Time Nick Message 00:24 harrison a minute to compile? 00:24 harrison slow 00:25 harrison i can compile my os itself in half that 00:30 Exio4 i don't own a server farm 00:30 Exio4 ;) 01:25 harrison i do not understand. how would a farm be relevant? 01:25 harrison when i talk about blocks and tackles 01:25 harrison frocks and spackle 01:25 harrison dogs and fractals 01:25 harrison or st francis with a flock of grackles 01:25 harrison what exactly happens to your hackles? 02:15 kahrl I think I found a fix for the bad design of the ScriptApi hierarchy: https://github.com/kahrl/minetest/commit/c1e35ba1d4c4585f32dfeeff9a880b62d388f94b 02:15 kahrl sapier: ^ 02:16 kahrl (why do these commits always end up so large?) 02:26 Tesseract kahrl: Overview? Also, perhaps MenuScriptApi instead of ScriptApiForMainMenu. 02:28 kahrl I'm a bit busy now, but I'll write an overview later 03:41 kahrl having two different classes ScriptApiMainMenu and MenuScriptApi might be confusing 03:42 kahrl not that ScriptApiMainMenu and ScriptApiForMainMenu is much better :P 03:42 kahrl Tesseract: any ideas? 03:43 Tesseract Hmmm, I guess I don't understand this heiarchy... 03:43 kahrl ScriptApiMainMenu contains the functions specific to the main menu scriptapi 03:44 kahrl ScriptApiMainMenu contains *everything* available in the main menu scriptapi 03:44 kahrl *ScriptApiForMainMenu <-- getting confused myself 03:45 Tesseract Yea, perhaps they could be renamed. MenuOnlyScriptApi? 03:46 kahrl 1 sec 03:54 kahrl http://paste.dy.fi/AyR 03:54 kahrl ScriptApiMainMenu works like the others (ScriptApiDetached, ...) so it shouldn't arbitrarily have a different naming convention 03:56 kahrl (this diagram is a bit simplified, I omitted ScriptApiNodemeta and ScriptApiItem, but they're not important here) 03:58 kahrl note that some day one might want to add a ScriptApiUtil that is used by both ScriptApiForGame and ScriptApiForMainMenu 06:19 celeron55 kahrl: looks reasonable to me; i think some of that explanation should be put in a comment in somewhere in the code in a place where someone is likely to look at when looking for how the scriptapi stuff is constructed 06:20 celeron55 (if such isn't already there) 06:24 kahrl celeron55: I changed it so that ScriptApiForGame is called GameScripting and ScriptApiForMainMenu is called MainMenuScripting 06:25 celeron55 but really, i don't even care; i trust that you do sane things 06:25 kahrl and they live in script/scripting_{game,mainmenu}.{cpp,h} and initialize the required mod API modules 06:25 kahrl so I got rid of all source files in script/cpp_api that don't start with s_, and all in script/lua_api that don't start with l_ 06:26 kahrl improves tab completion :P 06:28 kahrl is the dev wiki a place somebody is likely to look at? 06:29 kahrl I was going to put a whole page there 06:30 celeron55 well that's fine 07:00 kahrl https://github.com/minetest/minetest/pull/867 updated 07:02 kahrl ugh, github can't do angle brackets... fixed 07:04 nore_ kahrl, I have a question about omnicleanup 07:05 nore_ if i understood correctly, it would be a base for client-side mods? 07:45 kahrl nore_, no 07:45 kahrl that's still a long ways away 11:18 PilzAdam kahrl, would this be the correct page to describe your changes: http://dev.minetest.net/Script_Engine ? 13:33 nore could someone look at #858 and #859 ? 13:33 nore see the logs of yesterday 13:33 nore there is a discussion about those 16:26 sapier did noone realize the texture pack mainmenu commit doesn't match mainmenu coding style or didn't just anyone care? 16:27 thexyz sapier: what commit? 16:28 sapier the mainmenu changes that did add texture pack selection support 16:28 sapier it's quite good but neither variable nor function names match style of everything else 16:28 thexyz https://github.com/minetest/minetest/commit/383153419bef744af4bfa5f5d54c4bee663a5ce6 ? 16:28 sapier yes exactly 16:29 thexyz i guess no one cared because there are no style guidelines for lua code 16:29 thexyz (are there?) 16:30 sapier I don't think so but in that case I try to use the style of already existing code ;-) 16:30 sapier maybe we should add some hint like that to guidelines 16:31 thexyz oh 16:32 thexyz you renamed me to "IIya" 16:32 sfan5 "IIya" 16:33 thexyz "how to offend people" 16:33 sapier htop 17:22 sapier1 https://github.com/minetest/minetest/pull/870 should make updatepo.sh work again 17:45 nore could someone look at #869 and #868 ? 