Time Nick Message 00:10 paramat now pushing to mtgame 00:16 paramat complete 00:17 VanessaE mmmmh 00:17 VanessaE stop giving me reasons to rebase :) 00:17 est31 lol 00:18 est31 db bases on really really old mtgame, no? 00:18 VanessaE not anymore 00:19 VanessaE my last resync was a bit over a month ago 00:24 paramat :) 03:12 ShadowNinja est31: That usit test is actually broken to begin with. Both "a,b" and "b,a" are valid return values for minetest.privs_to_string{a=true, b=true}. Lua leaves the iteration order undefined. In fact when I updated the Lua version to 5.2 that usit test broke because the orderind wasn't consistent anymore. 03:13 ShadowNinja s/usit/unit/g 03:13 ShadowNinja VanessaE: That issue wasn't platform-specific. The hyphen was missing from the clients allowed-chars list. 03:14 VanessaE ShadowNinja: odd, because it actually caused the win client to hang outright 03:14 VanessaE but it worked in linux 03:14 ShadowNinja In any case, this is all server-side, the client doesn't know anything about any files. 03:14 ShadowNinja (in this case) 04:49 est31 in 5 mins: https://github.com/est31/minetest/commit/b8a8be9c86cf6249d7c1be1c01d27deb02bcb58c 04:49 est31 *pushing* (in 5 mins) 07:57 hmmmm we need a meeting 07:58 hmmmm a big long, long meeting where everybody is here, so we can talk about the status of things, future plans, and decide on which items take priority 07:58 hmmmm but alas, timezones 07:58 hmmmm and everybody's busy 08:01 VanessaE hmmmm: and those of us who ARE here can't help. :-/ 08:01 hmmmm too bad we don't have the luxury of holding conferences or summits like other big open source projects 08:02 VanessaE yeah 08:02 * est31 is there 08:02 hmmmm i dunno, it seems once again i've been pushed into this managerial role 08:02 hmmmm because nobody else is consistently around 08:03 kahrl I could probably attend a meeting this or next week 08:03 hmmmm maybe this level of activity is appropriate for what minetest is 08:03 hmmmm and it's not a problem 08:03 rubenwardy A TestCon would be awesome. 08:04 hmmmm you were going to say MineCon, but that's already taken, isn't it 08:04 hmmmm I was thinking for a minute that maybe, just maybe, we could arrange a meetup at minecon whenever that is 08:04 rubenwardy Yeah 08:04 hmmmm but then I have to somehow arrange for travel 08:04 VanessaE can you imagine the hate?? 08:05 hmmmm and this isn't some big corporation that can just eat airfare/hotel/rental car/food costs 08:05 rubenwardy It's very true. We're all quite spread out. 08:05 rubenwardy Virtual TestCon? 08:05 hmmmm lol 08:05 rubenwardy All meet up in game? 08:05 hmmmm time zone issue 08:05 est31 hehe 08:05 rubenwardy oh yeah 08:05 hmmmm that's the primary problem 08:05 paramat 'ConTest' 08:06 hmmmm half the developers are GMT +5 or +6, half the developers are GMT+12, half the developers are GMT+23 08:06 hmmmm that's three halves 08:06 rubenwardy lol 08:06 hmmmm erm, UTC i meant 08:06 kahrl aren't there lots of german ones too, that would GMT+1 or GMT+2 08:06 hmmmm yeah 08:07 Calinou make a monthly meeting at a timely date 08:07 hmmmm perhaps 08:07 hmmmm see, I'm not so sure if this is the correct thing to do 08:07 Calinou if they are too common, people won't join 08:07 VanessaE roughly 8pm CET seems to be the highest activity point in my servers, if that helps any 08:07 hmmmm people work on minetest because there are no commitments 08:07 rubenwardy It would be good to see a map of where people are, developer wise. 08:08 hmmmm they can just work on it when they want to and that's it 08:08 Calinou 8pm CET is a good time for online meeting 08:08 VanessaE (i.e. the most random people available to meet somewhere) 08:08 Calinou we probably should make a specific channel for it? 08:08 hmmmm if we start pushing developers to make commitments, that could turn people off to the whole idea 08:08 VanessaE problem is that leaves out most US-based people unless it's done on a weekend 08:09 hmmmm adopting business practices in general would make development so much more efficient but violates the entire principle of what minetest is 08:09 hmmmm so are people satisfied with the current amount of development activity? 