Time Nick Message 11:17 nrzkt ShadowNinja: i haven't time to finish it, but i'll test it this week-end i think. 11:18 nrzkt i will create the dev-0.5 branch for the breaking changes (because network is the only change, authenticiation, and maybe more), agree ? celeron55 told me to create network-rework, but i think dev-0.5 is better 11:19 celeron55 dev-0.5 is ok 11:22 nrzkt ok, i create it and merge PR later 11:28 nrzkt ShadowNinja: fixed #2376 for CONTENT_AIR under CONTENT_AIR, now i think it's good, can you review ? 11:28 ShadowBot https://github.com/minetest/minetest/issues/2376 -- Add find_surface_nodes_in_area Lua call which permit to only get the nodes which are under air nodes by nerzhul 11:28 nrzkt thanks celeron55 13:22 Zeno` I re-opened #2384; closing it is silly IMO 13:22 ShadowBot https://github.com/minetest/minetest/issues/2384 -- Mapgen v6 generates floating desert stone 13:39 kilbith thanks for that Zeno`, and totally agree 13:39 nrzkt +1 13:41 kilbith i'm afraid that paramat wants to transform v6 in a copy of v5 or v7, with floating things, etc. 13:43 kilbith anyways, closing an obvious bug is not correct from a core-dev... 13:44 kilbith *bug report 14:12 nrzkt we are not perfect :) 14:26 nrzkt https://github.com/minetest/minetest/pull/2325 is good for me. Zeno` ? 14:26 nrzkt i tested it and it works very well 14:27 * Zeno` looks 14:27 est31 nrzkt, what parts of inventory do you think need to be rewritten? 14:28 nrzkt hmmmm i don't know but the stacking / destacking is very very slow and very complicated 14:28 est31 so mouse click -> actual stacks? 14:29 est31 or sth else 14:29 Zeno` geez, the whole thing is unnecessarily complicated if you're talking about the controls (a simple state machine would be better than hardcoded logic) 14:29 Zeno` anyway I'm generally ok with 2325 14:29 est31 Zeno you mean adding it to f7? 14:29 nrzkt the functions to stack in inventory.cpp 14:30 Zeno` est31 I just meant how controls are handled 14:30 Zeno` (I wasn't referring to 2325) 14:30 est31 ah ok not pr issue but general 14:30 Zeno` yeah 14:32 Zeno` nrzkt, I'm not sure but I think there is an open issue by Sokomine regarding some stuff probably related to what's in inventory.cpp 14:32 nrzkt then if we talk about 2325 Zeno`, can we merge ? 14:33 Zeno` 2325 the only negative thing I have to say about it is the g_settings->...() calls, but that's a problem that needs to be solved independently IMO... so I have no objection ;) 14:34 nrzkt ok perfect, i merge 14:35 nrzkt any new comments on #2376 ? 14:35 ShadowBot https://github.com/minetest/minetest/issues/2376 -- Add find_surface_nodes_in_area Lua call which permit to only get the nodes which are under air nodes by nerzhul 14:39 Zeno` hmm, that one I'd like to look more at 14:40 Zeno` it's tricky because of block boundaries 14:41 est31 it should have documentation too 14:41 est31 inside lua_api.txt 14:43 Zeno` on line 606 it would probably be quicker to have if (c != CONTENT_AIR && csurf == CONTENT_AIR && filter.count(c) != 0) { 14:43 Zeno` so that filter.count is not even called if the two cheaper conditions are not true 14:47 nrzkt est31: i agree for lua_api.txt, forgot it 14:48 nrzkt right Zeno`, good improvement, and faster, because CONTENT_AIR could be many blocks 14:49 kilbith *nodes 14:50 nrzkt yes, yes :) 14:50 nrzkt sorry i'm at work, building new JAVA environments for my web developers :p 14:54 nrzkt est31, Zeno`: rebased, lua_api & condition optimisation 14:55 nrzkt can you review another time and tell me if we could merge ? 14:56 Amaz I've just tried out the cinematic mode, and noticed that the enabled/disabled text starts with a lowercase letter, rather than an uppercase letter, which is the case for the other options that show text in the middle of the screen... 14:59 kilbith uh, no. others options starts by a lowercase letter 15:00 Amaz I think that is because they are variable names... f1, f2, f3, f5, f6 all have uppercase letters. 15:00 kilbith (eg. fast, fly, noclip, etc) 15:01 kilbith that doesnt matter anyways 15:01 Amaz It is a discrepency... 15:01 Amaz There should be one case used, at least for the f keys. 15:06 est31 nrzkt, so your definition of surface node is any arbitrary node that has air above? 15:06 nrzkt it's the goal 15:06 nrzkt because the mods which need this API want to find nodes where place some things 15:07 est31 so they do the same check just in lua 15:07 est31 and the speed enhancement is because its c++? 