeno` yeah... odd, but I doubt it's related
12:46 gregorycu The more writes, the more chance of problems
12:46 gregorycu Actually, a file write probably triggers anti-virus
12:46 gregorycu Then the second write comes along
12:47 gregorycu Then bam!
12:48 Zeno` yes, of course
12:48 Zeno` But this stuff hasn't changed so I wonder why it's taken so long to show up :/
12:49 Zeno` infuriating
12:55 Zeno` maybe it's been there for years and just nobody noticed
12:55 Zeno` because they were on different tabs
12:59 Zeno` the delete first could be simply avoided of course (i.e. check if server is already in the local list and not insert it again if it is)
13:00 Zeno` for multiple duplicates deleteEntry() could be called at some other time as part of some fancy maintenance regime
13:01 Zeno` just needs another condition in insert()
13:01 Zeno` seems like the more sane way to do things
13:03 Player_2 joined #minetest-dev
13:06 Zeno` dunno about "managed_to_rename" as a variable name, but looks good
13:07 Zeno` your #else condition is wrong
13:07 Zeno` >>>> managedToRename = rename(tmp_file.c_str(), path.c_str()) == 0;
13:07 Zeno` gregorycu ^
13:07 gregorycu Yeah
13:07 gregorycu Fixing now :)
13:07 gregorycu What's a better name?
13:08 gregorycu rename_success ?
13:10 Zeno` yeah I like that much better
13:11 Zeno` I think the else { return true can be removed as well
13:12 Zeno` I don't like functions dropping off the end without actually returning
13:12 Zeno` i.e. just have return true on line 719
13:13 Zeno` will ReplaceFile() delete the temporary file?
13:13 gregorycu Yes
13:13 Zeno` nifty
13:14 gregorycu I've fixed it up
13:14 gregorycu bb in 30 min
13:16 damiel joined #minetest-dev
13:21 Fixer LOL
13:22 Fixer foobar2000 accessing favourites.txt?
13:22 Zeno` wtf is foobar2000?
13:22 Zeno` oh I see
13:22 Fixer music player
13:22 Zeno` really?
13:23 * Zeno` looks behind him. Hmm... no demons there
13:23 Zeno` This is why music should be banned
13:24 Fixer https://i.imgur.com/54mzjNT.png
13:25 Fixer Zeno`, player is monitoring that folder with minetest for music ^_>
13:25 gregorycu lol totalcommander
13:26 gregorycu You must be european
13:26 Fixer ofc
13:26 Fixer it rules
13:26 gregorycu Please
13:26 gregorycu The only person I know using total commander is serbian
13:26 gregorycu Which says A LOT about TC
13:26 Fixer in my country it is everywhere since ages
13:27 gregorycu Anyway, foobar fucked your game lol
13:27 Fixer yeah, lol
13:27 gregorycu I wonder wtf it is doing
13:27 Fixer searches for music and adds to library
13:27 Fixer good player, btw
13:28 Fixer gregorycu, and total commander is made by guy from Switzerland
13:29 Fixer in my country it is "Dos Navigator -> Far Manager -> Total Commander" history of filemanagers
13:29 Zeno` a music player
13:29 Zeno` *sigh*
13:29 Zeno` Add that to the PR comments please
13:30 Fixer me?
13:30 Zeno` Fixes conflict with foobar2000 lol
13:30 Zeno` nah, I'm not serious
13:31 Zeno` you might add it to the bug report though
13:31 Zeno` it's good
13:34 Zeno` At least it's appropriately named (fubar)
13:37 iangp Can someone check if I found a bug?
13:37 iangp When you change the font using the advanced settings, this mess up everything on GUI... only restarting minetest fixed the problem...
13:38 iangp Default font to fonts-japanese-gothic.ttf
13:41 Fixer o_0
13:42 kaadmy joined #minetest-dev
13:44 iangp Fixer: Is that reaction for me? Should I open a issue?
13:46 Fixer i guess so
13:46 Fixer please open
13:50 iangp Ok, thanks I'll
14:01 davisonio joined #minetest-dev
14:03 gregorycu_ joined #minetest-dev
14:13 Hunterz joined #minetest-dev
14:38 est31 joined #minetest-dev
14:40 gregorycu Debug is so damn slow
14:42 gregorycu Collision detection
14:42 gregorycu I thought someone fixed that
14:53 nore sofar: for game#1060, I like the tnt combine solution better, since it allows tnt to destroy already-placed items
14:53 ShadowBot https://github.com/minetest​/minetest_game/issues/1060 -- TNT: Fix multiple explosions erasing drops by sofar
14:54 jin_xi joined #minetest-dev
15:02 Void7 joined #minetest-dev
15:10 hmmmm joined #minetest-dev
15:17 gregorycu Um...
