Time Nick Message 02:10 jordan4ibanez !tell 02:10 jordan4ibanez !hel 02:10 jordan4ibanez !help 02:10 jordan4ibanez Oh wrong channel... 02:17 VanessaE RealBadAngel: caching cracks meshes? 02:17 VanessaE how many meshes in total are we talking about here? 09:11 rubenwardy celeron55, broken: http://www.minetest.net/donate 09:18 celeron55_ lol how is that even possible 09:19 rubenwardy I think custom HTML was disabled 09:19 rubenwardy So it displays as plaintext 09:19 celeron55_ yes but how did it get disabled, i haven't done that 09:19 rubenwardy Neutrinos? 09:20 celeron55_ i'm looking at the dokuwiki template that is there, and it specifically has the line if ($ID == 'donate') {$conf['htmlok'] = true;} 09:20 celeron55_ this makes zero sense 09:24 rubenwardy Weird things happen: int doom[1]; for(int x = 0; true; x++) { doom[x] = NULL; } 09:24 rubenwardy x goes 1, 2, 3 and then loops. o_O 09:24 celeron55_ eh what 09:25 celeron55_ i re-saved the page with a whitespace modification 09:25 celeron55_ now it works 09:25 rubenwardy Well done 09:25 celeron55_ i won't even try to guess what has happened 11:42 Zeno` ok, fixed the broken inventory 11:42 Zeno` well the part that annoyed me anyway 11:42 Zeno` I don't know if the rest is intended behaviour or not 11:42 Zeno` and "annoy" == "clearly broken" 11:52 Zeno` https://github.com/minetest/minetest/pull/1704 11:53 Zeno` maybe reverting f6321e2 is the best option :( 11:53 Zeno` it seems the PR is totally broken 11:54 Zeno` Mine doesn't fix it either (only fixes it for the craft grid) 12:03 Zeno` kahrl 12:11 Zeno` anyway this bug is a blocker 12:11 Zeno` it's completely broken and I don't have time to fix a bug introduced to fix a non-existent problem :( 12:12 Zeno` plus I cannot understand the new logic, heh 12:29 kahrl I can't reproduce the problem 12:29 kahrl Zeno`: ^ 12:30 Zeno` you can't? 12:30 Zeno` place a stack of 99 things in the craft grid 12:30 Zeno` now divide it into two 12:31 Zeno` using RMB 12:31 kahrl works fine 12:31 Amaz So right click the stack, and then click elsewhere? 12:31 Zeno` no, don't click elsewhere just hover over an empty slot 12:31 Zeno` and don't release the RMB 12:31 kahrl well, release the RMB then? 12:31 Zeno` no 12:32 Zeno` just right click on a stack, move mouse over empty slots 12:32 Zeno` without releasing 12:32 kahrl dropping items when RMB is held is intended 12:33 Amaz But what need is there to hold the RMB down? 12:33 Zeno` well it never worked that way before and it's now impossible to craft, say, a locked chest 12:33 Zeno` it's not impossible just very, very annoying 12:33 kahrl I would say it's easier now 12:33 Amaz ↑ 12:33 Zeno` no it's not because say you move over the steel ingots in the middle 12:33 Zeno` you pick them up! 12:34 kahrl ok, I don't do that but I agree that part is annoying 12:34 Zeno` they should stay in the middle not be replaced by 1 wood 12:35 kahrl for now you could place the wood first and then the steel 12:35 Zeno` halving is annoying as well for that same reason (when there is only 1 spare slot in the grid) 12:35 Zeno` it should not place "one" item automatically 12:35 Zeno` just because you move the mouse 12:35 kahrl it doesn't, just don't hold the RMB down 12:36 Zeno` most players do not play like that 12:36 Zeno` they hold down RMB, drag to where they want the half and then release 12:36 kahrl have you done a study? 12:37 Zeno` how can moving an item in the grid (to your hand, and swapping it with what's in your hand) be user friendly? 12:38 Zeno` just because you happen to move the mouse over that item as you're transferring half a stack from the bottom right to the top left? 12:39 Zeno` that's like right-clicking and dragging a folder to a new window (in Windows) but if your mouse happens to cross another window it automatically drops into the non-destination 12:39 Zeno` I don't need to do a study to work out that that's silly :) 12:40 Zeno` I don't expect moving the mouse to drop breadcrumbs in all the windows it happens to cross as I move the cursor 12:40 kahrl I agree f6321e2 is not optimal but reverting it doesn't make things better 12:40 Zeno` it makes things work 12:41 Zeno` anyway, if you like then release it... I can guarantee 100% that you will get many reports of it being broken but *shrug* 12:41 Zeno` at least then you'll believe me ;) 12:44 Zeno` put more than one item type in the grid 12:44 Zeno` and try and half another item as you move the mouse across the extra item 12:44 Zeno` halve* 12:45 Zeno` it's an exercise in frustration 12:46 Zeno` you have to carefully make sure the mouse pointer does not cross the item that's different to the one in your hand 12:47 kahrl or just release the RMB :P 12:52 Zeno` no that doesn't work 12:52 Zeno` *sigh* 12:53 kahrl hm? it does for me 12:53 Zeno` honestly, people do not expect things to happen without them somehow initiating that action 12:54 Zeno` when you copy a file using drag and drop do you press the RMB, release, move mouse and... ?? 