Minetest logo

IRC log for #minetest-dev, 2014-07-08

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
00:05 Taoki What the? In minetest.sound_play, toplayer is being ignored
00:05 Taoki A sound that was supposed to be triggered for another player was triggered for me
00:05 Taoki Any idea what is wrong?
00:06 Taoki minetest.sound_play(def.ambience, {toplayer = name, loop = true})
00:06 Taoki That didn't loop for player with name "name", also for me
00:09 BlockMen its "to_player" https://github.com/minetest/minetes​t/blob/master/doc/lua_api.txt#L219
00:09 Taoki oh
00:09 Taoki My mistake then, sorry
00:10 Taoki BlockMen: It takes a player name as an argument also, right? Not necessarily the player entity
00:11 Taoki Sorry, the API said that already >.>
00:22 Taoki celeron55: Something caused your server to shut down. I could only see the **Server shutting down** message as an user. But from what I know, that typically happens when it crashes.
00:22 Taoki It wasn't my mod though, honest! It's his fault *points to random other dev* (don't think it was Creatures though, but if it was let me know the error please)
00:23 BlockMen left #minetest-dev
02:46 ShadowNinja sfan5: So, the _game rules are largely identical to the core rules.  Mostly just approval changes by the look of it.
03:06 alexxs joined #minetest-dev
03:25 kaeza joined #minetest-dev
03:37 cg72 joined #minetest-dev
04:10 Megaf well, I managed to compile minetest for android here, thing is, I think it compile only the ARM version, how and where I change the CPU architecture I want to support?
04:14 Piggybear87 joined #minetest-dev
04:24 Piggybear87 left #minetest-dev
05:10 hmmmm I'm wondering how beneficial it would be to check for the presence of a block in the db using a bloom filter
05:11 hmmmm could somebody who has LevelDB and/or Redis check the average amount of time it takes to check the presence of a block (e.g. average time a loadBlock failure takes)
05:11 proller joined #minetest-dev
05:12 hmmmm if a similar structure is constructed by the databases and maintained in memory, it wouldn't be beneficial, and instead i should use the database to check for the presence of a block in the db
05:13 hmmmm it'd also be nice to find the amount of blocks popular servers have in their maps.
05:16 hmmmm according to wikipedia, m = (n * ln p) / (ln 2)^2, where p = desired probability, n = number of items, m = optimal array size
05:16 hmmmm given a sane value of n, i can optimize for the perfect compromise between speed and memory usage
05:41 Hunterz joined #minetest-dev
05:57 nore joined #minetest-dev
06:17 LemonLake joined #minetest-dev
06:29 Garmine joined #minetest-dev
06:29 celeron55 Taoki: i didn't have yet time to make a restart script; i'm now making one and it will be soon up again
06:35 celeron55 up
06:36 celeron55 copied error message on #minetest (a bug in worldedit, probably)
06:37 Guest28706 joined #minetest-dev
06:57 Garmine joined #minetest-dev
07:07 sfan5 ShadowNinja: yes
07:17 BlockMen joined #minetest-dev
07:18 BlockMen ShadowNinja, well, sfan5 forgot to mention that _game is maintained (for now) just by nore, him and me.
07:18 * BlockMen adds that to wiki
07:27 VanessaE I'm starting to see the occasional "ERROR[MeshUpdateThread]:  Waiting for texture xxxxxxxx timed out." (always seems to be some normalmap, and always different each time)
07:27 VanessaE someone asked the other day if others had seen this happening.  I had not...  until today.
07:27 VanessaE with these timeouts come HUGE mesh generation lag spikes
07:33 VanessaE (the lag spikes started earlier, but today's the first time I noticed the timeout message)
07:46 BlockMen left #minetest-dev
08:05 PenguinDad joined #minetest-dev
08:07 RealBadAngel VanessaE, i will move that file access outta meshupdate, same as i did with shaders
08:11 celeron55 could we make a second official download recommended to newcomers which contains carbone instead of minetest_game? does anyone else than me feel that this is a good idea?
