Time Nick Message 01:14 paramat hmmmmm, i added line comments to your latest 2 PRs, bbl 02:53 jhcole sofar: https://github.com/minetest/minetest_game/pull/1067 05:38 nore ~tell paramat +1 to game#1067 05:38 ShadowBot nore: O.K. 05:45 hmmmmm wth 05:46 hmmmmm a comparison between minetest and voxellands is definitely off-topic for this channel. i can tolerate small back-and-forths of off-topicness, but when it becomes the entire topic for like 100 pages of text and drowns out legitimate dev chat, then i'd say that shouldn't be allowed 05:47 hmmmmm in any case, if that guy wants to implement javascript, scheme, lisp, etc. extensions then he's totally free to 05:53 galex-713 I was trying to know the differences and the differences of goal between the two projects, since the goals of a project are applied by the devs and that’s the canal of the devs, I thought I would get more relevant answers here, and I mostly did 05:53 galex-713 hmmmmm: ^ 05:55 galex-713 sorry for the noise I thought #minetest was for the in-game-concerning things 07:41 nore lisac: are you there? 07:41 lisac Yep. 07:41 lisac I saw the pull request :D 07:42 lisac It won't run on my PC, though. 07:42 nore could you produce a backtrace please? :) 07:42 lisac Sure :D 07:45 lisac nore: http://pastebin.com/enQBiLGZ 07:46 nore hm... 07:46 nore *that* is strange 07:47 nore lisac: what does valgrind give? 07:47 lisac moment. 07:47 lisac what do I type? I'm quite new to debugging. 07:48 nore just, valgrind bin/minetest 07:48 nore like with gdb 07:49 lisac Ok. 07:50 lisac nore: http://pastebin.com/7p1mbVsf 07:50 nore ok, thanks 07:52 lisac np :D 07:52 lisac I'm going to shop, I'll be back in ~15 minutes. 07:59 nore lisac: could you test with lastest version please when you're back? :) 08:21 lisac nore: I'm back. 08:21 nore cool :) 08:22 lisac Compiling. 08:23 lisac nore: Seems to work, thanks! 08:23 nore ah, good 08:24 nore I reverted some changes, so could you try with the latest commit too please, where I re-added them? 08:24 nore (but in a different way) 08:24 lisac sure. 08:26 lisac Crashes on starting a game. 08:26 lisac but main menu works. 08:26 nore hm... 08:26 nore could you do a backtrace too please? 08:26 lisac sure. 08:28 lisac http://pastebin.com/g3J8Bqz9 08:30 nore hm, I guess dynamic_cast does strange things 08:30 nore I'm almost sure it is that, even 08:31 Zeno` interesting 08:32 Zeno` valgrind has the answers 08:33 Zeno` statictext.h line 176 has an invalid read 08:34 Zeno` why does it need to be cast anyway? 08:35 Zeno` if you remove the cast it crashes elsewhere 08:39 Zeno` oh yucky 08:41 nore Zeno`: I need to distinguish between irrlicht static texts, and ours 08:42 nore dynamic_cast works fine on my machine, but not lisac's, apparently 08:42 nore which is *strange* 08:42 lisac I can give you my system specs, if you want. 08:42 nore (I do not even get errors in valgrind) 08:42 nore lisac: yes, I would be interested 08:42 lisac Pentium D 820. 08:42 nore and you cmake command too 08:42 nore +r 08:42 lisac 3GB of DDR2 RAM... 08:43 lisac GeForce 210. 08:43 nore hm, most important would be OS I think 08:43 lisac D2151-A1 Motherboard.... 08:43 lisac Kubuntu Linux 14.04 08:43 nore ok 08:43 lisac So... I'm using cmake ../ 08:44 lisac * cmake .. 08:45 nore found that comment in code 08:45 nore / no dynamic cast possible due to some distributions shipped 08:45 nore / without rtti support in irrlicht 08:45 nore --> that's the answer 08:45 lisac So, what do I do? 08:47 nore lisac: that means I've got to change this code 08:47 lisac Ah. Well, ok. 08:48 Zeno` did you want valgrind output, nore? 08:48 Zeno` it doesn't tell you much more than the trace though 08:48 Zeno` (if anything more) 08:48 nore Zeno`: I don't need it, I just need to avoid that dynamic_cast 08:49 Zeno` 2016-05-03 18:41:52: ACTION[Server]: singleplayer joins game. List of players: singleplayer 08:49 Zeno` RTTI symbol not found for class 'irr::gui::CGUIStaticText' 08:49 Zeno` yes :) 09:10 nore lisac: Zeno`: could you try again please? 