Time Nick Message 01:43 VanessaE interesting glitch I noticed earlier: 01:44 VanessaE turn on 3rd person mode, first-press of F7 (where you're looking from behind your character). place a torch. now try to break it. 01:44 VanessaE I was unable to actually highlight it to punch it. 07:37 ShadowNinja ~tell sapier Text doesn't show up in text fields until you select a different field in your latest Android release build. 07:37 ShadowBot ShadowNinja: O.K. 13:25 Megaf Morning all 13:26 Megaf celeron55: Hi there, I think you might know this already, I just came across this and I found it rather interesting. https://cmr.github.io/lanes/ 13:26 Megaf I wonder how hard would it be to run each mod on it's own lua thread 13:27 hmmmm very hard and nonsensical 13:28 hmmmm we've had plenty of discussions about this before 13:28 Megaf really? 13:28 hmmmm if you want to make adding more threads a viable thing, you should start with making Map non-locking 13:29 Megaf I see quite a lot of advantages on running each mod on it's own thread, the biggest one is security and stability, a mod could crash without brining the server down 13:29 hmmmm non-locking is awesome, non-waiting is even better but may not be possible 13:31 Megaf hmmmm: non-locking map? What would that be and how would it work? 13:32 hmmmm accesses to Map need to be synchronized. i suggest you look into it more and you'll see that it's perhaps minetest's biggest problem as of right now. 13:34 Megaf I quite like synchronized things, they are more predictable and stable 13:34 Megaf The real limitation for that is hardware 13:39 hmmmm so then you like minetest's problems of threading effectively being useless 13:40 * hmmmm claps 13:43 Megaf lol, I just done my first multithreaded software and my first lua test at the same time 13:43 Megaf $ ./test.lua 13:43 Megaf 2 4 13:53 sfan5 this is not releavant for minetest development 14:03 hmmmm wow 14:03 hmmmm well that was a complete waste 14:03 hmmmm i get to work, there are absolutely no cars in the parking lot. building is locked down. this "forced" holiday is taking up 8 hours of paid time off don't want to spend it. 14:04 hmmmm so i guess today is a minetest day 14:06 * Megaf has to work 14:06 Megaf at least I can use IRC here 14:07 hmmmm i used to be able to irc from work but i got it banned 14:07 hmmmm i got overconfident and started SCPing like GBs of music from this computer 14:07 Megaf I use Quassel now a days, in Quassel you can run the core at home and the client at your workplace, and use the port 443 for example 14:08 Megaf evereything is done over SSL 14:08 Megaf actually, here I'm usin a SSH SocksV5 tunnel 14:08 Megaf then I connect the client to core using the tunnel 14:09 Megaf and so my web browsing 14:09 * hmmmm looks at logs 14:09 hmmmm minetest has 3rd person mode now!? 14:09 hmmmm sapier if (id == (u32)-1) { ?? ... not the bug anyway 14:10 hmmmm that is valid 14:10 hmmmm typecasting -1 to a unsigned integer sets all the bits 14:11 hmmmm the way the C standard describes integer storage types basically locks it into the familiar binary format so you'll never see a case where that doesn't work 14:12 hmmmm and (unsigned int)-1 is used in POSIX standards as well, so it's kosher 14:19 hmmmm hmm no zat 14:20 hmmmm well whatever. zat: the proper way to do lua callbacks is to store it in a field in the global "minetest" 14:21 hmmmm so what you want to do is make a wrapper for the decoration registering function in Lua that adds the callback to some global such as minetest.deco_callbacks[id] 14:22 hmmmm you can find good examples of how to perform lua callbacks in src/script/cpp_api/s_env.cpp 14:22 hmmmm id should be just an incrementing count 14:24 hmmmm sapier I always wonder for what reason someone whats to add the * to the variable ... to me it's a modifyer to the type ... does anyone have an explanation for this quite insane rule? 14:24 sfan5 hmmmm: char* a, b create a char * called a and a char called b 14:24 hmmmm it's the K&R way, and it also prevents mistakes like what sfan5 mentioned, but the pointer syntax in C is just stupid IMO 14:25 hmmmm I am actually using my spare time at work to come up with a nu-C language that's basically C, but incompatible and superior 14:26 sfan5 please make type comparison work like in PHP 14:26 hmmmm lol. 14:26 hmmmm anyway pointers are @ making the unary * operator free for something else 14:26 sfan5 do you plan to add support for it to gcc or clang or will it get converted to c for compliation? 