08:12 celeron55 granted, it doesn't have much deeper gameplay, but it generally looks better and has more stuff, and isn't broken
08:13 VanessaE RealBadAngel: ok
08:13 celeron55 (and sounds better)
08:13 VanessaE celeron55: seems fine to me; I haven't played it, but from the screenshots, it looks like a good game.  for sure, I like the HUD :)
08:14 celeron55 and it's also mostly compatible with minetest_game (most mods work)
08:14 RealBadAngel whats wrong with one download and many games included?
08:15 VanessaE celeron55: there's one caveat:  keeping that download up-to-date with upstream Carbone.
08:15 celeron55 RealBadAngel: it doesn't make sense to me because minetest_game and carbone overlap so much
08:16 celeron55 and there is nothing in the engine or in the games to help users understand why there are two and how they differ and why it is so
08:16 celeron55 and the name of the other one is "Minetest" so which are the ones not knowing anything going to choose?
08:16 celeron55 minetest_game of course
08:16 celeron55 so it completely beats the purpose
08:17 RealBadAngel well, there are tooltips, there could be added text area with more detailed game descriptio, kinda "about"
08:17 celeron55 yes, if there was an about screen, it could make me think otherwise
08:18 celeron55 one that new users will see without clicking anything
08:18 RealBadAngel thats pretty easy to add
08:19 RealBadAngel i can agree on carbone and mt being pretty the same, but there are also games to come totally differnt, and those definitely will require some about and how-to
08:19 celeron55 simply anything that is bundled in the official package must make sense to new users - it's the main criterion, because old users know how and where to download a mod or a subgame and what they are and why they would do so and what exacty they can find in them that isn't found in minetest_game
08:20 smoke_fumus joined #minetest-dev
08:20 celeron55 RealBadAngel: that's kind of a different thing, because a howto for eg. nodetopia or realtest isn't going to fit in an "about" box 8)
08:20 VanessaE hah
08:20 RealBadAngel at least simplified
08:21 VanessaE celeron55: a howto for Realtest wouldn't fit in a FORUM post for that matter :P
08:21 RealBadAngel i have creatures by taoki in mind
08:21 celeron55 of course it could say "you don't know how to play this; please use google now"
08:21 RealBadAngel text area could display pages, thats not problem, just flip the pages if text is bigger
08:22 sfan5 nobody wants to read a 7 page howto in the main menu
08:22 celeron55 well, make a prototype for such a thing with new users in mind and show it here
08:22 celeron55 let's see if we can make it good
08:22 RealBadAngel sfan5, ofc most wont read it, but then you can just say rtfm ;)
08:22 nore well, some kind of in-game tutorial could be made
08:23 nore detecting when a player has "completed" a step, and displaying tutorial for the next one
08:23 celeron55 nore: that's not important for now; the game about text in the menu is important
08:23 RealBadAngel nore, that seems to be related to achievements rather
08:24 nore RBA: it's between a tutorial and achievements :)
08:24 nore celeron55, so in the "about" text, there should be a short description of the game?
08:25 RealBadAngel but info "use pick to mine stone" in main menu would be more than stupid ;)
08:25 celeron55 nore: yes, optimized for new users to understand the difference between minetest_game and carbone
08:25 celeron55 that should be good for many other things too then
08:28 RealBadAngel btw, one could start adding tooltips for main menu
08:29 RealBadAngel that is definitely needed imho, labels and text on buttons is sometimes just not enough
08:31 RealBadAngel i cant for example put in "bumpmapping' label an explanation whats that and what it needs to work
08:34 celeron55 tooltips are fine too, but not that useful for the purpose of making new users understand multiple similar subgames
08:34 RealBadAngel ofc
08:35 RealBadAngel btw, im thinkin where to put button to read info buout subgame
08:35 RealBadAngel any ideas?
08:35 RealBadAngel in world configuration menu?
08:36 celeron55 it might require menu redesign, now that i think of it
08:37 celeron55 like, when opening MT, ideally the user should be faced with a choice between connecting a server, and choosing between locally installed subgames
08:38 celeron55 and then the choice is locked in the next screen
08:38 celeron55 (with a back button)
08:39 celeron55 this is what everyone is already really doing, but the menu just doesn't represent it well at all
08:40 celeron55 the "connect to server" option would have a hardcoded about box explaining it, and the games would have whatever they specify in a file added for this purpose
08:42 RealBadAngel an idea, i could display info bout game on double click its icon
08:42 RealBadAngel then above the games bar open textarea with back button
08:43 celeron55 oh god no, nobody will know how to do that
08:43 RealBadAngel tooltip?