08:10 rubenwardy If it were a business / studio, I would think that it was going too slow. 08:10 VanessaE the amount of activity is fine by me, hmmmm. it's the parts of the project at which that activity is directed that may be in question. 08:10 est31 the question is: how to rise 08:10 rubenwardy But it's not, and we're volunteer based, so it's the right amount 08:10 paramat yeah, people don't have much time, so i don't expect a high activity 08:10 hmmmm hmm 08:11 hmmmm it's just that every time I hear somebody complain about something that I've been wanting to get fixed for a long time, it sorta hurts 08:11 Calinou game needs a quite a bit of fixes and improvements :P 08:11 hmmmm i want that badly too, you know! 08:11 hmmmm yes 08:11 hmmmm first off all the damn bugs 08:12 hmmmm then it needs an infinite amount of polish on everything 08:12 hmmmm I tried to make what I generally work on as polished as possible 08:12 rubenwardy 8pm is a good time to get Polish patches 08:13 hmmmm alright then 08:13 hmmmm 8pm CET 08:13 kahrl that was a pun, I think :P 08:14 VanessaE where is RBA when we need him :P 08:14 hmmmm wait a minute 08:15 hmmmm alright, 10 AM EST 08:15 rubenwardy Yeah, it was a pun. XD 08:15 hmmmm I think we can do this 08:15 hmmmm well at least I can 08:15 VanessaE 10am EST? where'd you figure that from? 08:15 * hmmmm scratches head 08:15 rubenwardy 10 AM EST is better for me than 8pm CET, but I'm not a core dev 08:15 hmmmm CET is UTC +1 08:15 VanessaE 8pm CET would be 2pm EDT I think 08:15 hmmmm EST is -5 08:16 VanessaE we're in EDT now :) 08:16 hmmmm so the difference is 6 08:16 hmmmm oh 08:16 hmmmm so -4? 08:16 VanessaE I think 08:16 rubenwardy 10am EST is 3pm CET 08:16 hmmmm 8 pm CET - 5 == 3 08:16 VanessaE yep 08:16 VanessaE it is 08:16 hmmmm whoops 08:16 hmmmm i added instead of subtracted 08:16 VanessaE EDT = UTC-4 08:16 hmmmm 3pm is very doable 08:17 hmmmm when do you want to have these planning meetings/summits/whatevers 08:17 hmmmm not right away, we need time to get the word out or whatever 08:17 kahrl I don't ever attempt to convert timezones in my head, it's just too error prone 08:18 VanessaE kahrl: I looked it up :P 08:18 VanessaE hmmmm: twice month on the weekends initially, maybe? 08:18 VanessaE drop that to once a month as things get ironed out 08:18 hmmmm maybe once a month 08:18 hmmmm make it seem exclusive 08:18 hmmmm a surprising amount can get ironed out just by sitting down you know 08:19 VanessaE yep 08:19 kahrl once a month sounds fine to me 08:19 paramat monthly seems good 08:19 VanessaE the irony of that statement is that "just talking" has been said more than once to be a waste of time. I think we know better though :P 08:20 hmmmm it's all a balancing act 08:20 hmmmm even i say that "less talk, more code" but we definitely need to regroup and synchronize 08:20 VanessaE yes. 08:20 hmmmm there are major bugs that apparently nobody is assigned to 08:21 hmmmm there are new features floating around but we don't have any ETAs on 08:21 hmmmm we have no timelines 08:21 hmmmm there should probably be a 0.4.13 release soon but who knows