17:47 sapier1 868 is wrong for singleplayer 17:47 nore where? 17:47 sapier1 i guess it was wrong before too :-) 17:48 sapier1 oh wait 17:48 sapier1 forget about it you're right 17:49 sapier1 but you should change " world_doubleclick" to something different if it's not really a doubleclick anymore 17:51 nore sapier: done 17:51 sapier1 869 changes view of mainmenu text 17:52 sapier1 this may be a problem in favourite server description as well as mod download store did you check if the results are resonable after not escaping newline and cr anymore? 17:54 sapier1 237 what do you do if first char is a escape charß 17:55 nore what happens then? 17:55 sapier1 your code will fail to unescape it 17:55 nore why? 17:56 sapier1 because you'll never realize it 17:56 sapier1 your for loop is from 1 to size thus skipping first char ... why do you even use the fore loop it's by no means better than using stl find 17:57 nore nope, it works 17:57 nore and what for loop? I did not add code, just used the old one 17:57 nore tested with first char = \, no bug 17:58 sapier1 guiFormSpecMenu.cpp L 234 17:58 nore with ; it works too 17:59 nore ah, you mean the very first char... 17:59 sapier1 exactly 17:59 nore it looks like this cannot happen because the first char is always position 18:00 sapier1 you know your code may result in a full copy of string for any single char? 18:00 nore it is not my code, it is the old one 18:01 sapier1 ohhh now I see what you've done 18:01 sapier1 ... I hate that diff viewer 18:01 nore and what do you mean, a full copy? 18:01 nore those lines are shown modified because I removed one level of indentation 18:02 sapier1 I didn't realize this before but the for() loop may result in a full copy of the std string when adding each char 18:02 sapier1 but it was "wrong" before so not your fault 18:03 nore yes, it is a full copy of each char of the string 18:03 nore but it cannot be done in another way 18:04 sapier1 I'm talking about current += s.c_str()[i]; in worst case this is done by relocation of full string 18:04 sapier1 I don't think our strings are big enough for this to happen 18:05 sapier1 and the i=1 error was there before too 18:05 sapier1 but you should fix it as you already change it 18:06 kahrl with C strings, you *could* avoid reallocation by just replacing all separators with NULs and returning a list of pointers to the beginning of each part 18:06 kahrl but it's really not worth the hassle here 18:07 nore I did not think to replace separators with 0s,... 18:07 sapier1 I assume stl will allocate reasonable big strings by default (hope I'm right with this asumption) 18:07 nore my idea was to put counted strings instead of this, but this is way too complicated 18:07 sapier1 you'd need to rebuild the string afterwards 18:08 sapier1 if this was called hundreds or thousand times a second it'd be worth it but it's a function used in gui ... if ppl click that fast we should slow them down for their own wellbeeing ;-) 18:08 kahrl btw, why s.c_str()[i] and not just s[i]? 18:09 nore done the change 18:09 sapier1 good question ... I guess there's no reason for it 18:10 sapier1 did you push it? 18:11 nore yes 18:11 sapier1 I hate web page caching :-) 18:12 nore I hope I did not make errors... 18:12 sapier1 did you check if drop of \r\n replacement has sideeffects in modstore and favorites list? 18:13 nore how do you access favourites? 18:14 sapier1 I just added this to get reasonable server description in favorites list knowing it'd be better for server owners to update their descriptions ... but they where already a little bit reluctant due to mainmenu changes that time 18:14 sapier1 do you have curl compiled in? 18:14 nore I don't think so 18:15 nore just -DRUN_IN_PLACE=1 18:15 sapier1 you should you can't check those things that may be affected by your changes 18:15 sapier1 -DENABLE_CURL=1 if I'm correct 18:15 nore I will perhaps add conversion \r\n -> \n, and \r -> \n 18:16 nore it should be enough, no? 18:17 sapier1 the problem was some ppl used \n\n\n\n in their descriptions as those were shown on webpage only by that time ... so nothing was visible within minetest ... maybe it's fixed by now so we can drop the removal completely ... but you need to check it to tell 18:17 nore and replacing \n\n with \n? 18:18 sapier1 good question I'd prefere \\n to \n 18:18 nore in fact, no... that could break some things 18:18 sapier1 but not sure what'd be best 18:18 kahrl just do it for the description field? 