14:27 hmmmm I suppose it'd be a clang frontend 14:27 hmmmm i haven't made it that far yet however 14:27 sfan5 anyway pointers are @ 14:27 hmmmm right now it's just a huge collection of ideas that are being sorted out 14:27 sfan5 char @a, @b; like that? 14:28 hmmmm yeah. 14:28 hmmmm however type modifiers ALWAYS come before the type name 14:28 sfan5 I was about to say that it would not make it better 14:28 hmmmm so an array of pointers would be like char []@foo 14:29 hmmmm you wouldn't be able to do char @foo, not_a_string; however 14:29 sfan5 char is the type name so it would be @char a, b; wouldn't it? 14:29 hmmmm you'd only be able to combine variable declarations with types that are completely alike 14:30 hmmmm so char @foo, bar; would make foo and bar both char pointers 14:30 sfan5 writing it like that would be confusing then 14:30 hmmmm maybe there needs to be some syntactical separation to make that more obvious 14:30 sfan5 char@ foo, var; would be less confusing in that case 14:30 hmmmm it's only confusing because you're used to C/C++ 14:31 hmmmm I think Ada does what I'm talking about 14:31 hmmmm where the type names are syntactically separated from the variable names 14:31 sfan5 maybe, maybe not; I think it does not make sense for a modifier to affect something that is seperated (by a ,) 14:32 hmmmm you call it a modifier, but I say that a pointer is a different fundamental type 14:32 sfan5 like this? 14:32 sfan5 char@ a, b; 14:33 hmmmm that'd be better, yeah. 14:33 sfan5 type variable name, variable name; 14:33 hmmmm Go tries to counter this by putting type specifiers after the identifier name 14:34 hmmmm e.g. "type foo char" 14:34 hmmmm I don't like it though 14:34 hmmmm although it does make it less lexically ambiguous 14:34 hmmmm easier parsing means faster parsing 15:14 CraigyDavi Is it intentional that the black background box upon death has been removed? https://cdn.mediacru.sh/SEtQcbuRa4TW.png 15:37 rubenwardy Any comments on Add item eat callback ( https://github.com/minetest/minetest/pull/1318 ) ? 18:23 Piggybear87 sfan5, I replied to you in the forum with the lspci you asked for. PS this is Tedypig. 18:23 sapier VanessaE I'm sorry but adding compat code to generic stats would result in adding more code then generic stats are in total. Imho that feature is only worth adding in case we already drop compat for something more valuable. 19:14 ShadowNinja sapier: https://github.com/minetest/minetest/issues/1320#issuecomment-44211609 19:15 sapier what exactly happens there I can't reproduce the issue 19:15 sapier how do you reproduce it? 19:15 LemonLake Just look at a nodebox? 19:16 sapier doesn't happen to me 19:17 sapier so what's the difference? 19:20 ShadowNinja sapier: Look at a nodebox and the selection box stops working on all nodes. 19:22 sapier nope ... not for me 19:23 ShadowNinja sapier: I'm using Irrlicht 1.8.1 BTW. 19:24 sapier ok I'm on 1.7 19:26 sapier that comit changes the draw order but that's not supposed to cause a bug like that 19:28 PenguinDad 1.8.1 works just fine here 19:29 sapier ok I can reproduce it on one of vanessae's servers ... so why is this relate do the game running?? 19:41 ShadowNinja hmmmm: v6 seems to ignore the 'flat' MG flag. 19:54 hmmmm shadowninja, the last time someone said that it's because they didn't turn off caves as well 19:54 hmmmm caves are responsible for some terrain deformation 20:10 sapier I think I found the selectionbox issue 20:13 sapier could someone having the selection box issue verify #1336 fixes it? 20:13 ShadowBot https://github.com/minetest/minetest/issues/1336 -- Fix missplaced selection box in some situations by sapier 20:24 ShadowNinja hmmmm: I meant that it ignores nocaves. My map_meta seems to have caves set though... 