Time Nick Message 06:21 Lord_Evil Hello. Need some help with debuging Android build. 06:21 RealBadAngel Lord_Evil, you shall talk to sapier then 06:22 Lord_Evil RealBadAngel, where can I find him? 06:22 RealBadAngel here, but a bit later propably 06:23 Lord_Evil RealBadAngel, thanks :-) Spent two days building it, but now it sigfaults on me =( 06:24 RealBadAngel he shall be there evenings and nighttime 06:25 RealBadAngel CET 06:26 Lord_Evil Ah, I see. Ok, I'll wait. 06:30 nore thoughts on https://github.com/minetest/minetest/pull/1441 ? 07:27 hmmmm I wonder if it'd help if we used OpenGL texture compressio 07:27 hmmmm compression 07:28 VanessaE YES 07:28 VanessaE G*D YES 07:29 VanessaE I have been asking for that for the longest time, hmmmm. 07:29 hmmmm thing is that irrlicht doesn't seem like it supports that 07:29 VanessaE and I do believe there is at least one compression loader for irrlicht 07:29 VanessaE EDS or something? 07:29 VanessaE no, kds 07:29 VanessaE ran into it while looking for something only tangentially related. 07:30 VanessaE http://irrlicht.sourceforge.net/forum/viewtopic.php?t=45988 07:30 VanessaE ktx 07:30 VanessaE that's the one. 07:30 hmmmm yeah I am reading that same thread 07:31 VanessaE btw, I have an AMD R9 280X (3GB GDDR5, OC edition) on the way, if that's of any interest to you. 07:33 hmmmm well 07:33 hmmmm the KTX loader on that page was some guy's third party irrlicht modification 07:33 hmmmm like i said irrlicht doesn't support it 07:33 VanessaE yeah, that's what I gathered too 07:33 hmmmm so I don't get it 07:33 hmmmm do we dump irrlicht and start packaging our own vesrion 07:34 hmmmm that isn't screwed beyond belief? 07:34 VanessaE it is. 07:34 VanessaE it's not a good idea at all 07:34 VanessaE but that raises another question: 07:34 VanessaE texture compression is very old tech 07:34 VanessaE I mean, liek 15+ years old 07:34 VanessaE like* 07:34 hmmmm irrlicht is just highly deficient 07:34 VanessaE if irrlicht can't and won't support it, what do we do? 07:35 hmmmm well we went over this with something else a while back 07:35 hmmmm i forget 07:36 hmmmm I guess we're going to have to mix raw GL with irrlicht 07:36 hmmmm as for the other renderer options, texture compression will be unsupported ... 07:38 VanessaE I presented the options to Abe (he used to be a coder back in the stone ages). His suggestion is what amounts to ripping out irrlicht and rebasing against something else. 07:39 hmmmm well i hope abe has a lot of free time for rewriting minecraft 07:39 VanessaE heh 07:39 hmmmm minetest i mean 07:40 VanessaE well I suggested three: rip out and rebase (e.g. could be OGRE or something else as the new target), or supply a patched irrlicht with all the new features we seem to need lately, or do something like you're thinking, which I think we can both agree will be ugly and hacky from a code standpoint 07:40 hmmmm the latter of those three is actually the cleanest 07:41 VanessaE in the short term it would be, yes 07:41 VanessaE but I wonder how maintainable it will be in the long term? 07:42 VanessaE RBA has gone through hell just making some of the simplest shader effects work correctly in the majority of cases (and that still falls down for some people) 07:42 hmmmm providing a patched irrlicht vs. maintaining a bit of raw OGL that may need to touch some of irrlicht's internal interfaces 07:43 hmmmm =/ 07:43 hmmmm seems pretty obvious to me 07:45 VanessaE well I shouldn't say a great deal more about this, lest I jump in over my head 07:46 hmmmm texture compression is something that would help a lot but it's going to take a lot of work to acheieve 07:46 hmmmm there's a reason why it wasn't already in minetest 07:47 VanessaE I assumed it simply hadn't been considered at all 07:47 VanessaE I mean, it wasn't needed 07:49 hmmmm heh 07:49 hmmmm once we get real lighting in minetest, glowmaps are next 07:49 hmmmm there's so much potential 07:50 VanessaE nice 07:51 VanessaE I guess RBA will do those 07:51 VanessaE oh btw 07:51 VanessaE he's starting to attack the one-scenenode-per-some-minetest-element idea now 07:51 hmmmm who RBA? 07:51 hmmmm i was gonna do that 07:51 VanessaE yeah 07:51 VanessaE only he's pointing his efforts at a direct 1:1 translation 07:51 VanessaE one scene node = one minetest node 07:51 hmmmm oh.. =/ 07:51 hmmmm hope it works out, i doubt it will though 07:51 VanessaE except idk how much of that has been done, if anything 07:52 * VanessaE pokes RealBadAngel 07:52 hmmmm well on the bright side 07:52 hmmmm if it's viable, then you can remove meshmaking entirely 07:55 VanessaE hm, another compression thread was linked off of that last one, http://irrlicht.sourceforge.net/forum/viewtopic.php?t=10488 07:55 VanessaE the download link therein still works 08:09 nore does anyone have thoughts on #1441? (is it small enough to go in before 0.4.10?) 