18:18 sapier1 you need to try what works and what doesn't 18:19 sapier1 favorites description as well as mod download description 18:19 nore or I add a function remove_newlines, and add it for those? 18:20 sapier1 is another option but it's a little bit anoying to call multiple string cleanup functions 18:20 sapier1 just try if it really is a problem first maybe we don't need to do anything 18:21 sapier1 if everyone fixed his description we don't need to support it any longer 18:23 kahrl offtopic: have you seen my omnicleanup changes sapier1? 18:24 sapier1 not yet I'll have a look 18:26 nore installing libcurl... 18:27 sapier1 :-) it's always "add a nother layer of indirection" ... but I think it's a good solution in this case 18:29 nore how do I see modstore and favourites? 18:29 nore this time I have a build with CURL 18:29 kahrl sapier1: not really, in this case one cheese layer and one tomato layer at the same level ;) 18:30 kahrl where it was a single layer before 18:31 sapier1 oh :-) misinterpreted it due to my thoughts ... 18:31 sapier1 not as good as I hoped to be but I guess no chance to get a better solution ;-) 18:32 kahrl oh, did you have a solution in mind? 18:32 sapier1 no details 18:32 kahrl ok 18:32 sapier1 I thought about some base class that a initialize function would be passed to 18:33 sapier1 but haven't tried if this wouldn't result in other problems 18:34 kahrl the scriptapi and modapi modules need to be able to access the Server*/Environment*/GUIEngine*, that's hard when the base class doesn't know about them 18:35 sapier1 that was one of the things I meant with "other problems" 18:35 sapier1 I guess your solution is best to get without major redesign 18:37 kahrl at least when you try to initialize mainmenu scripting without a GUIEngine* the assert in l_get_scriptdir will quickly remind you 18:37 kahrl < did that once 18:38 hmmmm woah, all this lua api stuff 18:38 sapier1 :-) could be difficult to share data :-) 18:38 hmmmm remind me to not start on designing client-side lua until all this is less volatile 18:38 sapier1 hmmmm and we didn't even start to do the tough things ;-) 18:39 sapier1 client side lua isn't related to this at all 18:39 kahrl well with omnicleanup it should be easier to add a third API 18:39 hmmmm well of course not, but you're changing the structure of the api from underneath 18:39 sapier1 I will fight with hands and feet against using this code for client side lua ;-P 18:39 hmmmm (I mean actually coding it, not designing things) 18:40 hmmmm by the way, there was something I was working on to clean up weather somewhat that modifies a bunch of scriptapi things 18:40 kahrl sapier1, what is the problem with it? just don't do luaL_openlibs maybe and just use custom API modules specifically designed for it? 18:40 hmmmm i'll hold off on it until you get done with what you're doing so there are no conflicts 18:40 sapier1 client side lua is server executing unknown code at client ... thus we have a completely different security level than mainmenu or mods 18:41 hmmmm oh yeah, sapier, this is an awesome way for you to add in security to the lua api 18:41 hmmmm i guess you can design it in a way so that different lua api sets have different security levels 18:41 sapier1 kahrl if you leave this out there's nothing left ;-) 18:42 hmmmm sorry the previous two didn't work out, but this is an excellent opportunity for adding security 18:42 sapier1 that's been the plan client side lua can't be done without security so maybe the features can pass back to modapi after they are done 18:43 kahrl sapier1, you could load the table, string and math libs only 18:43 kahrl http://stackoverflow.com/questions/4551101/lual-openlibs-and-sandboxing-scripts first answer 18:43 sapier1 I assume not even those who fight against mod security are willing to allow a server to delete all files on their client :-) 18:44 Exio4 how would you make a way to only allow some mods to access os. for example? 