20:28 Megaf sapier: I've been wating for that fix 20:28 sapier does it work? 20:29 Megaf I'm at my workplace, I will tell you in one hour when I get home 20:29 sapier ok 20:32 Megaf Can someone remind me how I can apply this patch to my minetest here? https://github.com/minetest/minetest/issues/1336 20:33 sapier either add my git repository as remote and do a cherry pick or just clone it to a separate folder and checkout that branch 20:33 Megaf ShadowNinja: Are you dropping support for netbsd? 20:33 sapier or just edit that file ;-) 20:33 hmmmm ahh 20:33 hmmmm I'll have to check it out 20:33 ShadowNinja Megaf: No? 20:33 Megaf https://github.com/minetest/minetest/commit/31f9912e9e9f0d970725788af27f3a731415a57d 20:33 hmmmm ShadowNinja, are you sure that's a brand new map that you created with nocaves specified? 20:34 CraigyDavi Is it intentional that the black background box upon death has been removed? https://cdn.mediacru.sh/SEtQcbuRa4TW.png 20:34 CraigyDavi ^anyone? 20:34 ShadowNinja hmmmm: No. It apparently had caves set after your rework, I think I only set it in minetest.conf. But when I looked at the MGv6 code it DID seem to be missing a caves check, wven if caves were enabled. 20:35 ShadowNinja even* 20:35 sapier not really CraigyDavi wonder how it's lost 20:36 ShadowNinja Megaf: That fixed support for BSD threadnames. It includes NetBSD. 20:36 ShadowNinja fixes* 20:36 ShadowNinja +#elif defined(__NetBSD__) 20:36 hmmmm ShadowNinja, https://github.com/minetest/minetest/blob/master/src/mapgen_v6.cpp#L474 20:39 CraigyDavi Ok I created an issue on the issues section 20:40 ShadowNinja hmmmm: O.K. I guess not... 20:40 hmmmm well 20:40 hmmmm try to reproduce the problem 20:41 hmmmm put nocaves in minetest.conf, then try to create a new world, see if it has caves with mgv6, then go and check the map_meta 20:41 hmmmm see if that has nocaves as expected 20:43 ShadowNinja hmmmm: It works. Sorry for the false alarm. 20:43 hmmmm np 20:45 ShadowNinja sapier: Your fix works. s/missplaced/misplaced/ though. 20:48 sapier better? 21:05 ShadowNinja Comments on #1318? 21:05 ShadowBot https://github.com/minetest/minetest/issues/1318 -- Add item eat callback by rubenwardy 21:05 ShadowNinja #1149 too. 21:05 ShadowBot https://github.com/minetest/minetest/issues/1149 -- Add AppData file by davidgumberg 21:17 VanessaE sapier: it's weird, I just can't seem to reproduce that selection box glitch in singleplayer. 21:18 sapier if that fix helps it's some missing irrlicht scene config ... but I don't know it good enough 21:19 sapier my fix basicaly restores something beeing lost while moving around code 21:19 VanessaE I haven't tried any fixes yet 21:19 VanessaE (in fact I didn't even know it was looking at nodeboxes that was the trigger, but I *still* can't reproduce it in singleplayer) 21:20 sapier I didn't have to look on nodeboxes to cause it 21:20 sapier at least on your server 21:20 VanessaE oh ok 21:20 sapier I couldn't reproduce it on any other server 21:21 VanessaE it's very very strange 21:21 VanessaE my servers are ..*shrug* bog-standard minetest engine 21:22 sapier not really most likely it depends somehow on what is drawn in which order thus setting a defined state prior drawing selectionbox is quite reasonable to be a fix ... at least that's what I figured out .... RBA most likely could explain better 21:24 VanessaE did you notice the bit I mentioned where the select box gets drawn off into the distance/ 21:24 VanessaE ? 21:24 VanessaE like, off to the side, and of the wrong size 21:25 sapier yes that was the hint I needed to find it 21:25 VanessaE oh ok 21:25 VanessaE good :) 21:26 ShadowNinja Comments on #1277? I split it into two commits so that you can see what I actually changed. 