Minetest logo

IRC log for #minetest-dev, 2016-03-12

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

All times shown according to UTC.

Time Nick Message
00:11 paramat ok i'll merge #3843 soon, any objections to me +1 and also merging #3503 ?
00:11 ShadowBot https://github.com/minetest/minetest/issues/3843 -- Documentation: Auto-update conf.example and settings_translation_file.cpp by paramat
00:11 ShadowBot https://github.com/minetest/minetest/issues/3503 -- Nodebox: Allow nodeboxes to "connect" by sofar
00:16 paramat will merge one or both PRs in an hour
00:20 est31 paramat, there is still open discussion on #3503
00:20 ShadowBot https://github.com/minetest/minetest/issues/3503 -- Nodebox: Allow nodeboxes to "connect" by sofar
00:42 sofar I replied to shadowninja, I think the warning is sufficient. maybe improve the warning a bit?
00:43 est31 wouldnt it be less complex documentation wise and code wise to simply not do warnings, and only accept the written out values
00:47 sofar you know, I could just change the whole thing to top/bottom/front/back/left/right at this point
00:47 sofar I just want it in
00:47 sofar here I was thinking to be nice and display a warning and be a bit flexible and make checking cheap by just looking at the first letters :(
00:50 Fixer do that serialisation errors still appear on the screen in 0.4.13-dev? because I see them from time to time when playing on some older servers
00:51 sofar I'm getting really tired of repushing changes and digging up more discussion, though
00:51 sofar so I'll do it permitted that I actually get an approval
00:51 est31 discuss that with ShadowNinja :)
00:51 sofar ShadowNinja: speak up ;)
00:52 sofar est31: would you +1 if top/bottom/front/left etc. is used throughout?
00:52 est31 yes
00:52 sofar ShadowNinja: you?
00:53 sofar (I also have to find out where canonically is shown what side front is, lol
00:53 sofar is front E? +X ?
00:54 ShadowNinja sofar: Yep.
00:54 sofar ok, then I will resubmit everyone tonight
00:55 ShadowNinja sofar: Front depends on facedir, but would default to +Z I suppose, whatever side the front texture is put on on nodes without facedir.
00:57 Void7 joined #minetest-dev
00:59 sofar right, with facedir=0
01:01 paramat ok
01:02 paramat yes i prefer that nesw was dropped
01:02 cat5e joined #minetest-dev
01:03 paramat at this point it's more useful to have this merged, used and tested
01:10 sofar good, that's a plan then
01:12 paramat i removed milestone from #3759
01:12 ShadowBot https://github.com/minetest/minetest/issues/3759 -- Remove 'new style water' and waving water.
01:16 sofar ShadowNinja: confusing part is that the "back" side of the chest is placed with +Z if facedir=0, so the "back" side of the chest is facing +Z :)
01:16 Void7 joined #minetest-dev
01:17 ShadowNinja sofar: O.K., so make back = north.
01:17 sofar before I code it, I'll have to check if there's other uses to make sure it's consistent
01:18 sofar map.cpp:v3s16(0,0,-1), // front
01:18 sofar hmm
01:18 sofar ok, so front is -Z
01:19 sofar alrighty then
01:20 sofar I'll also change order of the horizontals to facedir = 0, 1, 2, 3 order (S, W, N, E)
01:20 sofar dis gonna be hilarious
01:27 * sofar sed's
01:38 est31 so front is -Z always?
01:38 est31 not depending on the facedir
01:38 est31 ?
01:38 est31 i have no idea tbh
01:38 est31 about all that facedir magic
01:39 sofar front is Z-- with facedir = 0
01:40 sofar IOW, with facedir=0, the face of the node that is FACING south is the front
01:46 ssieb joined #minetest-dev
02:02 Void7 joined #minetest-dev
02:17 sofar c++ part done, now testing with my mt_game patches
02:17 sofar I'll have to redo the rot[] table
02:17 sofar dinner time, 3 hungry kids and cranky SO
02:18 sofar afk ;)
02:31 paramat left #minetest-dev
02:31 lubuntu_ joined #minetest-dev
02:33 ShadowNinja Chat console fix (#3831): http://sprunge.us/HNWB?diff
02:33 ShadowNinja "Esc stops working when in F10 mode and form invoked, and F10 pressed again" -- https://github.com/minetest/minetest/issues/3831
02:44 yang2003 joined #minetest-dev
02:57 Miner_48er joined #minetest-dev
03:00 everamzah game#909
03:00 ShadowBot https://github.com/minetest/minetest_game/issues/909 -- Doors that were removed by being burned (fire mod) are leaving hidden nodes.