08:44 celeron55 i'm going to say nothing else would really be user-friendly except that i just described
08:44 celeron55 s/that/what/
08:45 VanessaE bbl
08:56 RealBadAngel celeron55, http://i.imgur.com/r2TpDP6.png
08:57 celeron55 the issue is, new users probably won't go and hover their mouse over those icons
08:57 celeron55 so they just end up with whatever game happened to be selected and be like "oh, this is minetest i guess"
08:57 Calinou joined #minetest-dev
08:58 celeron55 they have to be faced with an explicit choice, then they will take the 5 seconds to read the about box and immediately see what the choice is about
08:58 RealBadAngel its already worthy to click that button, it changes backgroud, could play also subgame music
08:59 celeron55 you don
08:59 celeron55 *you don't understand
08:59 sfan5 new users don't know they can click those buttons
08:59 sfan5 +might
08:59 sfan5 not*
09:00 RealBadAngel oh cmon, users are not as that stupid, then can see those buttons ;)
09:00 celeron55 this is the first thing a user sees in minetest; it must give them as good a start as possible
09:01 celeron55 we can slack in UI design elsewhere
09:04 RealBadAngel so you mean game button bar because user may not know they can click it?
09:04 RealBadAngel *is unusable
09:05 RealBadAngel so what for we do have it?
09:05 sfan5 for people who know how to use it?
09:06 RealBadAngel remove ugly lookin game selection pop up list, and make game button bar common for windows where user can select game
09:07 RealBadAngel that would be way nicer
09:07 kaeza joined #minetest-dev
09:21 alexxs joined #minetest-dev
09:22 celeron55 no, the game selection should only be in a single place
09:22 celeron55 and the place is "at first, before the other UI is entered"
09:44 alexxs joined #minetest-dev
10:23 RealBadAngel celeron55, http://i.imgur.com/GVMSMDC.png
10:23 RealBadAngel what do you think about such way?
10:23 celeron55 those icons are horrible
10:23 sfan5 aside from the icons being horrible it looks ok
10:24 RealBadAngel icons are quick find, just examples
10:24 RealBadAngel but, thx to that we can have lotsa space, tooltips ofc will explain them better for those who cant think ;)
10:25 RealBadAngel also, another buttons like gamebar will become obvious they are clickable too ;)
10:26 celeron55 this does not fix the UI flow problem
10:27 RealBadAngel but allows to add extra button with ? next to another world related buttons
10:28 RealBadAngel and there will be still lotsa place
10:29 RealBadAngel not to mention such buttons would be way better than current ones for android menus imho
10:56 proller joined #minetest-dev
11:07 ImQ009 joined #minetest-dev
11:35 proller joined #minetest-dev
11:39 casimir joined #minetest-dev
11:54 proller joined #minetest-dev
12:00 PilzAdam joined #minetest-dev
12:08 proller joined #minetest-dev
12:10 Megaf_ joined #minetest-dev
12:11 RealBadAngel i do have one thought about that:
12:11 RealBadAngel os<<tsrc->getTextureName(p.t​ile.texture_id)<<"^[crack";
12:11 RealBadAngel this means that we are producing new textures by adding cracks to it??
12:12 RealBadAngel so each texture is multiplied by number of crack frames?
12:12 celeron55 yes; also it causes the mapblock mesh to be regenerated
12:13 celeron55 it's crude but works well in practice
12:13 RealBadAngel but are those textured stored or dropped?
12:14 celeron55 no; minetest has no mechanism for determining whether textures have references to them and it cannot know whether any of them can be safely dropped
12:14 RealBadAngel so thats why our textures demands so much ram, yes?