08:29 Calinou what's the potential uses? knowing if a ore is registered or not? 08:39 nore Calinou, for mapgens 08:40 nore i.e., if you make a lua mapgen, you can spawn the ores that would have been spawned 08:40 nore you can even use a custom spawning code... 08:40 Calinou oh, this is cool then 08:41 Calinou if nothing goes wrong while testing, why not merge it? 08:47 nore well, I need another core dev to agree... especially since we're in feature freeze now 11:11 celeron55 10:53:07 < VanessaE> one scene node = one minetest node 11:11 celeron55 lol what 11:12 celeron55 that isn't going to work at all, i don't think RBA is that stupid 11:12 celeron55 one scene node = one mapblock could work 11:12 VanessaE celeron55: well I believe that's what the proposal was. I never said it would work though :) 11:15 VanessaE [07-02 15:02] node shall be a single mesh 11:15 VanessaE [07-02 15:02] with own properties 11:16 VanessaE prior to that, 11:16 VanessaE [07-02 14:38] we just have to go back 11:16 VanessaE [07-02 14:38] node = node 11:16 VanessaE so yeah. 11:22 Jordach if this solves the aging problem of bad meshes beyond 32km then i'm for it 11:41 crazyR hey all, im trying to create a button in the unified inventry. how can i close the unified inv when the button is pressed. the reason i ask is because it needs to load a diffrent formspec on button press 11:42 crazyR oh wrong channel sorry 11:56 RealBadAngel crazyR, waypoints are a perfect example of using UI's API 11:56 RealBadAngel in waypoints.lua you have all the code examples you will ever need 17:27 nore does anyone have thoughts on #1441? 17:28 nore (is it small enough to go in before 0.4.10?) 17:33 casimir in the tool capabilites what is max_drop_level for? 17:36 PilzAdam nore, no; feature freeze 17:36 nore ok 17:36 PilzAdam also people might be confused because changing the table doesn't affect the mapgen 17:36 PilzAdam and doc is missing 17:36 nore ^ ah, forgotten that 17:36 PilzAdam casimir, nothing; its use is to be defined by mods or games 17:37 nore and about changing the table: it is the same as changing the nodes table 17:38 PilzAdam no, nodes table is read after all mods ran, while the ore def gets loaded directly when calling register_ore() 17:40 nore hm... and what about ABMs? 17:41 PilzAdam same as nodes IIRC 17:41 nore well, anyway, for biomes it is as for ores 17:44 PilzAdam maybe both should be changed to work the same way as nodes 17:44 PilzAdam hmmmm, ^ 18:22 A_Zman705 hi everyone 18:23 A_Zman705 any1 feeling neighborly today 18:23 A_Zman705 im a complete newb just warning 18:24 sfan5 this is the development channel 18:24 A_Zman705 so far i used synaptics pack thingy and installed 18:24 sfan5 for general talk use #minetest 18:24 A_Zman705 yea im trying to git it to work 21:09 hmmmm PilzAdam, that's your call if you want to do something like that. 21:09 hmmmm just so that you're aware, though, ores don't have a unique ID other than their internal array position, i.e. the order in which they were registered 21:17 BlockMen do we have someone for building with MSVC now (for release)? 21:18 BlockMen if not i would do it 21:18 hmmmm xyz? 21:18 BlockMen i thought he quit? 21:18 hmmmm he did? aww 21:20 hmmmm I don't understand this commit https://github.com/minetest/minetest/commit/c410e9182d322a8c095ef94fbadf4d8f541e6b98 21:21 hmmmm why is this guy replacing the sysv sem_* functions? 21:25 hmmmm http://ck-hack.blogspot.com/2013/09/unnamed-semaphores-and-pososx.html ahh.. wtf 21:25 hmmmm anyway he should mention that in the body of his commit message 21:28 hmmmm lol. the excuse for OSX not having sysv semaphores is: "if Apple implemented unnamed semaphores, how many XServes would you buy?" 21:47 hmmmm https://github.com/minetest/minetest/commit/ea6c3835debfffa22163f048eae57928311143c4#diff-c03ca828c6b8a7695f2cd7e52c316a3cR778 21:47 hmmmm oh noes, what is this 21:47 hmmmm it looks like random bits being set and checked that should be constants 21:49 hmmmm i would like to see something like bool H_merge = param2 & GLASSFRAME_FLAG_EDGE == 0; and param2 &= REST_OF_STUFF_MASK; 22:03 sapier I've got a msvc2012 build system ...32 bit only 22:03 BlockMen btw, do we release with leveldb? 