03:07 sofar everamzah: mention my name in that bug plz
03:07 sofar I have a fix for that already
03:09 everamzah sure, you mean like, from now on?
03:09 sofar no, just in that bug
03:09 everamzah oh, alrighty
03:24 est31 joined #minetest-dev
03:56 paramat joined #minetest-dev
04:01 paramat now merging #3843
04:01 ShadowBot https://github.com/minetest/minetest/issues/3843 -- Documentation: Auto-update conf.example and settings_translation_file.cpp by paramat
04:10 paramat merged
04:49 paramat left #minetest-dev
05:38 sofar OK, I'm all done redoing #3503 and all 3 dependent game PR's
05:38 ShadowBot https://github.com/minetest/minetest/issues/3503 -- Nodebox: Allow nodeboxes to "connect" by sofar
05:38 sofar it does make it nicely consistent now
05:38 sofar ShadowNinja: ^^
05:49 Hunterz joined #minetest-dev
06:50 ssieb on_after_place is called even if the place didn't succeed?  why?
07:49 Krock joined #minetest-dev
07:55 ssieb sofar: it would help so much if there was engine support for multi-block nodes
07:56 nrzkt joined #minetest-dev
07:57 Amaz joined #minetest-dev
08:00 srifqi joined #minetest-dev
08:10 srifqi joined #minetest-dev
09:14 Obani joined #minetest-dev
09:26 ud1_ joined #minetest-dev
09:30 Hunterz joined #minetest-dev
09:47 DevBox joined #minetest-dev
09:53 blaze joined #minetest-dev
10:00 VargaD_ joined #minetest-dev
10:19 Calinou joined #minetest-dev
10:26 linkedinyou joined #minetest-dev
10:30 Guest68637 joined #minetest-dev
11:40 Fixer joined #minetest-dev
11:44 edgrey joined #minetest-dev
12:40 Sockbat joined #minetest-dev
12:42 numZero joined #minetest-dev
12:53 numZero joined #minetest-dev
13:05 Fixer joined #minetest-dev
13:08 turtleman joined #minetest-dev
13:26 proller joined #minetest-dev
13:52 Megaf joined #minetest-dev
14:18 Player_2 joined #minetest-dev
14:26 rubenwardy joined #minetest-dev
14:31 Dragonop joined #minetest-dev
14:52 SaKeLWorld joined #minetest-dev
14:58 kaadmy joined #minetest-dev
15:30 blaze left #minetest-dev
15:30 sofar_ joined #minetest-dev
15:30 Etzos joined #minetest-dev
15:34 fling joined #minetest-dev
15:35 thatgraemeguy joined #minetest-dev
15:35 thatgraemeguy joined #minetest-dev
15:40 Void7 joined #minetest-dev
15:43 srifqi joined #minetest-dev
15:43 ssieb joined #minetest-dev
15:48 rubenwardy joined #minetest-dev
15:54 blaze joined #minetest-dev
15:56 rubenwardy needs one approval label: https://github.com/minetest/minetest/pull/3826
16:05 hmmmm joined #minetest-dev
16:11 Robby_ joined #minetest-dev
16:11 misprint_ joined #minetest-dev
16:12 Calinou joined #minetest-dev
16:13 book` joined #minetest-dev
16:21 Obani joined #minetest-dev
16:52 ShadowNinja sofar_: https://github.com/sofar/minetest/commit/584e4896bab7f3b9b1be8838cff0a7461830aa21#diff-fc5bd7f599797bce163af517e57a314cR561  Why not just side == "top"?   https://github.com/sofar/minetest/commit/584e4896bab7f3b9b1be8838cff0a7461830aa21#diff-1305560bd8befb32862f0feeefabd02eR3481 Missing comment closer.  --  These can be fixed when merging though.
16:54 sofar_ ah
16:54 sofar joined #minetest-dev
16:55 sofar easy to fix.
16:56 ShadowNinja sofar: I'll fix it locally.
16:56 sofar are you merging?
16:56 ShadowNinja sofar: Yep.  #3503 for anyone who wants to comment.
16:56 ShadowBot https://github.com/minetest/minetest/issues/3503 -- Nodebox: Allow nodeboxes to "connect" by sofar
16:57 ShadowNinja I'll wait a bit of course.
16:57 sofar hell, go ahead :)
16:57 ShadowNinja I'll also push the console fix now.