12:14 celeron55 probably no
12:15 celeron55 you won't be digging most kinds of nodes in a single game session
12:15 RealBadAngel it becomes a real problem for HD textures
12:15 RealBadAngel it multiplies them by 16
12:16 celeron55 of course you can fix it, but it's not going to be easy
12:16 RealBadAngel i think i do have a solution
12:16 celeron55 well it's easy for the shader implementation
12:16 RealBadAngel same as i used for glasslike framed
12:16 celeron55 the other one you probably want to skip
12:16 RealBadAngel add separete mesh with offseted slightly textures
12:17 celeron55 i consider that a worse hack
12:17 RealBadAngel it will work without problems and for sure will remove need to multipy textures
12:18 RealBadAngel and it doesnt look like a hack, just copy vertices of dug node, add them to meshbuffer, with other textures
12:18 celeron55 make sure it works well for partially transparent things like grass
12:18 celeron55 the current implementation is really good for those
12:19 RealBadAngel im rewriting now mapblock_mesh anyway
12:19 RealBadAngel most of things that could be done once (mostly shader stuff) got kicked off there
12:20 celeron55 oh god
12:20 celeron55 i'm so sure there will be awful regressions that you are too lazy to fix
12:20 RealBadAngel rather not
12:22 RealBadAngel last time problems were so hard to fix because there were 4 pieces of code trying to do the same. Plus fucked up open source drivers.
12:42 vifino joined #minetest-dev
12:45 * VanessaE wonders how the result will perform
13:06 RealBadAngel definitely it will use less memory
13:06 RealBadAngel btw, when digging particles are here, who needs cracks? will code it, its just a side note
13:12 kaeza joined #minetest-dev
13:14 VanessaE RealBadAngel: frankly I like seeing both :)
13:15 VanessaE however, for many people, particles perform poorly, which is probably why they can be turned off in the menu I guess.
13:15 RealBadAngel you dont like current cracks, like really ;)  how big grass and dirt texture HDX has?
13:15 VanessaE wat
13:15 RealBadAngel ok, lets say just stone, how big it is?
13:15 VanessaE Grammar FAIL
13:18 RealBadAngel lets see, dirt is 512kb, how many frames crack has? 16?
13:19 RealBadAngel so once you dig a single dirt node, it will 8704kb in memory
13:19 RealBadAngel *take
13:19 sfan5 VanessaE: where was the world I could reproduce this with again? https://github.com/minetes​t/minetestmapper/issues/8
13:20 SmugLeaf joined #minetest-dev
13:20 SmugLeaf joined #minetest-dev
13:21 VanessaE sfan5:  http://digitalaudioconcepts.com/vanessa/hob​bies/minetest/worlds/Creative_World.tar.bz2
13:21 VanessaE warning, that's a big file
13:21 sfan5 wasn't there a smaller one?
13:21 sfan5 ..
13:21 sfan5 wasn't it the nostalgia map?
13:21 Megaf_ Hi everyone
13:21 VanessaE the Nostalgia map might be enough too, s/Survival/Nostalgia/
13:21 * Megaf_ roars
13:21 VanessaE it's about 258 MB
13:22 VanessaE fark, s/Creative/Nostalgia/  rather :)
13:22 VanessaE http://digitalaudioconcepts.com/vanessa/hobb​ies/minetest/worlds/Nostalgia_World.tar.bz2
13:22 CraigyDavi`` joined #minetest-dev
13:22 * Megaf need to learn to see the chanel he is talking in
13:22 VanessaE really any map that exceeds ±2048 in the Z direction should do it.
13:22 rubenwardy joined #minetest-dev
13:25 Megaf 1.3Gmap.sqlite
13:28 Megaf It seems like maps dont get as big now a days as they used to get
13:28 Megaf Was that some optimization or something?
13:29 Megaf I had a server with a 5 GB map
13:30 VanessaE Megaf: my worlds are all leveldb, which is why they're notably smaller than they used to be
13:30 VanessaE maybe levelsb uses some kind of compression, or just that it's database format is more efficient.  idk.
13:30 VanessaE leveldb*
13:32 sfan5 VanessaE: GD Warning: product of memory allocation multiplication would exceed INT_MAX, failing operation gracefully
13:32 sfan5 :(
13:33 sfan5 damn libgd
13:33 VanessaE eep
13:33 Megaf VanessaE; indeed leveldb use compression by default, yes
13:33 Megaf VanessaE; but even sqlite maps seems to be smaller
13:33 sfan5 INT_MAX is only 4GB
13:33 sfan5 I have more then 4GB memory..