when that'll be 08:21 hmmmm ahgh 08:22 VanessaE *head explodes* 08:22 hmmmm if only we could enslave people for 5 days per week 08:22 hmmmm all this crap would get done so fast 08:22 kahrl to be most efficient, the meeting should have a clear agenda that is known to all attendants at least a day before (ideally earlier) 08:23 kahrl so that everyone can read about the issues that will be discussed 08:23 hmmmm i have a rough idea 08:23 hmmmm i'll write an outline 08:23 hmmmm soo yeah 08:23 hmmmm monthly status and synchronization meetings 08:23 hmmmm weekly pull request grooming meetings 08:23 est31 yesyesyes 08:23 hmmmm any other meetings? 08:24 VanessaE I think those alone are enough 08:24 est31 yea 08:24 Calinou https://etherpad.net/ 08:24 Calinou https://ethercalc.org/ 08:24 hmmmm all this "planning to plan" stuff is so corporate 08:24 VanessaE anything else I could suggest could just as easily be addressed in one of those two categories 08:24 Calinou if you need tools for planning, use those 08:24 hmmmm i would rather be coding right now 08:25 hmmmm it's 4:30 am and I'm awake talking about minetest meetings 08:25 hmmmm lolol 08:25 hmmmm later, I need to go to bed right now 08:25 VanessaE nooooooo 08:26 hmmmm i'm gonna be really tired for work tomorrow 08:26 hmmmm and i already wake up an hour late 08:26 VanessaE btw 08:26 VanessaE this is another fault we keep running into 08:27 VanessaE :) 08:27 VanessaE I mean the only-ever-on-late issue, could be theoretically addressed by meetings with defined times 08:28 VanessaE (but that doesn't help much if not everyone can meet then) 08:29 * paramat pets the issues *purr* 08:36 VanessaE my turn to fade away too. 08:36 VanessaE night. 08:37 * VanessaE pets her cat, since the issues have rough edges :P 08:48 est31 Ideas for: https://github.com/minetest/minetest/pull/2751 08:48 est31 ? 08:48 est31 API design is soo hard. 08:49 est31 esp. finding an API which is modular but still easy to use 08:49 est31 and fast 08:49 est31 and maintainable 08:50 est31 and without too much complexity 08:58 est31 ah thats why 08:58 est31 lol 08:59 est31 first I thought, why does this code do res = some_function() \n if (res) blabla 08:59 est31 but now I understand 08:59 est31 if you do if (res= some_function()) blabla, it might be shorter, but people may mistake 09:59 est31 can ppl have a look at https://github.com/minetest/minetest/pull/2751 10:00 Megaf_ Hi all 10:00 Megaf_ https://github.com/minetest/minetest/issues/2697#issuecomment-107903174 10:01 Megaf_ #2697 10:01 Megaf_ Mouse seems laggy in 0.4.12 (and 0.4.11) on OSX #2697 10:02 est31 I can't fix it because I cant reproduce 10:02 est31 what about submitting a PR 10:02 est31 and getting people with macs to test it 10:02 est31 if they say its ok, it can be merged 10:03 Megaf_ Im still testing the issue but I have no idea on how to fix it 10:05 est31 no problem, even triaging is helping 10:05 Megaf_ est31: problem is fps_max 10:05 Megaf_ instead of limiting the view range, it limits everything 10:06 Calinou if we disable fps_max on OS X, people will complain about overheating 10:06 Calinou and low battery life 10:06 Megaf_ if I set fps_max to 40, game becomes unplayable 10:07 Megaf_ Calinou: so we need to fix fps_max 10:07 Megaf_ to make it not limit the camera movement 10:09 Megaf_ AFAIK fps_max worked limiting the viewing range 10:12 rubenwardy https://github.com/minetest/minetest/pull/2094 10:31 rubenwardy est31, code changed 10:34 est31 rubenwardy, best to have it cached. 