18:44 Exio4 sapier ^ 18:44 kahrl Exio4, that's more difficult because those run in the same interpreter as other mods 18:44 sapier1 won't be allowed at all unless someone really has a legit uscase to store server data on client machine 18:44 Exio4 yeah kahrl 18:45 sapier1 yes mod isolation is even more tough than client side lua security 18:45 kahrl if you mean for server-side scripting, we could provide a minetest.getlualib("os") that can only be called during the initialization phase and checks the current mod name 18:46 sapier1 kahrl there's more we need to disable debug too as you can access anything within lua stack by this functions 18:46 kahrl yeah, stuff like metatables 18:46 sapier1 another open issue 18:47 sapier1 as far as I know serialize requires metatables atm 18:47 kahrl at least in client side scripting, if you don't load the os, io modules etc. at all you don't have to try to hide via metatables 18:47 sapier1 mod isolation would have another sideeffect, same thing required for isolation can be used for mod profiling 18:48 sapier1 yes client side will be a very very stripped down lua env 18:48 sapier1 the only usecase I know atm is scripting of some environment behaviour e.g. door rotations 18:49 sapier1 maybe entity movement prediction too 18:49 kahrl that, sounds, and hud 18:49 sapier1 yes so basicaly things that affect (client side) temporary data only 18:51 sapier1 but that'll have to wait until autumn now it's time to update mobf first ;-) talking about it what about 774? 18:56 hmmmm storing server data on a client machine is completely useless 18:57 sapier1 that's what I intended to say ;-) 19:00 kahrl are there any opinions on merging omnicleanup? 19:01 hmmmm i think it's great 19:01 sapier1 I'm fine with it 19:01 kahrl c55 also approved, so merge now? 19:02 hmmmm yup 19:03 sapier1 anyone to check 870? it's should fix i18n for mainmenu again 19:05 kahrl sapier1: looks good on first sight 19:05 kahrl I assume xgettext is clever enough to not break horribly from the lua syntax? 19:06 sapier1 it compains about not knowing what to do with *.lua and using c therefore 19:06 sapier1 along with some warnings but seems to work despite of this 19:07 sapier1 of course I can't say it's gonna work with any strange language construct lua supports 19:24 PilzAdam kahrl, /home/adam/Minetest/minetest/src/server.cpp:920:9: warning: deleting object of polymorphic class type 'EmergeManager' which has non-virtual destructor might cause undefined behaviour [-Wdelete-non-virtual-dtor] 19:32 sapier1 pilzadam you know how updatepo.sh should work could you check it does now what it's intended to do? 19:32 PilzAdam sure 19:33 sapier1 if you see any texts I missed tell me 19:33 sapier1 https://github.com/minetest/minetest/pull/870 this should help 19:34 proller PilzAdam, https://github.com/minetest/minetest/pull/833 -- who approved it? 19:40 PilzAdam proller, RealBadAngel 19:41 PilzAdam I dont see how this is relevant, though 19:41 proller it have division by zero 19:41 PilzAdam hm? 19:41 proller and make BIG BUG when you join to old servers 19:42 proller and why it touches src/content_mapblock.cpp 19:42 proller why not make it only in map? 19:42 PilzAdam the drawtype needed to be adjusted 19:43 proller very bad when range == 0 19:45 PilzAdam just change the rangelim() to 1 then 19:45 PilzAdam there shouldnt be any flowing liquid nodes with range 0, though 19:45 proller https://github.com/minetest/minetest/issues/840 19:45 proller and its broke by other way 19:46 proller old server have no rangelim 19:46 proller ups, range 19:46 nore PilzAdam: what do you think of #868 and #869? 19:46 PilzAdam well, guess why there is a protocl version bump 19:49 proller PilzAdam, and here - https://github.com/minetest/minetest/blob/master/src/content_mapblock.cpp#L264 - its mispell in comment or bug in code? 19:51 Exio4 uhm? 19:52 Exio4 the !? 20:04 PilzAdam sapier1, the tabs are not translated 20:04 PilzAdam in the "config" window "World:" at the top and "Depends:" translations do not work 20:05 sapier1 config window? 20:05 sapier1 world config? 