09:10 nore the code does not use dynamic_cast anymore, it should work fine now :) 09:10 lisac sure, gimme a moment. 09:12 lisac How'd I use the colored chat? :D 09:13 nore lisac: you can't really "use" it 09:13 nore but you can do /help 09:13 lisac Ok. 09:13 nore also, if you have an error (like a missing texture), it should show up in red 09:13 lisac /help all 09:13 lisac it gives all white. 09:14 nore :( 09:14 nore this is *strange* 09:14 * nore checks the version that is committed 09:15 nore it *should* be colored 09:15 lisac Well, I'm using min. dev. test. 09:15 nore chat commands are in builtin anyway 09:15 lisac It might be the issue due to fact I'm not using run in place. 09:16 nore hm, I don't think so 09:16 lisac Ok. 09:28 nore lisac: could you re-test, and tell me what you get on actionstreamN 09:28 nore s/N/?/ 09:28 nore i.e. do you get a lot of "COLORED TEXT" printed, or not? 09:31 nore lisac: wait, what's your irrlicht version? 09:36 nore lisac: could you test with latest commit too please? :) 09:37 * nore is going to eat, bbl 09:46 Zeno` nore, I'm eating dinner and a few other things right now but I can test in about an hour or so if you still need it 09:47 nore Zeno`: that would be perfect :) 09:48 nore Also, if you can try with the last commit, and without it, it would be good :) 09:54 lisac Moment, pleae. 09:54 lisac *please 09:54 lisac I was playing super tux kart. :D 09:55 lisac nore 09:56 lisac nore: 2016-05-03 11:56:34: ACTION[Main]: COLORED TEXT 09:56 lisac I get ton of those. 09:58 lisac Version: 1.8.1+dfsg1-1 10:08 gregorycu lads 10:10 Zeno` gcu: masterchef is on and you're here? :-o 10:10 gregorycu Fuck masterchef and everything it stands for 10:10 Zeno` strong sentiment... haha :) 10:11 Zeno` nore, apart from the actionstream COLORED TEXT as lisac noted (I assume you added it for debugging) everything seems fine 10:12 nore lisac: Zeno`: so you get colored text? 10:13 nore Also, what happens when you remove the last commit? 10:17 gregorycu My bishop is broken 10:17 Zeno` Oh, I haven't tested that it actually works :D Just that it doesn't crash and valgrind doesn't complain 10:18 Zeno` I'll try removing the latest commit. I'm not properly back yet (maybe 15 more minutes) 10:20 lisac Nope. 10:20 lisac Doesn't work. 10:21 Zeno` nore, I thought you were in France. Don't they spell colour properly their either? :p 10:21 Zeno` s/their/there 10:22 nore Zeno`: well, I alternate between UK and US spelling 10:22 lisac It is faster to write 'color'. 10:22 gregorycu It's faster to write c 10:22 gregorycu Doesn't make it right 10:22 nore I learnt British English, but I've read a lot of American English on the Internet... 10:23 gregorycu Make sure you right the correct thing, always 10:23 nore :D 10:29 gregorycu Hooray, my bishop is now behaving himself 10:30 gregorycu Next is birds 10:30 nore gregorycu: pathfinder? 10:46 lisac bye everyone 10:46 nore bye 10:48 Zeno` nore, you want me to revert "Another test to fix the bug"? 10:48 nore yep 10:49 Zeno` ok 10:49 Zeno` no segfaul 10:50 Zeno` t :) 10:51 Zeno` and valgrind is clean also 11:44 Zeno` #4072 anyone? 