22:04 sapier hmm in this case I'd have a problem I never built leveldb for windows 22:05 BlockMen i could do both builds (with leveldb aswell), but c55 want others to agree that i do BOTH builds 22:06 BlockMen *or another person that does one of the builds 22:08 sapier is there a reason why you shouldn't do both? 22:08 BlockMen no specific, he wants as many ppl be involved in building on windows 22:09 BlockMen honestly idk what/why he wants 22:09 BlockMen but well, hes the boss 22:09 sapier guess it's to do at least a little bit of checking 22:09 BlockMen checking what? 22:10 sapier checking that windows version behaves correct ;-) 22:11 BlockMen pff^ 22:12 sapier additionally there's another thing we have to decide ... what games to add 22:12 BlockMen first: how many at all 22:13 sapier I guess good old minetest is still set, is it? ... minetest_next is next one to be set, carbone most likely too ... nodetest maybe? and I personally like realtest too but that hastn't been suggested by anyone by now 22:14 BlockMen i think for now we should just add suggested games 22:15 sapier looking at vote and suggestion I guess calling this idea a complete failure isn't most wrong interpretation 22:16 hmmmm could anybody explain this loop to me? https://github.com/minetest/minetest/blob/master/src/sky.cpp#L464 22:17 hmmmm ohhhhhh i didn't see the reference to j in the conditional, nevermind 22:27 sapier imho if blockmen wants to do those builds let him do them 22:29 PilzAdam minetest_game, minetest_next and carbone? 3 times almost the same game? 22:30 sapier we could drop minetest_game 22:30 sapier guess noone would miss it 22:30 sapier and I suggest not to include minimal in release 22:31 BlockMen ^ that 22:31 PilzAdam of course not; almost all worlds depend on minetest_game 22:31 BlockMen minimal needs to be droped in releases 22:31 sapier hmm didn't think about that issue PilzAdam you're right 22:32 sapier still minetest_game is dead 22:32 sapier could minetest_next renamed to minetest_game replace it? 22:32 BlockMen yes 22:32 BlockMen but thats also not best solution 22:32 BlockMen and it could break worlds because jungle is enabled 22:33 sapier I'm just evaluating options right now 22:33 sapier anyone better ideas? 22:34 BlockMen keeping the zombie (_game) until comapitbility break (e.g. 0.5)? 22:35 sapier my now I assumed we'd not break map compatibility ... at least haven't I seen any suggestion that'd require to break it 22:36 BlockMen well, if we want keep compatibility for minetest_game for always we can 1) always include a not longer developed game or 2) reopen _game 22:37 BlockMen but both is not really good IMO 22:37 sapier that's why I asked if minetest_next renamed could replace it 22:37 sapier but right now we need other opinions to this topic 22:38 BlockMen i never saw it myself, but someone always said that jungles can break an existing world 22:46 ShadowNinja sapier: Do you have a Minetest for android APK available? 22:47 sapier not with current version no ... but that's a matter of minutes https://forum.minetest.net/viewtopic.php?f=3&t=9389 here's latest version 22:48 sapier guess I should build current official version one last time prior release 22:53 kahrl sapier: how about an option to only show "recommended" games (i.e. minetest_next but not minimal, minetest_game)? 22:53 kahrl sapier: if this option is enabled the "not recommended" games will only be shown if there is an existing world with that game 22:53 sapier possible but a little bit hard to get done in time ;-) 22:53 kahrl dunno, seems easy 22:53 BlockMen nice idea, kahrl 22:55 sapier well it's easy if we hardcode what is recommended but if you want some more usefull generic way we have to think about how this could work first .. and then agree to a variant of doing it 22:55 kahrl we could even hardcore the names minetest_game and minimal for now and figure out a more general way later 22:55 kahrl there aren't really any compatibility issues there 22:56 sapier if hardcoding is accepted changes most likely are minor true 22:57 BlockMen actually its already hardcoded, but other way round: https://github.com/minetest/minetest/blob/master/src/defaultsettings.cpp#L191 22:57 BlockMen *well, "hardcoded" 23:01 sapier hmm hardcoding would have to be done in gamemgr.lua Line 64 23:02 sapier we'd have to remove those games from list after we got them from server ... but I'm not sure what's gonna happen if you try to configure them 23:02 kahrl wait, are games received from a server now? 23:02 sapier no from core 23:02 sapier sorry 23:03 sapier was a little bit to exciting afternoon today ;) 23:03 kahrl heh 23:03 kahrl how would the user configure them if they aren't in the list? 23:06 BlockMen wait, we have now tooltips. how about tooltip: "not updated anymore" and make the icon half transparent, but letting the games in the list (for configuration)? 