10:38 rubenwardy No where else does 10:39 est31 its a map lookup for nothing 10:39 est31 and that is expensive 10:39 rubenwardy https://github.com/minetest/minetest/blob/master/src/hud.cpp#L344 10:40 est31 yea, current code is bad 10:40 rubenwardy Adding a caching system is beyond the scope of this commit 10:40 rubenwardy unless 10:40 est31 cache it with static float hud_scaling = g_settings->getFloat("hud_scaling"); 10:40 rubenwardy static s32 scaling = ... 10:40 rubenwardy inb4 10:40 rubenwardy :( 10:41 est31 in the longer term we should switch to an enum based approach 10:42 est31 I have a PR for that, but people dont like it 10:42 est31 for good reasons 10:42 rubenwardy Can you change hud_scaling in the main menu? 10:42 est31 (technical details) 10:42 rubenwardy Caching it would mean you'd have to restart the game. 10:43 est31 yes 10:44 rubenwardy code updated 10:44 est31 thats why I made the PR the first place 10:45 est31 but it should be more separate, and generic 11:08 rubenwardy https://github.com/minetest/minetest/issues/2501 11:57 est31 pushing in 5 mins: https://github.com/est31/minetest/commit/ba3ff5ef39d768da4a0511d22a220b9aafa3d3f6 12:04 est31 pushed 16:40 est31 man this auth.cpp needs refactoring 16:40 est31 esp. overloading is making stuff hard 16:40 est31 if you have to count the arguments, you know you did sth wrong in the design 16:40 est31 also, no camelCase 16:41 est31 but thats the smaller issue 16:43 hmmmm by the way, do people actually like camelCase? 16:44 est31 its faster for the reading, same additional steps to do for the writing 16:44 est31 same amount 16:45 rubenwardy I use it for functions, and underscores for locals 16:45 est31 on my keyboard its even faster 16:45 est31 but minetest isn't very "unified" in this matter 16:46 est31 e.g. the lua api has l_do_this_and_that 16:47 est31 meh, coding style is important, but shouldn't become too important 16:47 hmmmm it's only important if it creates a barrier to writing the code 16:48 hmmmm or others reading it 16:48 rubenwardy Isn't being super picky making a barrier to writing code, though? 16:50 hmmmm here's something i wouldn't permit: https://github.com/RealBadAngel/minetest/blob/f0f29f7dc6b1a44424ed75e2226fd800b2b43d18/src/minimap.cpp#L267 16:50 hmmmm a lot of est's code is not conformant with the official style, but at least i'd permit it because it's still easily readable 16:51 rubenwardy Wow 16:51 rubenwardy Indentation is important like that 16:52 hmmmm also here's another example of something that isn't quite a dealbreaker, but I frown very much upon it: https://github.com/RealBadAngel/minetest/blob/f0f29f7dc6b1a44424ed75e2226fd800b2b43d18/src/minimap.cpp#L94 16:52 RealBadAngel https://github.com/RealBadAngel/minetest/commit/d7f15aae3b72975156b4989d219985149612fd86#diff-f20c82d93887e9a0f4d4eeb66cc7eaa7R242 16:52 hmmmm people who change their own code style mid-function 16:53 hmmmm now note that this code here is not final 16:53 hmmmm I'm just using it as an example of what would be acceptable or not acceptable 16:53 hmmmm yeah I know you fixed it 16:53 hmmmm sorry, that was NOT meant to be a rag on your minimap 16:53 RealBadAngel np :) 16:54 hmmmm I don't have any examples off-hand of blatantly wrong indentation 16:54 RealBadAngel sandboxes are often badly formatted, no matter what 16:54 hmmmm oh, actually I do, there's that one game that has possibly the messiest code imaginable that I pasted here 16:55 hmmmm if only I can remember the name of it 16:55 RealBadAngel i do have another flower 16:56 RealBadAngel https://github.com/minetest/minetest/blob/master/src/mapblock_mesh.cpp#L1237 16:56 RealBadAngel why the heck whole if is not commented out? 16:57 kahrl is it actually possible that m_mesh is NULL there? 