11:44 ShadowBot https://github.com/minetest/minetest/issues/4072 -- Fix holding down F10 (open console) causing GUI to freeze by Zeno- 11:47 nore Zeno`: sorry, I must leave now, so can't test: 11:47 nore s/:/ :\// 11:47 nore but looks fine 11:47 Zeno` np :) 12:18 gregorycu yeah, pathfinder 12:33 nore gregorycu: is it committed anywhere btw? 12:34 gregorycu A copy is 12:34 gregorycu Not the latest 12:34 gregorycu I'll commit the latest 12:34 nore I'll look then :) 12:34 nore How fast is it now? 12:36 gregorycu Depends on the length 12:36 gregorycu For < 15 distance, about 500 microseconds 12:37 gregorycu Most of the slowness is fetching block data 12:37 nore Wasn't it 250 yesterday? 12:37 gregorycu Yeah, that was very short distances, about 9 12:37 nore Ah, ok 12:37 nore So, quite fast then 12:37 gregorycu Also depends on the rule 12:37 gregorycu Yeah, fast enough 12:37 nore What rules du you have for now? 12:38 gregorycu Also, I know a way to sacrifice speed for accuracy 12:38 gregorycu Or rather, accuracy for speed 12:38 gregorycu I will make that an option 12:38 gregorycu Animals don't have to be 100% optimal all the time :) 12:38 nore Indeed 12:38 nore What's your idea? 12:38 gregorycu But yeah, by default, will be perfect 12:38 gregorycu Just a factor, 1 to 10 12:38 gregorycu 10 being very accurate, 1 being less 12:39 gregorycu Basically amplifies the heuristic 12:39 gregorycu Makes the algo favour things that approach the goal faster 12:39 nore Ah, I see 12:39 gregorycu https://github.com/gregorycu/minetest_mods/tree/master/pather 12:39 gregorycu If you look at the code here, some of the options are in there 12:40 gregorycu minetest.find_path(src,dest,1000,{swimming_above=true}) 12:40 nore It's not optimal indeed, but that's not a problem and it is even more realistic for animals :) 12:41 gregorycu https://upload.wikimedia.org/wikipedia/commons/8/85/Weighted_A_star_with_eps_5.gif 12:41 gregorycu That's what I mean 12:41 gregorycu Suboptimal path, but also not horrible 12:43 gregorycu I added a new rule before 12:43 gregorycu diagonally=true 12:44 gregorycu And the C++ changes: https://github.com/gregorycu/minetest/tree/better_pathfinder 12:45 nore gregorycu: also, how does needed height work? I.e. if there is a step between 12:46 gregorycu Needed height is for things more than one block tall 12:46 gregorycu aka. players 12:46 gregorycu Or rather, ie. players 12:46 gregorycu It works... well 12:46 gregorycu Actually, I better check that 12:47 nore I mean, what happens in that situation (X is solid, . is not) 12:47 nore XX.. 12:47 nore .... 12:47 nore ..XX 12:49 gregorycu If required_height is 2 there 12:49 gregorycu That should be fine 12:49 gregorycu An extra X at the top or bottom will mean not enough space 12:50 gregorycu The middle column, column 3 means a player can fall down 12:50 gregorycu Need like a shared online sketchpad 12:50 nore Well, it should not be possible 12:50 nore There is no middle column... 12:51 gregorycu http://flockdraw.com/cn3v8g 12:52 gregorycu Is that what you mean? 12:53 gregorycu Doens't take long for the dicks to come out 12:54 gregorycu Well, what do you mean? 12:55 nore Nvm 12:55 nore Gtg 12:55 gregorycu ok 14:59 nore ~gregrorycu what I meant precisely was that you can have a 2 node height over one surface, and same over another, that is next to it and one node below; but if this is the case, a player can't go through since they have to fit in a one node high hole to get through 14:59 ShadowBot nore: 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). 14:59 nore ~tell gregrorycu what I meant precisely was that you can have a 2 node height over one surface, and same over another, that is next to it and one node below; but if this is the case, a player can't go through since they have to fit in a one node high hole to get through 14:59 ShadowBot nore: O.K. 15:02 rubenwardy [Help] Command "gregrorycu": runs profilers and optimises code 15:03 rubenwardy Usage: coffee > gregrorycu minetest > profit 15:21 gregorycu If a player can't move directly to the side, it excludes all potential positions below that position. 