21:26 ShadowBot https://github.com/minetest/minetest/issues/1277 -- Update to Lua 5.2 by ShadowNinja 21:27 sapier ShadowNinja: for what I understand 1277 has that one issue we can't get full 5.2 support with luajit 21:28 ShadowNinja sapier: LuaJIT supports most of Lua 5.2's features. What it doesn't support we don't need to use. 21:29 ShadowNinja LuaJIT supports the strip argument to string.dump too, which is Lua 5.3. 21:29 sapier that doesn't really change the problem ... right now we can say Mintest supports lua 5.1 only 21:29 sapier everything else is modders fault 21:29 ShadowNinja If we keep Lua at 5.1 then the features that LuaJIT provides will be unusable because it won't be provided by both Lua and LuaJIT. 21:30 sapier later it will be "5.2 except of x, y, z, ...." 21:30 sapier imho that's more then ugly 21:30 ShadowNinja If we update to Lua 5.2 we will support "most Lua 5.2 features" until LuaJIT gets Lua5.2 support. 21:31 sapier I'd prefere 5.1 to "most of 5.2" 21:31 RealBadAngel not better include luajit and drop lua at all? 21:31 sapier no 21:31 ShadowNinja Better than a two year old copy of Lua. 21:31 sapier luajit is crap if you want to debug 21:32 RealBadAngel i never needed to debug lua at all 21:32 sapier well I did believe me you can't debug with luajit enabled 21:33 ShadowNinja Lua is in some ways better than LuaJIT, but LuaJIT's speed makes up for it's shortcomings. 21:33 sapier Speed is not an option if you break debugging 21:34 sapier you can run servers on luajit but for modders a non luajit version most likely is better option 21:34 RealBadAngel having mods two times as fast is not an option at all, its a must have 21:34 ShadowNinja sapier: Well, right now I'm discussing updating Lua, not removing it. 21:35 sapier removing is what RealBadAngel suggested as that'd be even worse then broken 5.2 support that's way more important to prevent ;-) 21:35 ShadowNinja RealBadAngel: Lua is quite usable on it's own. It's easier to debug and has some other advantages due to being less optimized. 21:35 sapier if you add 5.2 that's gonna cause mintests lua support to be broken on luajit 21:36 ShadowNinja sapier: How so? 21:36 Megaf sapier; the nodebox thing seems to be fixed 21:36 sapier missing 5.2 features is what is defined as "broken" 21:37 sapier I'm against merging 5.2 support until we can provide full 5.2 support for luajit too 21:38 sapier as always that's my opinion only there might be enough other developers to still merge it 21:38 VanessaE who says you have to advertise that it has partial 5.2 support? 21:38 VanessaE does it break 5.1 compat? 21:39 VanessaE put Lua 5.2 in, recommend to mod authors not to use 5.2 features for now. 21:39 VanessaE simple. 21:41 RealBadAngel if recomending not to use, why change to it in the first place? 21:41 VanessaE sapier, pull #1336 fixes the selection box problem for me also. 21:41 sapier adding new code always means adding new bugs 21:41 ShadowBot https://github.com/minetest/minetest/issues/1336 -- Fix missplaced selection box in some situations by sapier 21:42 VanessaE or it seems to from a quick test. 21:42 sapier ok as there are two ppl and me where it's working I'm merging it now 21:42 VanessaE certainly works properly in the problem area anyway 21:42 VanessaE will this cause a performance regression? 21:43 VanessaE (it's hard for me to tell with all this map loading/mesh gen activity right now) 21:43 sapier no it's been there prior my drawscene split too so it's at worst back to original speed 21:44 VanessaE ok 21:44 ShadowNinja sapier: As far as I can tell the only important thing that LuaJIT is missing is _ENV. It supports the env argument to load though. 21:44 VanessaE it seems to work fine with that patch. 21:45 sapier ShadowNinja: that's exactly what I mean, we don't even know exactly what is missing 21:45 sapier we just know it's not complete 21:45 ShadowNinja sapier: And there's a patch for that: http://www.freelists.org/post/luajit/A-patch-for-ENV-support 21:46 sapier problem is still almost all luajit implementations of distributions are built in lua 5.1 support mode so there's more missing 21:52 VanessaE that's a problem. 