17:04 Obani ShadowNinja, https://github.com/minetest/minetest/pull/1689
17:05 Obani also https://github.com/minetest/minetest/pull/817
17:06 ShadowNinja sofar: Hmmm, it seems like i is also unused in that loop, and the one above...
17:07 sofar lol
17:07 sofar caught red handed copying
17:07 sofar ShadowNinja: yah, nuke it
17:07 Obani joined #minetest-dev
17:07 sofar I wonder why clang didn't spot that
17:08 sofar ah, we increment it
17:08 sofar nuke it
17:25 rubenwardy joined #minetest-dev
17:32 Megaf joined #minetest-dev
17:33 ShadowNinja sofar: Doesn't seem to be working with Technic cables...  Also, does it default to connect_sides = all?
17:37 ShadowNinja Oh, maybe I chould set type=connected...
17:58 ShadowNinja Having only the fixed nodebox for the selection box makes it harder to place wires.
17:58 ShadowNinja Not really a reason to block though.
18:01 sofar there's no default for connect_sides
18:02 sofar we could set it to 63 (all)
18:03 ShadowNinja It seems like the default is all in practice.
18:03 sofar it is?
18:04 sofar reset() puts it to 0
18:05 sofar I should try and modify some of the technic cabling to see if it could work
18:07 sofar ShadowNinja: connect_sides defaulting to all doesn't make sense, by default nothing will connect to random non-connected nodes
18:08 sofar remember, connect_sides is not a property of connected nodeboxes, but of any nodebox
18:08 sofar any node, even
18:09 ShadowNinja sofar: Yes, but if I ommit "connect_sides" in the definition for eg, a solar array, it will connect on all sides.
18:09 ShadowNinja That's what it should do though.
18:09 sofar ahh yes
18:09 sofar but now add connect_sides
18:10 sofar yeah technic cables should convert nicely
18:15 ShadowNinja sofar: Can you make it work with selection_box?  This makes cables a lot harder to place otherwise.
18:15 ShadowNinja Pushed.
18:16 sofar worth the fix, right?
18:16 sofar I'll do the technic cable conversion too so I have a good test bed
18:16 sofar can you pastebot your technic diff?
18:16 sofar so I don't need to redo the lot
18:17 ShadowNinja sofar: Sure, one sec...
18:18 ShadowNinja sofar: http://ix.io/rq4  --  There's some onrelated cleanup here, but it's small and there's no harm in applying it.
18:19 sofar I think you're patching an older technic
18:19 sofar is is there a dev branch?
18:19 sofar it's fine, I can work with this
18:25 sofar lol, crash when connecting switching station
18:25 sofar it's gonna take some work :)
18:25 ShadowNinja sofar: Actually, you'll have to remove that cleanup stuff -- half of it isn't in that diff...  Moment and I'll get you a working one.
18:27 ShadowNinja sofar: Try this: http://ix.io/rq5
18:28 sofar that's clean
18:28 sofar it's nice using the groups, right? :)
18:29 sofar don't you have to add the switching station to each cable's connects_to ?
18:29 sofar I'll look
18:30 sofar oh, you just put it in each tier
18:30 sofar solar array doesn't connect yet... I'm sure there are more
18:31 sofar heh it works, just doesn't visually connect
18:31 ShadowNinja Actually, a better design would be having it connect directly to the switching station, or an 'all tiers' group, that way you could add groups after the switching station is registered.
18:32 ShadowNinja Solar array should work.  I only connects on the bottom though.
18:34 sofar the arrayed ones work
18:34 sofar not the "small" one
18:38 sofar got it
19:08 damiel joined #minetest-dev
19:23 Soni joined #minetest-dev
19:31 turtleman joined #minetest-dev
19:34 ShadowNinja Anyone have any idea why this would happen on OSX? https://github.com/minetest/minetest/pull/3824#issuecomment-193836619
19:35 ShadowNinja It fails randomly and then works.
19:36 ShadowNinja If we can't figure this out we'll just have to disable that test on OSX and hope nothing breaks when you use it.
19:47 Void7 joined #minetest-dev
19:47 kahrl ShadowNinja: maybe a race condition on Thread::m_thread_id?
19:47 kahrl look at https://github.com/ShadowNinja/minetest/blob/1fae76b6ceca01f274720fe31a0fbb98ce02fe5e/src/threading/thread.cpp#L134
19:48 kahrl perhaps the thread runs (after being started by pthread_create) before m_thread_id gets set by the parent thread
19:50 ShadowNinja kahrl: Yes, that might be it.  I'll add a test.