15:22 gregorycu You can try it out and see :) 15:24 gregorycu There is a minor bug where I don't detect head collision when the height is >1 15:25 gregorycu Which is fixed locally 15:25 rubenwardy I fail, coffee | gregrorycu minetest | profit 15:26 gregorycu Lol 15:26 gregorycu Cheers Ruben 15:26 hmmmmm lol 15:26 hmmmmm too many instances where fixing one thing breaks another thing 15:27 * gregorycu is what it is 15:27 gregorycu Isn't a perfect command 15:28 hmmmmm hmm i'm referring to the chat console 15:28 hmmmmm https://github.com/Zeno-/minetest/commit/8c951cae5bcfa715e227d47d122cc2de45d70a63 15:28 gregorycu Ahh ok 15:28 hmmmmm commit that breaks it 15:28 hmmmmm ironically another 'fix' 15:28 gregorycu (obviously you were not talking about my pathfinding code) 15:34 gregorycu Hey ShadowBot, you forgetting something? 15:34 VanessaE no, it isn't. nore can't spell :) 15:35 gregorycu Lol 15:35 gregorycu Did that answer your question nore? 15:37 gregorycu I'm going to bed in 5 min. Otherwise you can ask me in 6 hours time. 15:44 paramat any reviews for simple PR #4072 ? 15:44 ShadowBot https://github.com/minetest/minetest/issues/4072 -- Fix holding down F10 (open console) causing GUI to freeze by Zeno- 15:47 Zeno` yeah I think it's great 15:48 paramat well, you can self-approve :) 15:48 Zeno` already have :P 15:48 hmmmmm paramat, PLEASE stop that. 15:49 hmmmmm the idea of self-approval is bunk. 15:49 hmmmmm it holds no weight whatsoever. 15:49 hmmmmm I don't know about you, but whenever I see "one approval" and then I look at the PR thread and see nobody actually approving, I say to myself, "oh, it's one of those self approvals. aka means NOTHING at all" 15:50 hmmmmm it's kind of like using yourself in a citation 15:51 paramat we've been through this before though. i really don't mind not being able to self-approve, but labelling your own approval is useful for searching by label. so let's actually decide 15:51 hmmmmm if you don't think your PR is ready, why don't you use the WIP tag? 15:52 paramat yeah i'm happy with that 15:52 paramat well 15:52 paramat searching labels for 'two approvals' would fail with that approach 15:53 paramat but i really don't mind either way, as long as we decide on a methid 15:53 paramat (method) 15:53 Zeno` if it has two approvals the person should just merge it 15:53 Zeno` so the two approvals tag is useless for the majority of PRs 15:54 hmmmmm right my point is you can call it whatever you want 15:54 hmmmmm you just need to be aware that most likely others won't see it the way you see it 15:54 paramat no, very often a PR will get 2 approvals but it can't be merged until later 15:54 paramat (due to dev business) 15:54 paramat (busyness) 15:55 paramat hmmmm yes self-approving does upset people :) 15:55 paramat as they get stuck on the literal meaning instead of knowing the usefulness 15:57 paramat however labelling one's own approval is helpful as a label, that's why i'm doing it 15:57 PilzAdam hmmmmm, there seems to be a consensus among all other core devs that self-approving is legit 15:57 Zeno` PilzAdam, it *is* legit. But there is no point in adding a label to self-approve 15:57 paramat well it's legit, but actually showing it in a label is the issue 15:57 Zeno` otherwise why would you even make the PR? 15:57 PilzAdam ah, okay 15:58 paramat nope, often the contributor will make a PR but not be confident about it 15:58 Zeno` paramat, well they