21:52 VanessaE and luajit team doesn't like the idea of bundling it with a project either. 21:52 VanessaE (personally I don't care if they like it or not...) 22:08 ShadowNinja sapier: Well, not my distro. I found a list of unimplemented 5.2 features and all but one of them have been implemented by now (NUL ("\0") in regex, which can be done with "%z".) 22:09 ShadowNinja ^We can fake that in 51compat.lua too, but that would be a bit ugly, and that feature is rarely used. 22:09 sapier "not my distro" ... 22:10 sapier for what I remember it's 2:2 right now ... celeron and me against vanessae and you pro 22:11 sapier the main problem is "we are not completely compatible to 5.2" ... forcing 5.2 back to 5.1 as solution isn't a reason to add 5.2 support at all ;-) 22:12 ShadowNinja The only major feature missing is _ENV. 22:13 sapier for your distribution, for mine it's way more 22:14 sapier for what I remember you're on arch, true? 22:15 ShadowNinja sapier: Correct. The missing features aren't all that major though, they mostly deal with metatable checks. 22:16 sapier it doesn't matter, no one wants to read and interpret 5 different locations one beeing package file of his distribution to find out what's supported and what not 22:16 VanessaE inb4 RBA says he plans to use those features :P 22:18 sapier if we add this I already know we cause a lot of discussions "update your distribution" .. "works for me" ... things like that 22:20 sapier and arch is by no means a benchmark about supported features, to me arch is bleeding edge code I'd only touch with gloves and for sure wont use on a development machine or server ;-) 22:23 Megaf sapier; I totally agree 22:24 VanessaE the only way to solve that is to bundle Luajit with minetest. 22:24 sapier which is both against suggestion of luajit and our own intention to add as less dependencys as possible 22:25 VanessaE I know. 22:25 VanessaE but we did/do the same with Lua 22:25 VanessaE and for the same reasons, right? 22:25 sapier what do we win on adding lua 5.2 support? is there anything worth it? 22:25 sapier even on adding luajit we only get close to 5.2 support but still would have broken 5.2 support 22:26 VanessaE I can't say because I don't know 5.2's feature set (except it has some bitwise operations), but having out-of-date Lua is guaranteed to be a problem sooner or later. 22:26 sapier We're not a commercial company which needs marketing claims like "Brand new Minteste now with lua 5.2 support" 22:27 VanessaE wanna bet? 22:27 VanessaE commercial, no 22:27 VanessaE but marketing is always important if you want something to be used. 22:27 VanessaE I'm not saying to go start running TV commercials here 22:27 sapier important enough to get bad moding experience due to issues caused by half implemented 5.2 features? 22:28 VanessaE I'm just saying that people will begin to notice "fuck, this only has lua 5.1? what a piece of shit" eventually. 22:28 sapier of course they will ... and I don't wanna tell "stay at 5.1 forever" ... I just believe it's to early to switch 22:31 sapier And unless someone provides good new reason to switch I don't think I'm gonna change my mind about it. And to prevent false assumptions I don't expect my vote to be honored more then any other core devs vote 22:44 ShadowNinja sapier: #1337 22:44 ShadowBot https://github.com/minetest/minetest/issues/1337 -- Death screen missing the background box 22:44 sapier Yes already seen it 22:44 sapier I wonder what makes the box around it invisible ... guess it's another layering issue 22:49 kaeza dat issue number 22:50 sapier well quite a good fit ;-) 23:35 LemonLake I know it's not quite core-dev related, but can I get some help with something, please? 23:36 LemonLake Basically, I'm running Linux Mint 16 in a VM. I currently have 3D acceleration enabled. 