23:07 kahrl thing is, the tooltip is only shown a while after you start hovering, so it's easy to miss 23:07 BlockMen i thought its configurable 23:08 BlockMen *the delay 23:08 kahrl ah could be, I didn't look into it much 23:08 kahrl that would solve it yes 23:09 BlockMen oh, only in default settings 23:09 BlockMen not per element 23:09 sapier what about a big fat label diagonal above button? 23:09 sapier LEGACY 23:10 BlockMen or a dialog with info when chosing that game 23:10 BlockMen i thing with 48x48 legacy is not easy to read 23:11 kahrl you have to make sure the dialog is understandable if the user clicks "OK" immediately, by instinct ;) 23:12 sapier maybe we should try the legacy option first 23:12 sapier users are used to just press ok on warning dialogs 23:14 BlockMen true. but what makes you think they notice the label? 23:14 sapier it's different 23:15 kahrl how about a warning in the world creation dialog 23:15 sapier and to be honest I think this is more about telling them later "you'd just have to read" ... and label is way more easy to implement 23:16 ShadowNinja sapier: The DPI scaling is messed up on my cousin's phone. The scrollbars are about 2px and the inventory is bigger than the screen. 23:16 sapier world creation ... wait 23:16 sapier maybe we could just remove them from dropdown 23:16 sapier what phone does your cousin have? 23:17 sapier for what you tell I'd guess some extra high dpi device? 23:18 BlockMen personally i would prefer a dialog when chosing the game in world creation (since there are users neither on forum nor irc) 23:18 BlockMen instead simply removing, but i think world creation is the right place for this 23:19 sapier as mainmenu is already fstk I guess it'd not be a big deal to add this dialog ... but I personally aren't a friend of dialogs noone reads :-) 23:19 ShadowNinja sapier: On my N10 too. (300 DPI) 23:20 ShadowNinja sapier: His is a N5 (high DPI too) 23:20 BlockMen concerning dialogs: irc.minetest.ru/minetest-dev/2014-07-04#i_3794724 :P 23:20 BlockMen the first part 23:20 sapier you could try changing the gui_scaling value ... sadly I can't do anything about the scrollbars those aren't even scaled 23:21 sapier I haven't even seen a setting to scale them in irrlicht 23:22 sapier gui scaling should help in game ... but might cause initial menu to be to big ... guess you've got a small screen too? 23:22 ShadowNinja sapier: Well, you'll have to fix that. 2px-equivalent scrollbars are useless. 23:22 ShadowNinja sapier: My screen's 10 inchs, his is 5. 23:23 ShadowNinja +e 23:24 sapier hmm should work ... guys ... I don't HAVE to fix it ;-) there are other developers that could help too ... still if it's a irrlicht issue there's not really much we can do about it ... no adding our own implementation is crap it's gonna be broken for any new irrlicht version 23:24 sapier so let's hope irrlicht does support scaling it somehow ... still it's gonna be for next version 23:24 sapier or some bugfix android release 23:25 sapier I'm quite sure it's not gonna be the only issue popping up once android versions are spread 23:35 ShadowNinja sapier: The "Config MODs" page is very squashed too. 23:37 sapier would there be more room do display? 23:46 ShadowNinja sapier: Yes, loads of it. 23:47 sapier can you post screenshots? maybe I can change the automatic scaling factor to better fit screens like that 23:48 ShadowNinja sapier: http://i.imgur.com/C5eBNLV.png 23:49 sapier oops 23:49 sapier that's strange 23:49 ShadowNinja sapier: And the "Minetest 0.4.9-123-g0123456" is cut off. 23:50 sapier yes I see it 23:51 sapier I'll do a new build for you writing dpi information to debug.txt ... this is way to small ... do you have gui_scaling set manually? 23:51 Taoki GUI scaling option doesn't seem to do anything for me. Or is it for the Android version only? 23:51 sapier should work everywhere (except üc main menu) 23:51 Taoki sapier: Also, I saw your commit about the Android enabled native build. Where can I get Minetest for ARM from? I have two devices, and testing it is the least I could do\ 23:52 sapier nowhere 23:52 sapier android yes native arm no 23:52 Taoki Yeah, needed to restart Minetest. Had the impression it would be applied automatically :P 23:57 Taoki Any chance of it ever ending up on Google Play? 23:57 Taoki It's easiest way to install, and would likely make us yet even more popular! 23:58 sapier yes but not for 0.4.10 ... hopefully 0.4.11 will be good enough 23:59 Taoki Awesome