12:54 Zeno` and then what to place? 12:55 kahrl minetest is not a file manager 12:55 Zeno` never said it was 12:55 Zeno` but people have expectations on how things work 12:56 Zeno` this new behaviour is baffling because I have *never* seen anything similar 12:57 Zeno` I can't even think of a metaphor 12:57 Zeno` you pick up something to move, you hold it 12:57 Zeno` and then release when it's there 12:58 Zeno` that's logical 12:58 Zeno` you don't pick up something, move, and expect it to fall out of your hands 12:58 Zeno` before you reach where you want to put it 12:59 Zeno` even worse the object falls out of your hands and suddenly you pick up what it fell onto 13:16 Zeno` isn't the game supposed to be for the players? 13:40 kaeza any news on weblate/whatever? 13:45 Zeno` WHOA! 13:45 Zeno` Hold the horses! 13:45 Zeno` https://github.com/minetest/minetest/blob/master/src/debug.h#L75 13:45 Zeno` that does that mean? 13:46 Zeno` I just checked Leveldb and they use assert() correctly 13:55 VanessaE OH nonononono 13:56 VanessaE that is NOT gonna work 13:56 VanessaE kahrl: you must revert that shit NOW 13:56 VanessaE (or some other dev) 13:57 VanessaE I just fetched and built latest just to see what's going on here. 13:57 VanessaE right click, hold and drag to move half a stack, the intent of that action is to drop that half at its destination when you release the button. the patch that was committed does NOT work as intended and causes a severe playability regression 13:58 VanessaE and it's not just the crafting grid, it leaves a trail of single items in every inventory slot it crosses. 14:01 VanessaE well whether that was the intent or not it's a huge playability problem 14:01 VanessaE no one, and I mean NO ONE expects click-and-drag to do anything but drag-and-drop, and we even use that in the creative inventory. 14:01 VanessaE this is inconsistent and broken in every way that matters to the average user. 14:01 VanessaE revert it or fix it. 14:05 DusXMT I think dragging out a half is neat, but not that it leaves items in the grid it crosses, it's a unique approach 14:06 PenguinDad Am I the only one who's got no problem with this? 14:06 kaeza MC does leave a trail, but only if you right click on a slot *after* grabbing a stack 14:06 kaeza (right click and drag) 14:07 kaeza maybe that's what the patch intended to do? 14:08 DusXMT Or maybe I'm not understanding this (that's quite possible, as for example, it took me a good 10 minutes to realize RMB is the right mouse button) 14:08 VanessaE PenguinDad: right click and drag a stack and move it around over your inventory 14:08 VanessaE just move it randomly 14:08 VanessaE you will suddenly have your inventory totally randomized and G*d only knows what will be in your hand after the move 14:10 kaeza http://youtu.be/quFMqmHJl2o?t=4m30s 14:11 VanessaE kaeza: right, and that doesn't work here 14:11 VanessaE not even close, and I even tried doing exactly that 14:11 VanessaE stacks don't re-combine with this patch 14:11 PenguinDad VanessaE: I never move stacks while holding LMB or RMB 14:11 VanessaE they just swap 14:11 VanessaE PenguinDad: this is a new "feature" that is horribly busted. 14:14 VanessaE ShadowNinja: *poke* you committed this 14:18 kaeza comments? https://github.com/kaeza/minetest/commit/7b0a31ce2134cdc01b1b57423eafabedc50bb811 14:18 kaeza needed since builtin was reorganized into subdirectories, so builtin/*.lua does not suffice anymore 14:19 kaeza it seems updatepo.sh was not run in a looooong time 15:12 zat wat 16:03 zat someone knows how to change vm->m_data[vi].getContent() != c_place_on so groups are also taken in consideration? 16:11 hmmmm you want to know how to tell if a node is in a certain group or not? 16:11 hmmmm from within the core 16:13 zat hmmmm: right 16:14 zat so I can specify groups for decorations place_on. 16:15 hmmmm oh, you're modifying decorations... it would be nice if you made a specific option to check for a group or a certain node 16:15 hmmmm checking for presence in a node group is a much more expensive operation 16:17 zat hmmmm: I dont want to have to register many decorations for biomes with different surfaces 16:18 hmmmm check out INodeDefManager::getIds() 16:32 zat hmmmm: you are in charge of v7 or mapgens in general? 