15:18 gregorycu minetest.find_path appears to be broken...
15:20 kaadmy gregorycu: i think that function has been replaced with a faster one in latest git
15:20 kaadmy not sure though
15:20 hmmmmm joined #minetest-dev
15:20 gregorycu That's the one I'm using
15:32 gregorycu And by broken, I mean suboptimal path
15:37 Zeno` And the previous version gave an optimal path?
15:37 gregorycu I'm not sure, will check
15:38 Zeno` hmmmmm, since you had so much input into the original conversation... #4038
15:38 ShadowBot https://github.com/minetest/minetest/issues/4038 -- Handle particle spawners in env and delete expired ids by Zeno-
15:38 gregorycu It's severely unoptimal
15:41 gregorycu wtf lua
15:41 gregorycu ~=  is not equals
15:41 ShadowBot gregorycu: Error: You must be registered to use this command. If you are already registered, you must either identify (using the identify command) or add a hostmask matching your current hostmask (using the "hostmask add" command).
15:41 gregorycu My bad bro
15:41 Zeno` hmmmmm (if you look at the patch you'll see that it's not really my PR)
15:42 hmmmmm heh
15:42 hmmmmm why -1024?
15:42 Zeno` it seemed like a nice number
15:42 Taoki[mobile] joined #minetest-dev
15:42 Zeno` there is no reason
15:42 hmmmmm fair enough
15:43 hmmmmm i have a request
15:43 Zeno` -2 would have done
15:43 Zeno` yes, request away
15:43 jin_xi i think 1338 is used for similar purposes elsewhere
15:43 hmmmmm i don't know if this is too nitpicky or what
15:43 hmmmmm i would prefer it if addParticleSpawner took a bool for 'expires'
15:43 Zeno` I just want to get this thing out of the way so that it's no longer me looking after it :)
15:44 Zeno` hmm... /me looks
15:44 hmmmmm u32 ServerEnvironment::addParticleSpawner(float exptime, bool is_permanent) or something
15:44 Amaz joined #minetest-dev
15:44 hmmmmm m_env->addParticleSpawner(blah, blah > 0.f);  ?
15:44 hmmmmm I don't know
15:44 hmmmmm what do you think about it
15:47 Zeno` Seems reasonable but then the check really just gets moved to server.cpp
15:48 hmmmmm the check just gets moved, but the design of the particle spawner interface is improved
15:50 hmmmmm meh nevermind, i'm being too nitpicky
15:50 Zeno` nah I might change it
15:50 Zeno` I was trying to limit *my* changes but what you do does make sense
15:51 Zeno` the function is only called from that once place anyway
15:51 hmmmmm right, it's only being called from one place
15:51 hmmmmm meaning it's not much of an interface to begin with
15:52 Zeno` yeah :)
15:52 hmmmmm it's just another chunk of code that does *thing*
15:52 Zeno` can we leave it for now?
15:52 hmmmmm yes
15:52 Zeno` I do agree with you that it would be better but I'd just rather leave the commit as close to original as possible. Thanks
15:55 Zeno` I'll add a comment to -1024
15:57 gregorycu Is there an event which triggers when a node becomes loaded?
15:57 gregorycu In lua?
15:57 hmmmmm kinda
15:57 hmmmmm do you mean a node of a specific type or at a specific position
15:58 gregorycu node of a specific type
15:58 hmmmmm nope, not at all
15:58 gregorycu When i place a node I set a repeatable timer
15:58 gregorycu Which gets broken when I reload the map
15:59 hmmmmm you gotta save the node positions
15:59 gregorycu In a file manually?
16:01 hmmmmm yup
16:02 Zeno` will push 4038 shortly...
16:04 Zeno` hmmmmm, you need to add a +1 to the PR :P
16:04 hmmmmm do you really
16:05 Zeno` I have no idea... the system confuses me
16:05 hmmmmm there is no system
16:05 Zeno` but you're ok with me merging it?
16:05 hmmmmm yes
16:05 hmmmmm go already
16:05 Zeno` lol, ok, ok, will do so shortly :)
16:07 Zeno` I can't remember how
16:07 Zeno` justkidding!
16:10 gregorycu http://www.gregorycurrie.com/fark.jpg
16:10 gregorycu That's the optimised pathfinder
16:11 Zeno` I like those dots
16:11 gregorycu Lucky for you, cause there is so many of them
16:12 Zeno` merged the particle thing
16:13 gregorycu The good news is that the optimised pathfind is optimised
16:13 gregorycu In terms of speed
16:14 Zeno` so the path is "go wherever you like as long as it's in the general direction?"