20:05 PilzAdam yea, the one to enable mods 20:06 sapier1 tab translation is a problem 20:06 PilzAdam also the translation of "Installed mods" in the "Mods" tab does not work 20:06 PilzAdam ("does not work" means that I have translated it in the .po file, but it simply ignores it) 20:07 PilzAdam "Depends:" in the "Mods" tab doesnt work either 20:07 PilzAdam and "Core devs", "Active contr." and "previous contr." in the "Credits" tab 20:07 VanessaE kahrl: can you check that omnicleanup commit against the vbo patch please? (co0nflict in mapsector.cpp) 20:08 VanessaE -0 20:08 sapier1 I guess I know why it doesn't work 20:09 VanessaE (the vbo stuff needs to go in ASAP before it gets lost) 20:11 PilzAdam sapier1, also, why are "mods_current" and "mods_available" added to the .po files? 20:11 PilzAdam they dont seem to be text that needs to be translated 20:11 PilzAdam proller, dunno... it works, so I better dont touch it ;-) 20:12 PilzAdam proller, I think we should make the range rangelim to 1 minimum 20:12 kahrl PilzAdam: about that warning, that shouldn't be any actual problem as EmergeManager is never delete polymorphically 20:12 kahrl deleted* 20:12 PilzAdam its still a warning... 20:13 kahrl add a virtual destructor to IBackgroundBlockEmerger? 20:13 proller PilzAdam, with finite liquid when flowing rare blocks sometimes disappear for one cycle 20:13 sapier1 core devs etc are a problem because those have a color key in 20:13 kahrl VanessaE: what do you mean by checking the commit 20:14 proller PilzAdam, and 1 make ugly flowing on old servers 20:14 VanessaE kahrl: something in your commit broke the vbo patch 20:14 sapier1 i think I can fix it in core 20:14 kahrl what's the conflict? 20:15 VanessaE I'm not sure, I suck at resolving such things, I only know it's in mapsector.cpp :P 20:15 kahrl probably some #includes 20:15 PilzAdam VanessaE, grep "<<<<" in that file 20:15 VanessaE src/mapsector.cpp:21 to be more exact 20:15 kahrl can you paste the entire mapsector.cpp? 20:15 VanessaE er 20:15 VanessaE hang on a sec 20:16 VanessaE https://github.com/PilzAdam/minetest/blob/7af479cba466f801c5cb648c8a449877957cb723/src/mapsector.cpp 20:16 kahrl well no, I mean the one on your disk that shows the conflict 20:18 VanessaE it's that ^^^ versus https://github.com/minetest/minetest/blob/4e1f50035e860a00636ca5d804c267119df99601/src/mapsector.cpp 20:18 sapier1 pilzadam escaping is a problem with i18n 20:18 kahrl you need to fix the mapblock_mesh.h include, I think 20:18 VanessaE (I git --reset and updated to HEAD) 20:20 PilzAdam proller, that visual glitch can be maybe fixed by catching if the server doesnt send range 20:21 kahrl VanessaE: try https://github.com/kahrl/minetest/commits/vbo 20:22 VanessaE that goes in clean 20:24 VanessaE lesse if it compiles :) 20:25 proller PilzAdam, my position is: range must not affect visual, only distrubution in map.cpp, and better to solved via 1/viscosity ;) 20:25 PilzAdam proller, have you tried the previous drawtype with range=2 or 3? 20:25 PilzAdam it just doesnt look good 20:26 PilzAdam VanessaE, updated my branch 20:28 proller PilzAdam, no, i play only with finite_liquid and this range totally against it 20:28 proller and it have many visual glitches, needed to fix 20:31 proller i think about rewriting visual to 8 triangle mode 20:37 PilzAdam so use the range in the drawtype only if normal liquids are used 20:38 VanessaE PilzAdam: since that vbo code clearly works as intended (I've been using it for a while now) and can be disabled by default, is there any particular reason it hasn't been merged yet? 20:39 PilzAdam VanessaE, because it leaks like hell? 20:39 PilzAdam it does not work as expecte 20:39 PilzAdam +d 20:40 VanessaE "work as expected" == "improves performance as it was originally written to" 20:40 VanessaE but okay 20:40 VanessaE if it leaks, fine. disable it by default and fix it later? 20:42 sapier1 pa plz recheck i18n 20:56 PilzAdam sapier1, "Core Devs" still doesnt work 20:56 sapier1 I guess everything in there? 20:56 kahrl this makes flowing liquid textures flow diagonally when the liquid flows diagonally: https://gist.github.com/kahrl/6235515 20:57 PilzAdam sapier1, umm... https://gist.github.com/PilzAdam/6235529 20:57 PilzAdam I clicked on a texture pack and got an error, then I closed Minetest and when restarting it crashes with this message 20:58 sapier1 hmm I thought I fixed that one 20:59 kahrl VanessaE, can you check if my patch fixes #853? 