23:36 RealBadAngel well, luajit has things that regular lua has not 23:36 LemonLake When it's enabled, it looks fine, runs fine, etc. As soon as I start minetest, though, this happens: https://www.youtube.com/watch?v=JlM2A7nyS3U 23:36 RealBadAngel like ffi 23:36 RealBadAngel http://luajit.org/ext_ffi.html 23:36 LemonLake Minetest works perfect when 3D accel is disabled, however. The only problem with it being disabled is that the whole system runs at ~3fps 23:36 RealBadAngel and bitwise operations comes with luajit not lua 23:40 RealBadAngel you say in a Virtual Machine? 23:40 LemonLake Yes 23:40 RealBadAngel lol 23:40 LemonLake My actual machine is dead. 23:41 RealBadAngel so you do have also vrtual 3d acceleration 23:41 LemonLake On my host bios? 23:41 RealBadAngel cmon, for christ sake, vm are not for gaming 23:41 LemonLake I'm aware. 23:42 LemonLake I'm not playing games on them, though. 23:42 RealBadAngel they dont have real gpu's 23:42 LemonLake But I've got this problem that nobody wants to help me with. 23:42 LemonLake And I can never compile on Windows. 23:42 LemonLake So I'm compiling on here to debug it 23:42 RealBadAngel just install normal OS and play 23:42 LemonLake LemonLake> I'm not playing games on them, though. 23:42 RealBadAngel or compile 23:43 RealBadAngel have you ever heard of dual boot? 23:43 LemonLake Yes and it's something that I'm not going to do. 23:43 RealBadAngel you dont have to, installer will do that for you 23:44 LemonLake What installer? 23:44 RealBadAngel any linux installer 23:44 LemonLake That's a thing? 23:44 RealBadAngel installing linux next to windows is as easy as a few clicks 23:44 LemonLake Do I actually have to run the ISO, though? 23:44 LemonLake Like on a disc 23:45 LemonLake Or is there an actual installer? 23:45 RealBadAngel download distro of your choice, burn the dvd (or cd) 23:45 RealBadAngel boot from it 23:45 RealBadAngel and choose installing next to windows 23:46 RealBadAngel i suggest pickin Xubuntu for example 23:46 LemonLake I previously didn't want to do that. 23:46 LemonLake Buuut 23:46 LemonLake I'm gonna do it, one question first though 23:46 LemonLake How does the dual boot ACTUALLY work? 23:46 LemonLake Like, how and when do I switch OS? 23:46 RealBadAngel when you turn on your machine you have simple menu to pick OS to load 23:46 LemonLake Ah. 23:47 LemonLake So I can't just come home, turn on my machine then go down to eat, I have to actually pick before it starts up. 23:47 LemonLake I'm fine with that though, got another 1.4tb to spare. 23:47 LemonLake Just gonna find a blank DVD to burn 23:48 LemonLake Gotta be honest, I'm sh*t scared. 23:48 RealBadAngel why so? installing linux is way easier than installing windows actually 23:49 RealBadAngel at least modern linuxes 23:49 LemonLake Lately I've been in the mindset where if I do one thing wrong in stuff like OS settings or I'm sure you can guess, the entire PC is a brick 23:49 LemonLake But balls to it. Gotta take risks. Here goes, burning now. 23:49 LemonLake Way quicker, too. 23:49 LemonLake Oh crap that's loud 23:50 RealBadAngel after you burn dvd, and boot from it 23:50 RealBadAngel you may choose live 23:50 LemonLake Or install 23:50 RealBadAngel which means you can try the distro without hdd even 23:51 RealBadAngel = no changes made to your disk 23:51 LemonLake Should I partition the disk, actually? 23:51 LemonLake I have a 2TB SSD (I think it's an SSD) split into two drives on Windows 23:51 RealBadAngel installer will do that for you 23:51 LemonLake Not sure how it works 23:51 LemonLake Ah ok 23:51 RealBadAngel you have just to pick correct option 23:51 RealBadAngel just read all the messages 23:52 LemonLake That's why I'm logging in on android :P 23:53 LemonLake Burned. 23:53 LemonLake That disc eject scared the crap out of me 23:54 LemonLake Okay. Let's do this! 23:55 lemonlake_ I hope the boot order is correct... 23:56 lemonlake_ Okay, it keeps launching windows 23:56 RealBadAngel so change it in bios 23:59 lemonlake_ No key opens bios