16:32 hmmmm anything that has to do with map generation 16:48 zat hmmmm: okay I did it 16:49 zat now groups are taken in consideration 16:53 zat hmmmm: what do you think, is this change commitable? 16:54 hmmmm i don't know unless i see the change 16:55 zat I'll find how to get a patch file 17:12 ShadowNinja kaeza: Have you tested your patch and made sure the other projects in src (eg, SQLite3) don't mess it up? 17:13 VanessaE ShadowNinja: please see the above ...discussion about the right-click-drag commit ... 17:14 ShadowNinja VanessaE: Seen, doesn't seem like a bug to me. Having a seperate rightclick action necessary to activate that could be done though. 17:14 VanessaE ShadowNinja: try it out yourself 17:14 VanessaE it's busted. bad. 17:15 kaeza ShadowNinja, it doesn't, but for some reason it's dropping some texts 17:15 ShadowNinja VanessaE: Seems to work fine for me. 17:16 VanessaE it doesn't. 17:16 VanessaE pick up half a stack but hold the button. 17:16 VanessaE drag. 17:16 ShadowNinja I think that you usually don't want to swap stacks with that though, so that could also be changed. 17:16 VanessaE watch it leave a trail everywhere and randomly swap with what's in your pointer as you drag 17:16 VanessaE this is BAD. 17:16 VanessaE this WILL generate complaints. 17:17 kaeza eh scratch that; the texts were disabled before 17:17 ShadowNinja VanessaE: I don't have a problem with that, but I'll agree to the two above changes. 17:17 VanessaE ok. 17:18 kaeza ShadowNinja, results of run are here: https://github.com/kaeza/minetest/commit/19f27f41bbd69c341627a76042229318e54e4d86 17:18 ShadowNinja kaeza: So, all good? 17:19 kaeza (that is, if you care to review a 25K lines patch) 17:20 ShadowNinja kaeza: Eh, I'll trust you on whether or not it works. 17:21 kahrl ShadowNinja: we should merge (or reject) the remaining translation pull requests first 17:22 ShadowNinja kahrl: The updatepo.sh change can be merged before or after, actually running it will have to be later though. 17:22 kahrl yeah, the 25K figure made me think the patch did both 17:22 ShadowNinja kaeza: All good and ready to merge (just the updatepo.sh change)? 17:22 kaeza yep 17:23 ShadowNinja kahrl: He linked the updatepo.sh commit seperately. 17:23 kahrl ah, ok 17:23 ShadowNinja (A long time ago, it's https://github.com/kaeza/minetest/commit/7b0a31ce2134cdc01b1b57423eafabedc50bb811 ) 17:27 kahrl kaeza: re: weblate, I managed to set it up on my test machine, but it still has some issues 17:27 kahrl http://dev.minetest.net/Installing_Weblate is what I did 17:27 kahrl it hasn't been put into production yet 17:28 kahrl actually, I think weblate 2.0 might come out soon(TM) so we might wait for that 17:29 ShadowNinja kahrl: Does https://github.com/minetest/minetest/pull/1598 look O.K.? That's the only one that does more than update a translation file. 17:30 ShadowNinja kahrl: Also check the other pulls at https://github.com/minetest/minetest/pulls?q=is%3Aopen+is%3Apr+label%3A%22%40+Translation%22 -- They all look good, but I don't know any of those languages. 17:30 kahrl 2d7a78eb56 should be skipped 17:30 ShadowNinja kaeza: Merged your patch, BTW. 17:30 kahrl (that's a commit in #1598) 17:31 ShadowNinja kahrl: If the rest good? 17:32 kahrl looking 17:39 kahrl I don't really like passing gettext()-ed format strings to (sn)printf 17:39 kahrl it seems like a security issue waiting to happen 17:40 kahrl I don't if there's an easy better way though 17:40 kahrl don't know* 17:44 kahrl the rest seems fine 17:44 kaeza kahrl, ah ok, good to know 18:27 zat hmmmm: for the beginning of DecoSimple::generate http://pastebin.com/bEDjjrrY 18:27 zat I hope I didn't mess up with the CONTENT_IGNORE part 18:29 zat ok I actually messed it up... when place_on is a group, c_place_on is CONTENT_IGNORE 18:31 hmmmm sure, great stuff, but I hope that bit of code is compartmentalized for only the case where place_on is a group 18:36 zat it actually works for both cases 18:36 zat would it be better if I made it work the old way if its not for a group? 