19:55 ShadowNinja If this is the issue, I'll have to rework Thread a bit (C++11 threads have to call the threads get_id instead of using a value set by the thread creator).
20:06 ShadowNinja kahrl: Seems that's it.
20:11 kahrl ShadowNinja: good to know the cause :)
20:13 kahrl ShadowNinja: well, the "obvious" solution would be to add a semaphore which the threadproc waits on and which the parent process posts (after setting everything up)
20:13 kahrl ShadowNinja: but if there is a simpler solution go for it
20:13 ShadowNinja kahrl: No need.
20:13 ShadowNinja See my updated PR.
20:14 kahrl ah, that should work too
20:18 kahrl hmm
20:18 kahrl does POSIX guarantee that pthread_create writes to its first argument before starting the thread?
20:18 kahrl can't find anything here: http://pubs.opengroup.org/onlinepubs/9699919799/functions/pthread_create.html#
20:21 Void7 joined #minetest-dev
20:21 kaeza joined #minetest-dev
20:21 sofar https://youtu.be/blrurijucN4
20:23 sofar can we convert that to core? It's just bone manipulation
20:23 sofar and it looks FREAKING AWESOME
20:26 sofar the body movement isn't perfect yet, but I think we can nail this out with rui
20:31 rubenwardy joined #minetest-dev
20:32 celeron55 if a mod can do it, then the only challenge is the interfacing it has to do with mod-provided stuff
20:33 celeron55 it should be pointed out though that including that in core might make no sense
20:33 sofar it's laggy through mods
20:33 sofar very, and lots of packets
20:33 sofar client-side would be highly optimized
20:33 sofar I see your point thought
20:33 sofar what if you could disable it?
20:33 celeron55 optimization is one valid reason to do it, yes
20:34 celeron55 but in theory it shouldn't improve lagginess
20:34 sofar I mean, look at the video, it's a localhost server and it'c choppy
20:35 sofar player:head_animation(bool enabled = true) ?
20:35 sofar idk, something like that
20:38 celeron55 that might indeed fit with the other interfaces
20:39 sofar that would make it per-player
20:39 sofar not a bad idea
20:40 sofar a mod could take over a player and make it's head sping
20:40 sofar spin*
20:40 sofar now there's an idea to scare my kids
20:43 Megaf joined #minetest-dev
20:56 Sokomine moving heads would certainly be nice :-) looking straight up used to create a headache by the clouds whirling around. that seems to have been fixed
21:14 ShadowNinja kahrl: I suppose it isn't guaranteed to be set first.  I doubt that's a concern in practical implementations though.
21:15 ShadowNinja If it is a coucern we could call pthread_self in threadProc.
21:28 anonmod joined #minetest-dev
21:29 anonmod the get_[heat|humidity] documentation pages say "The weather functions were dropped at Apr 18, 2014."
21:29 anonmod is there a proper way to sample the biome data in abm's?
21:30 kahrl ShadowNinja: I could imagine a pthread implementation where pthread_t is simply a process id, and pthread_create calls clone(), then stores the return value of clone into its first argument
21:31 kahrl ShadowNinja: so the process created by clone could execute before pthread_create has a chance to write to its first argument
22:06 Megaf joined #minetest-dev
22:11 ShadowNinja Merging #3826.
22:11 ShadowBot https://github.com/minetest/minetest/issues/3826 -- Allow selecting screenshot format. by kaeza
22:23 damiel joined #minetest-dev
22:31 est31 joined #minetest-dev
22:32 est31 nore, you spoke in favor of #3713 a few days ago. do you +1 it?
22:32 ShadowBot https://github.com/minetest/minetest/issues/3713 -- Resend blocks when modified while sending to client by Jeija
22:37 nore est31: it looks good, but I didn't test it
22:37 nore but since several people tested it and even said it fixed the meta setting, then +1
22:54 est31 fine, thanks!
23:10 sofar nice
23:15 Fixer_ joined #minetest-dev
23:22 Fixer_ tested pr3811 for some time, seems fine for me
23:37 yang2003 joined #minetest-dev
23:38 rubenwardy joined #minetest-dev
23:54 Megaf est31, would be cool having that merged I think
23:54 Megaf if theres no drawbacks
23:55 Megaf oh, was it merged already?
23:56 kahrl 3713 was merged, yes
23:56 Megaf partly
23:56 kahrl right
23:57 kahrl the second commit is just code style though, it doesn't affect functionality at all

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