15:10 nrzkt no 15:10 nrzkt it's because we can filter easily with C++ than LUA because we can get some informations faster 15:10 kilbith *Lua 15:10 nrzkt high level lua calls can be good, because mods are good, but if we could give them high level API then can be faster 15:11 est31 thats true. its good to have high level api 15:11 nrzkt s/then/they/ 15:12 nrzkt better mods = better API :) 15:12 nrzkt mods are not only good because we have an API with many functions, but also because we could give them high level functions to implement easily some things 15:14 Zeno` what's SN mean by every node checked twice? 15:14 Zeno` oh 15:14 Zeno` I see 15:14 nrzkt if you find a faster way, but i don't think we can improve that 15:14 Zeno` sec 15:15 nrzkt we need to find if a node has air over it 15:15 nrzkt then we check the node itself because we filter its content, then we check the upper node 15:17 nrzkt est31: fix lua_api.txt 15:17 Zeno` http://dpaste.com/0ANADQM <--- *****NOT TESTED***** 15:18 nrzkt no 15:18 Zeno` no? 15:18 nrzkt why decrease Y ? 15:19 nrzkt and why p_above is y -1 ? it's not above :p 15:19 Zeno` so you only have to call getNodeNoEx once in the inner loop 15:19 nrzkt hmmmm 15:19 Zeno` err yes, that should be + 1 15:20 nrzkt oh i see 15:20 nrzkt but i prefer lower to upper 15:20 nrzkt and it's not truc 15:20 nrzkt true 15:21 nrzkt because i need to test y and y + 1 , not y and y + 1 and y + 2 ... 15:21 Zeno` anyway that *should* be faster (when it's fixed) but I'm not sure if that's what SN was talking about 15:21 nrzkt surface node: P is dirt and P(y+1) is air 15:22 Zeno` it is doing that, right? (assuming the +1 is fixed) 15:22 nrzkt but P is not a surface node if P(y+2) is air. 15:23 nrzkt and if i understand your loop i test P with every node above 15:23 Zeno` x, y, z is the surface node 15:23 Zeno` hence push_v3s16(L, v3s16(x, y, z); (missing parenthesis there hehe) 15:23 nrzkt your test is wrong 15:23 Zeno` As I said, not tested but the idea is correct :P 15:23 nrzkt no 15:24 nrzkt if in the mapblock i have dirt AIR dirt AIR dirt AIR (stack Y) 15:24 nrzkt the 3 dirts on axis Y must return, with your loop only the first found 15:25 Zeno` I don't understand why, but I'm tired 15:25 nrzkt then you test only search the first node which has air above, then it's not correct 15:25 Zeno` no, I am testing both nodes 15:25 Zeno` that's what c = c_above is for 15:26 nrzkt oh 15:26 nrzkt i see, sorry , i have missed this 15:26 nrzkt this change everything :p 15:29 Zeno` :P 15:29 nrzkt ok i rebased it using your idea 15:29 Zeno` the idea is correct, but I've probably missed something (I wrote it in, what, 50 seconds? lol) 15:30 shadowzone Zeno`: look on #minetest 15:30 nrzkt erf, i forget to rewrite p 15:31 Zeno` s/wrote/re-arranged 15:34 Zeno` actually if you want to go really crazy you might be able to do if (c == CONTENT_AIR && c_above != CONTENT_AIR) { //go up two nodes} else if (filter.count(c) != 0) {...} but I leave that as an exercise heh 15:35 Zeno` no point in optimising it much further if it's not necessary 15:35 nrzkt let's keeping a readable code :p 15:35 Zeno` yep, that's what I mean 15:36 nrzkt s/keeping/keep 15:36 nrzkt then, are you right with the last version ? 15:36 Zeno` if it works 15:37 Zeno` but I think SN should be consulted because of his comment(s) 15:37 Zeno` it *looks* right :) 15:38 Zeno` still dunno about those map boundaries but perhaps it's not an issue 15:39 nrzkt not an issue yes 15:42 * Zeno` looks again 15:44 Zeno` seriously, I'm fine with it. But I *would* prefer to have ShadowNinja's input as well since he commented on the PR 15:46 Zeno` then again I think his issues have been addressed 15:46 Zeno` I do see how VanessaE's issue arises though 15:48 Zeno` If the player is moving into a new area then the adjacent area is (probably) not merged yet (in the x, z directions). That might cause the stripes? 15:48 Zeno` I dunno :( 15:49 Zeno` actually it shouldn't, so scratch that 15:53 Zeno` be back in a while :) 15:58 nrzkt VanessaE issue is unrelated, it's related to mapgen 17:26 ShadowNinja ~tell nrzkt Your pull still 1. doesn't support air nodes other than CONTENT_AIR (makes this useless for moon games and the like). 2. Requires a list of every non-air node to be passed to it. It should use nodedef fields to check if something's a solid/air node. 17:26 ShadowBot ShadowNinja: O.K. 21:13 MattJ Will there be new stable builds in the PPA? For Ubuntu 14.10? 21:17 kilbith MattJ: the 14.04 builds are compatible with 14.10 21:17 MattJ Ah, good to know, thanks 21:18 MattJ But 0.4.12 isn't built for any dist yet 21:18 kilbith MattJ: https://launchpad.net/~minetestdevs/+archive/ubuntu/daily-builds 21:18 kilbith ^ 0.4.12 21:18 MattJ I was worried about those, given that they track git and there are unstable changes coming 21:19 kilbith not that unstable, but certainly not adviced for mounting a server 21:25 MattJ which is what I was hoping to do :) 21:26 MattJ Might try rolling my own packages