should say so 15:58 paramat sure, but that won't show up in labels 15:58 Zeno` paramat, I'm sorry but AFAIK you're the only one who likes core devs adding the label to their own PRs 15:59 Zeno` If you need feedback, label it with "needs testing", or "WIP" or add a comment :/ 16:00 paramat it's useful when filtering for 'two approvals' 16:00 Zeno` just filter for one approval and the end result is the same 16:00 Zeno` (unless you don't know who the core devs are I suppose) 16:00 paramat ok 16:01 Zeno` anyway, whatever :) 16:01 paramat i'm not too bothered about it, will change behaviour 16:03 paramat so, inless it's made clear by a WIP label or in the description, we'll auto-assume self-approval if a core dev makes a PR 16:03 paramat (unless) 16:04 paramat or rather, 1 of the 2 required core dev agreements is assumed 16:05 paramat that's what the label is referring to 16:07 paramat so essentially, people get upset over actually stating this in a label, even though self-approval is still actually happening :} somewhat weird but ok 16:21 sofar sfan5: "This translation is currently locked for updates!" 16:22 sofar I guess I don't understand weblate yet. 16:29 sofar can't translate, no idea why 16:46 paramat #4080 is simple, any reviews? 16:46 ShadowBot https://github.com/minetest/minetest/issues/4080 -- Fix Android player_stepheight for snow block by MoNTE48 16:46 sofar I gave it a friendly thumbs up 16:46 hmmmmm that is so hacky 16:47 hmmmmm yeah i guess it's fine for now. 16:48 paramat what would be a less hacky fix? 16:49 hmmmmm i don't know.. 16:49 hmmmmm the main issue is that jumping on android is difficult 16:49 paramat ok, i'll merge this later 16:51 paramat and maybe 4072 also once i test 16:52 hmmmmm personally, i think minetest should be pulled from the Play store until there is a maintainer who can actually ensure the quality of minetest and fix bugs... 16:52 sfan5 sofar: but locked by who 16:53 sofar sfan5: no idea? est31 for updates? 16:53 hmmmmm releasing to a major platform when there is literally no support or QA is reckless 16:53 sofar all languages are locked 16:53 sofar not just dutch 16:53 sfan5 o.O 16:59 paramat i still somewhat wish we didn't bother with android at all, but then it may be better players use an official app rather than unofficial ones 16:59 paramat monte seems happy to help out with android 17:08 nore VanessaE: heh, I'm on a phone, it's harder 17:09 MoNTE48 Hi. #1454 and #2122 17:09 ShadowBot https://github.com/minetest/minetest/issues/1454 -- Irrlicht sometimes reports unexpected multibyte character 17:09 ShadowBot https://github.com/minetest/minetest/issues/2122 -- Volume key Android 17:11 MoNTE48 + #2571 17:11 ShadowBot https://github.com/minetest/minetest/issues/2571 -- Android: Taking a screenshot floods the chat with an error 17:15 paramat 2571 is a duplicate of 1454 17:16 MoNTE48 I know, but the essence. I fixed it last night 17:16 paramat ok 17:17 MoNTE48 Remove this https://github.com/minetest/minetest/blob/master/build/android/patches/irrlicht-back_button.patch#L7 and add https://gist.github.com/MoNTE48/0906326741c8b4bc39938c37869d5c3c on nativeactivity.java 17:19 paramat a PR would be much appreciated 17:19 hmmmmm how could any of these items be independently tested? 