17:00 RealBadAngel well, idk. maybe with all nodes made of air? 17:01 kahrl I don't see anything that would drop it other than the destructor 17:04 rubenwardy celeron55: https://gist.github.com/rubenwardy/4619a74d168f022697b3 17:08 RealBadAngel karhl, thats true, only destructor can drop it 17:31 Krock updated #2744 "Enable server when not building client …" 17:31 est31 just wondering 17:31 est31 Zeno has encouraged that PR, now he left 17:32 est31 and the commit message is still wrong. 17:32 est31 err sorry 17:32 est31 its only the PR title 17:33 est31 now better 17:33 Krock does it matter? finally, the commit counts 17:33 est31 no doesnt 17:33 est31 just thought it would be the same as the commit message 17:51 est31 what bout this: https://github.com/est31/minetest/commit/8f37af0d412bca645a33f57b57b14470ac5349a3 17:52 est31 for master 18:03 kahrl est31, I don't quite get it 18:03 est31 ? 18:03 kahrl since the external protocol is most likely implemented in an external program, with no connection to the minetest engine, how can it use the minetest API? 18:04 est31 irc_commands uses luasocket 18:06 kahrl hmm. that seems like a pretty specific usecase though 18:07 kahrl why does one input passwords via irc anyway? 18:07 est31 in order to log in via irc 18:07 est31 so that you can grant sb privs 18:08 kahrl I see 18:09 kahrl perhaps a more secure alternative would be: 1. connect to minetest server via a special init packet, 2. upon successful SRP login minetest server returns something like an oauth token, 3. you can use this token from irc to log in as that user 18:10 kahrl of course this token would be single-use, limited lifetime, to prevent replay attacks 18:11 est31 yea just very inconvenient 18:11 kahrl meh, I've been using that method for twitch IRC, it's not that bad 18:11 est31 I mean, you have to log in, join the server, copy the token, log out, and then send it over irc 18:12 kahrl it shouldn't be too hard to write a script for the irc client to do that 18:12 est31 is there an unified irc scripting language? 18:12 kahrl nope 18:13 kahrl you could do it as a shell script that copies the token to the clipboard, too 18:13 est31 yea thats what sfan5 suggested too 18:14 est31 (scripting the irc client) 18:14 est31 but perhaps we should deprecate minetest.get_password_hash then 18:14 est31 as its of no more use 18:14 kahrl isn't it already? 18:15 est31 no it isnt 18:15 kahrl I thought deprecating it was part of the move to SRP 18:15 est31 one use case, I admit, remains, which is the /setpassword command 18:16 sfan5 est31: get_password_hash can be used to do sha1 on data 18:16 est31 hacky stuff... perhaps it should be renamed then 18:17 * est31 puts the callback onto a branch where it can rot in piece 18:18 kahrl would it be possible to make /setpassword set a SRP, not a legacy password? 18:18 est31 yes it would 18:19 est31 but then people can't connect anymore with older clients 18:19 kahrl that's true 18:19 est31 also /setpassword doesnt benefit from SRP 18:19 est31 the server (even its owner) already know the password 18:20 sfan5 that's exactly what setpassword is for 18:20 est31 yes 18:21 est31 ok there are some benefits, like the removed replayability 18:21 est31 err no 18:21 est31 sorry 18:22 est31 replayability is already removed, even for "old" passwords 18:22 sfan5 hm 18:22 est31 "old" passwords get SRP too with the new protocol 18:22 sfan5 does changing the password remove the benefits of srp? 18:22 est31 with the hashes used as "passwords" 18:23 est31 sfan5, a player changing their password, or a server owner changing it? 18:23 sfan5 player 18:24 est31 a player changing or setting their password relies on some trust yes 18:25 est31 when they change, they send a new verifier, unencrypted 18:25 sfan5 what does that mean? 18:25 sfan5 like 18:25 sfan5 what can the server owner do w/ that verifier? 