18:37 zat it would be more efficient I think 18:42 zat hmmmm: http://pastebin.com/6K2KdvNB 18:56 hmmmm hrmm 18:56 hmmmm zat, while that is good code, and what I asked for, I just thought of a better way to do this 18:57 hmmmm perhaps there should be a c_place_on vector 18:58 hmmmm see the thing about it is, performance matters a lot here because this loop is happening thousands of times per call, and there are hundreds of calls per second 18:58 zat hmmmm: that wauld me much better yes lolololol 18:58 hmmmm but 18:58 hmmmm correct me if I'm wrong 18:58 zat yes, I always prefer comparing integers too 18:58 hmmmm place_on is currently a string, right? 18:58 hmmmm err rather, expected to be a string when read from Lua 18:59 hmmmm if it's read in as a table, or there's a group, then it becomes a vector 18:59 zat place_on_name in a string 18:59 hmmmm yeah 18:59 zat c_place_on is the id 18:59 zat in... is* 19:00 hmmmm so I'm thinking of something like this: 19:02 hmmmm place_on = ["group:soil", "default:water", "default:stone"] 19:02 hmmmm just as an example 19:02 hmmmm err... that's Python... s/[/{/ 19:03 hmmmm so then the place_on vector would contain all the IDs of soil, water, and stone 19:03 hmmmm or if it's just 19:03 hmmmm place_on = "group:soil", then it'll be a vector as well 19:03 hmmmm yeah, okay, for simplicity's sake I guess you'll have to use a vector every single time 19:04 hmmmm although you never really know what will be more beneficial without profiling it 19:06 zat I doubt there would be even a noticeable difference 19:06 hmmmm of course there won't be a "noticable" difference, but most things aren't noticable until they're scaled up which is when it starts to hurt 19:08 zat hmmmm: your idea is good, but... honestly.... 19:08 zat no idea where to start lol 19:09 hmmmm ModApiMapgen::l_register_decoration 19:09 hmmmm see deco->place_on_name? 19:09 hmmmm well first of all std::string place_on_name should turn into std::vector place_on_names 19:10 hmmmm lua_getfield(L, index, "place_on"); 19:10 hmmmm if (lua_isstring(L, -1)) { 19:10 hmmmm std::string name = getstringfield(L, index, "place_on"); 19:11 hmmmm if (name.substr(0, 6) == "group:") { 19:11 hmmmm ndef->getIds(name, result_vector); 19:11 hmmmm err 19:11 hmmmm ndef->getIds(name, deco->place_on_names); 19:11 zat im not getting it lol 19:11 hmmmm erm 19:11 zat are you busy? 19:12 hmmmm yes sorta 19:12 hmmmm I would just do it myself if I weren't 19:12 hmmmm oh shit 19:12 zat ¿¿ 19:13 hmmmm you can't do ndef->getIds there, you need to just keep it into a place_on_names vector 19:13 hmmmm and then resolve them with getIds in resolveNodeNames 19:13 zat it shouldnt be like that 19:13 hmmmm if you can come up with a better way I'd love that 19:14 zat I mean, integer ids shoud be always retrieved at startup time (by my standards) 19:14 hmmmm consider the fact that node aliases don't take effect until after the registration phase entirely 19:14 hmmmm and that nodes may be defined after that file gets loaded and that line of code executed 19:16 hmmmm as a result of the problems this caused, whenever you go to register some sort of node to do something in the lua api, you store the string name of it. if it's performance intensive, put the part where you translate it from a string to an integer id in a resolveNodeNames()-type function 19:16 iqualfragile_ idea: python api4minetest4life 19:20 VanessaE ew. 19:31 zat hmmmm: a cache for that sounds fine 19:31 zat feed the cache as soon as new nodes are registered 23:25 MinetestForFun Hi, my server runs on the lastest dev version of minetest, so only clients with 0.4.10-dev + version of minetest can connect to my server, is it possible to show a message when a client with the bad client version try to connect in my server, for explain to the player "he needs a more updated version of minetest" ? (i don't know if a variable of "minetest_version exist", and if my idea is possible) 23:26 MinetestForFun (my idea is a windows like the no_bad_names mod, but i need to check the client version of minetest) 23:27 kahrl MinetestForFun: you might want strict_protocol_version_checking = true in minetest.conf 23:28 kahrl (protocol versions are not exactly the same as minetest versions though) 23:29 MinetestForFun i've try the "strict_protocol_version_checking = true in minetest.conf", but it's not adequate, many clients in 0.4.10 can't connect to my server 23:30 kahrl the client version can't be checked with the mod api, not reliably at least 23:31 MinetestForFun ok, at least, i am sure... :/ 23:32 MinetestForFun thank you for yours answers 23:32 MinetestForFun your answers*