20:59 sapier1 1056 doesn't have a texture_pack_path variable? 21:00 sapier1 its 1052 what version do you have? 21:01 PilzAdam master + your #870 21:01 sapier1 very strange 21:01 PilzAdam what version do _you_ have? ;-) 21:01 sapier1 but I fixed it either .. so back to the Core Devs issue 21:04 PilzAdam sapier1, could you tell me how I fix that? I cant start Minetests menu anymore 21:04 sapier1 I already pushed the fix 21:04 sapier1 I think I found why coredevs isn't working too 21:05 proller PilzAdam, kahrl why in flowing bottom face is empty ? 21:05 PilzAdam sapier1, pushed? to what branch? 21:05 VanessaE kahrl: sure, hold a sec 21:06 sapier1 the pull request 21:06 sapier1 now the coredevs fix is in too 21:07 kahrl proller, dunno 21:07 PilzAdam kahrl, works fine 21:08 kahrl PilzAdam: it would look nicer with lava if the lava texture tiled diagonally, I guess 21:09 PilzAdam true 21:10 sapier1 pa did you find the fixes? 21:12 PilzAdam sapier1, it kinda works 21:12 PilzAdam but I get endless error messages 21:12 sapier1 for updatepo? 21:12 PilzAdam Could not open file of texture: /home/adam/Minetest/minetest/textures/TestBDcraft/screenshot.png 21:12 PilzAdam 23:12:22: ERROR[main]: GUIFormSpecMenu::drawMenu() Draw images unable to load texture: 21:12 PilzAdam 23:12:22: ERROR[main]: /home/adam/Minetest/minetest/textures/TestBDcraft/screenshot.png 21:13 sapier1 oops :-) 21:13 VanessaE whoa 21:13 VanessaE diagonally-flowing? 21:13 sapier1 file_exists(screenfile) 21:13 VanessaE iiiiiinteresting 21:13 sapier1 line 1053 21:14 PilzAdam 1503 is "f:close()" 21:14 PilzAdam *1053 21:14 sapier1 so +4 for your code 21:14 VanessaE kahrl: it certainly fixes it 21:14 sapier1 its if not file_exists then 21:14 sapier1 the parameter is missing :-) 21:15 PilzAdam sapier1, that works, but mainmenu.lua:1067: attempt to concatenate global 'no_screenshot' (a nil value) 21:16 PilzAdam you need to declare local no_screenshot before the if 21:17 PilzAdam "No information availible" should be translated too 21:17 sapier1 yes fixed ... can someone plz create a minetest repository with at least a single one of all optional addon types included 21:17 VanessaE kahrl: indeed diagonal tiling would look better if it can be achieved without a performance hit, but this is good anyway 21:17 sapier1 it's difficult to test without having thousands of things installed everytime 21:17 PilzAdam sapier1, are you too lazy to download texture packs? :-p 21:18 sapier1 of course who needs that crap? 21:18 sapier1 just slowing down startup 21:18 sapier1 you should know you start a lot when developing a mod ;-P 21:19 PilzAdam "World:" at top of the modmanager isnt translated too 21:19 sapier1 and of course I removed everything not beeing present due to issues with things not working when everything is present ... now I have exactly opposit problem ;-) 21:20 sapier1 this one is tough the text isn't static 21:24 sapier1 ok should work this way 21:31 PilzAdam works now 21:31 sapier1 ok do you se any other missing strings? 21:31 PilzAdam nope 21:32 PilzAdam wait, the vertical "SETTINGS" text is missing 21:34 sapier1 done 21:34 sapier1 I guess the error messages wont work too ... but I don't see a quick fix for that 21:34 sapier1 at least those error messages containing variable parts 21:37 nore_ PilzAdam, did you look at #868 and #869? 21:41 PilzAdam thexyz, can you push the weblate commits so I can run updatepo.sh? 21:43 kahrl I just checked in minecraft, and their lava also flows diagonally but has visible seams 21:43 kahrl so making it tile seems not trivial 21:43 VanessaE time to hide the seams :) 21:44 kahrl of course it's not as visible there because lava flows only 3 nodes (unless you're in the nether) and often only flows in 1 direction 21:44 PilzAdam nore_, no, Im obviously busy with other things 22:59 kahrl trivial fix for the virtual dtor problem, pushing in 15 minutes: https://gist.github.com/kahrl/6236578 22:59 PilzAdam why dont you push that now? 23:00 kahrl just in case somebody protests 23:01 kahrl don't know why they would but I'm not in a hurry