18:25 est31 they can brute force 18:26 est31 not even the best protocol can save you from that 18:26 est31 its a basic feature of login protocols :) 18:26 Calinou brute force prevention in core? 18:26 sfan5 so nothing more than they can do anyway? 18:26 est31 no 18:27 est31 a MITM can't even impersonate the user to the server with that verifier 18:27 est31 AFAIK 18:33 est31 yes 18:34 est31 that website just sais so: https://www.logintc.com/blog/2013-12-06-secure-remote-password.html 18:39 est31 also I guess it would be great for the irc_commands mod to have a user mask based authentication scheme 18:39 est31 so you won't need passwords at all 18:39 est31 at least not for minetest 19:25 est31 btw hmmmm whats the proper way to make the emerge thread make mapgen some day for some block? 19:26 est31 or just some way 19:32 hmmmm someday? 19:33 hmmmm oh some way 19:33 hmmmm if you really want things to be done synchronously i recommend bypassing the emergethread entirely 19:33 est31 I dont need it to be blocking, I can restart minetest then it should be generated 19:33 hmmmm huh? 19:33 hmmmm well 19:34 hmmmm first of all you have to have an active EmergeManager 19:34 hmmmm emerge->initMapgens() 19:34 hmmmm otherwise nothing will get generated, only loaded 19:35 hmmmm then, emerge->startThreads() 19:36 hmmmm then, emerge->enqueueBlockEmerge(peer_id, blockpos, true); 19:36 est31 why do I need a peer id? 19:36 est31 ah I gues I know 19:37 est31 the data gets sent to the client with the peer id 19:37 hmmmm the emerge thread marks blocks it just created as unsent for that peer 19:37 hmmmm so the next time there's a server step, it'll send those unsent blocks 19:37 est31 very nice, enqueueBlockEmerge is the method I've needed I think, thanks 20:17 BIrdyMeows Hello everyone :3 20:18 ElectronLibre BIrdyMeows.. Read the rules... 20:18 BIrdyMeows i have and? 20:19 ElectronLibre "Chit-chat goes to #minetest." 20:21 hmmmm I wonder if BIrdyMeows was about to engage in hardcore minetest core development discussion 20:21 hmmmm ..or if he was going to make a variety of emoticons and greetings 20:21 ElectronLibre I don't think so. I know her, which is why I helped her understand that this was not the appriopriate channel. 20:22 ElectronLibre I think, variety of emoticons and greetings.. 20:23 Calinou she was going to elaborate on deterministic, polymorphic Perlin noise functions with type inference. 20:49 jordan4ibanez if you do on_construct and then add the node you're constructing next to it, it'll do it infinitely and cause a segfault 20:52 hmmmm hmm 20:52 kahrl well... user error I'd say :P 20:52 hmmmm it's working as intended 20:52 hmmmm how can you possibly counter that? 20:52 hmmmm without restricting what users can do 22:11 celeron55 hmmmm: i just stumbled upon this: http://c55.me/random/2011-05/uloste4.png 22:12 celeron55 that's a version of minetest that clearly waited for all block neighbors to have stone before adding dirt and grass and trees to them 22:15 hmmmm cool 22:17 celeron55 well actually looks like the code didn't wait for anything 22:18 celeron55 it just generated stone to places that had CONTENT_IGNORE at a full 3x3x3 volume around the block being generated 22:23 twoelk the landscape looks like sporting large scars from something slashing at it, reminds of early mappainter cavegen 22:56 RealBadAngel https://github.com/minetest/minetest/pull/2755 22:57 RealBadAngel any comments? 22:59 RealBadAngel hmmmm, kahrl?