17:19 hmmmmm i'm sorry we're not just gonna take some person's word that it works, trust me 17:19 hmmmmm just "trust me" 17:22 MoNTE48 hmmmmm 17:22 MoNTE48 hmmmmm 17:22 MoNTE48 hmmmmm, it's not my problem. I found a solution to that error of 3 years, and this is a serious mistake. Check - your job. After all, you are recognized as the programmers, I do not 17:22 hmmmmm OK, that being said, don't get upset if your PRs sit around for a while 17:23 MoNTE48 I'm not said. I just work here https://github.com/MultiCraftProject/MultiCraft 17:27 paramat nore sfan5 PilzAdam #4083 can be closed? i think so. 17:27 ShadowBot https://github.com/minetest/minetest/issues/4083 -- use_texture_alpha =true should Be on by default 17:28 sfan5 yes 17:28 paramat ok 17:30 paramat closed 22:55 Fixer australia man is back 22:56 Fixer gregorycu, shadowninja managed to randomly reproduce lighting bug on pr 3770 22:56 gregorycu so essentially, people get upset over actually stating this in a label, even though self-approval is still actually happening :} somewhat weird but ok 22:57 gregorycu He's right you know. 22:57 gregorycu Either you want self approval, or you don't. Tags should reflect reality. 22:58 gregorycu Fixer, I came in just to say that. Now I gotta get ready for work :) 22:58 gregorycu That's good though 22:59 Fixer is not approval is by definition means approval of other persons? how you can self-approve? by making pr you already self-approve anyway $) 23:00 jhcole I'm confused. Can someone explain what happens to meta set on a falling node after it falls? 23:09 jhcole Its seems like the meta information does not 'fall' with the node. I'd really like to be able to attach extra information like meta to a node that might fall, and still be able to access it after it falls. Does anyone know a way to do this? 23:48 gregorycu jhcole that is a legitimate question. But with no answer. 23:48 gregorycu Try creating a bug for this 23:49 gregorycu Maybe it's behaviour that should change 23:49 gregorycu Fixer, in answering your question 23:49 Fixer gregorycu, everybody is sleeping, except australia, I guess 23:49 Fixer Australia* 23:49 Fixer i will go sleep in 5 minutes :) 23:50 gregorycu It's to make this true: all PRs that have one approval, have the one approval tag. 23:50 Fixer yes 23:50 Fixer self-appr does not count as tag 23:50 Fixer right? 23:50 gregorycu How do you find PRs that only need one additional approval? 23:51 gregorycu What tag do you search on? 23:51 Fixer with tag "one approval" 23:51 Fixer in ideal world 23:51 Fixer some devs may approve inside but without text :/ 23:51 Fixer erm without tag* 23:52 gregorycu But that won't pick up things that are self-approved if they don't tag it 23:52 gregorycu I generally agree with hmmmm with a lot of things, but I think he is wrong here 23:52 Fixer how you can post the code in PR that you don't self-approve? %) 23:52 Fixer is approve word means aggrement from outside? 23:53 gregorycu Maybe it's not your code? But that's irrelevant. If github had an auto tag feature, we should use that 23:53 gregorycu But hmmmm's problem isn't one of efficiency 23:54 gregorycu " I don't know about you, but whenever I see "one approval" and then I look at the PR thread and see nobody actually approving, I say to myself, "oh, it's one of those self approvals.  aka means NOTHING at all"" 23:54 Fixer i'm not even a dev to discuss this %) 23:54 gregorycu He has a problem with self-approval itself, not the tag 23:55 Fixer you can't self-approve yourself by definition, somebody needs to validate it and approve (from outside) 23:55 Fixer my kind of logic 23:55 gregorycu That's not really true 23:55 Fixer just my nondev opinion 23:56 gregorycu Self-validation is a thing 23:56 gregorycu It's a senantical argument anyway 23:57 Fixer whatever :) hmmmm will probably join irc soon 23:57 gregorycu Probably as I arrive at work 23:59 Fixer devs need developer clock that show time of other devs :) 23:59 Fixer europe is sleeping, usa is at work, australia goes to work 23:59 Fixer usa is finishing the work* 23:59 Fixer gn