16:14 gregorycu Even then, the screenie shows the pathfinder striping
16:14 gregorycu It's not clear from the dots, but it's going up and back, up and back
16:15 gregorycu Like the game snake
16:15 Zeno` I like that game
16:16 est31 gregorycu, i didnt say the optimized pathfinder is good
16:16 est31 i just said it was better than before
16:16 gregorycu No, it's much faster
16:16 gregorycu It's nice
16:16 gregorycu But isn't sofar using this for his sheep
16:17 PilzAdam snake sheep
16:17 est31 :)
16:19 Zeno` sheep?!
16:20 est31 Zeno`, yes lots of them https://www.youtube.com/watch?v=avkzrF7J96U
16:20 est31 https://www.youtube.com/watch?v=P7kACeC0O5g
16:21 Zeno` lol
16:21 Zeno` nice
16:22 Zeno` that reminds me of the thing my mum installed on Windows XP once, with sheep running around all over the desktop
16:22 gregorycu Yeah, very cool
16:29 gregorycu est31: Does the pathfinder have any persistent state?
16:32 est31 no
16:33 gregorycu I'm seeing how long 10k pathfinding calls take
16:33 gregorycu About 10 seconds
16:33 est31 every time you ask the pathfinder method you get a new pathfinder
16:33 gregorycu Sweet
16:33 est31 that's why it doesnt make much sense to expose the pathfinder class via the header
16:34 gregorycu I'm not :)
16:34 est31 gregorycu, my PR removes the pathfinder class from the header ant puts it into the cpp
16:34 gregorycu sorry
16:34 gregorycu lol
16:34 gregorycu I read expose as explore
16:35 gregorycu It's late
16:35 est31 np
16:35 est31 gregorycu, try pathfinding 10k without my patch, you will get an std::bad_alloc or something like that error
16:35 est31 or perhaps not
16:35 gregorycu Yeah, once was bad enough
16:35 est31 it will just loop for almost eternity
16:35 Void7 joined #minetest-dev
16:35 est31 trying to allocate a 10k * 10k * 10k block
16:37 est31 and its a very dumb optimisation
16:37 est31 just not allocating those nodes you dont read at all
17:09 kinokoio joined #minetest-dev
17:15 johnnyjoy joined #minetest-dev
17:17 Void7 joined #minetest-dev
17:25 Darcidride joined #minetest-dev
17:27 Krock joined #minetest-dev
17:28 hmmmmm oh boy
17:29 hmmmmm valleys is manipulating noise on its own
17:29 kinokoio joined #minetest-dev
17:29 Zeno` well, echoes are common in valleys
17:30 hmmmmm i think i might have left the mapgen go for too long
17:30 hmmmmm i have to make the same change in 12 different files and then not know what to do when somebody slightly modified it and didn't understand the abstractions the original code used
17:31 hmmmmm this is my fault
17:31 hmmmmm if i fixed things up to keep them tidy from time to time, there wouldn't be any need to start hard coding things like this
17:33 Zeno` It's all Lua's fault
17:38 sofar is greg looking into fixing the pathfinder code?
17:38 sofar I've seen paths like this (but only occasionalyy): http://www.gregorycurrie.com/fark.jpg
17:39 sofar most of the time it's ok (just look at my sheep videos
17:58 Krock joined #minetest-dev
18:07 est31 merging in 20 minutes https://github.com/est31/minetest.github.io/co​mmit/5566bb8d6cb7aceb3f5562230434da3450190b7a
18:07 est31 hmmmmm, ?
19:01 Megal joined #minetest-dev
19:05 proller joined #minetest-dev
19:19 hmmmmm yeah that's fine and all but...
19:19 hmmmmm is there any way to automate this so it doesn't have to be changed every time there's a version number increase?
19:32 est31 idk, there is a script in util/ to increase the version
19:32 est31 also there is a wiki page
19:32 est31 http://dev.minetest.net/Releasing_Minetest
19:33 proller joined #minetest-dev
20:06 proller joined #minetest-dev
20:19 troller joined #minetest-dev
20:33 proller joined #minetest-dev
21:25 turtleman joined #minetest-dev
21:52 DFeniks joined #minetest-dev
22:02 damiel joined #minetest-dev
22:24 proller joined #minetest-dev
22:25 proller joined #minetest-dev
22:47 troller joined #minetest-dev
22:50 proller__ joined #minetest-dev
22:54 rubenwardy joined #minetest-dev
23:33 Fixer !tell paramat this one is hanging with two approvals for some reason: https://github.com/minetest/minetest/pull/3977
23:33 ShadowBot Fixer: O.K.
23:33 yang2003 joined #minetest-dev
23:55 crazyR joined #minetest-dev

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