13:36 sfan5 than*
13:37 sfan5 VanessaE: can't reproduce, can you?
13:37 VanessaE I haven't tried
13:37 VanessaE not since the initial report, I went back to using a different branch just to make sure I had good imagery
13:37 VanessaE different fork(
13:38 VanessaE close it for now, if it crops up again, I'll report it.
13:38 sfan5 I would like you to try too
13:38 VanessaE lemme see..
13:39 sfan5 VanessaE: https://cdn.mediacru.sh/oJd1LQNj6hDE.png
13:39 VanessaE hm
13:39 VanessaE I seem to be using the official fork already
13:39 VanessaE and the maps generated with it are good
13:39 VanessaE what commit am I at?  hm.,
13:40 sfan5 <MinetestBot> GIT: sfan5 commited to minetest/minetestmapper: Fix --drawalpha e85a498 2014-07-08T15:08:07+02:00 http://git.io/dajduw
13:40 sfan5 that one probably fixed it
13:40 VanessaE commit dff45891521b8fd299f68f1ca1c36124e3b6616b
13:40 sfan5 actually let me try
13:40 VanessaE is where I'm at.
13:40 EvergreenTree joined #minetest-dev
13:41 VanessaE I must have bisected it back to this point and stayed here.
13:41 Jordach joined #minetest-dev
13:41 sfan5 hm, no
13:41 VanessaE *scratches head*
13:41 sfan5 it does not happen even with the commit before "Fix --drawalpha"
13:41 VanessaE odd, because I'm also not using --drawalpha
13:42 sfan5 I tried both with and without
13:42 VanessaE well lemme update to HEAD and just see what happens.
13:44 * VanessaE adds --drawalpha and runs the mapper script...
13:48 VanessaE wat
13:48 VanessaE --backend  is no longer recognized?
13:49 VanessaE guess it isn't needed anyway
13:49 VanessaE perhaps you should just ignore that switch rather than fail with an error.
13:58 VanessaE ok, it seems good now.
13:58 VanessaE if you reload the Creative World map from the main link, http://digitalaudioconcepts.com/vanessa/hob​bies/minetest/worldmaps/Creative_World.png , it looks good.
13:58 VanessaE and that alpha in the water is just plain beautiful
13:58 VanessaE kodus, sfan5
13:59 VanessaE kudos, too :)
13:59 sfan5 :)
13:59 sfan5 <VanessaE> and that alpha in the water is just plain beautiful
13:59 sfan5 this why I added that feature
14:01 hmmmm joined #minetest-dev
14:06 VanessaE I'd call it good, sfan5.
14:15 rubenwardy joined #minetest-dev
14:16 troller joined #minetest-dev
14:24 BlockMen joined #minetest-dev
14:24 BlockMen https://github.com/minetest/minete​st/pull/1328#issuecomment-48342751
14:25 LemonLake joined #minetest-dev
14:26 BrandonReese joined #minetest-dev
14:30 VanessaE BlockMen: as long as it's a new drawtype with new texture filenames, I'm okay with it.
14:30 BlockMen VanessaE, yes, as new drawtype
14:31 VanessaE _with new filenames_
14:31 sfan5 new filenames?
14:31 sfan5 you mean for the textures?
14:32 VanessaE yes
14:32 sfan5 this has nothing to do with the engine changve
14:32 sfan5 change*
14:32 LemonLake joined #minetest-dev
14:32 VanessaE sfan5: well from an engine standpoint no, but when it gets applied to the default game, which it will, that's where I'm talking about.
14:34 BlockMen VanessaE, if you want support the new drawtype you need other textures for your pack anyway?
14:34 BlockMen sfan5, nore: what you think about removing this line: https://github.com/minetest/minetes​t_game/blob/master/minetest.conf#L4 ?
14:35 sfan5 BlockMen: if it fixes the not-being-able-to-jump-place annoyance, yes
14:35 VanessaE BlockMen: correct.  I just want to avoid a style conflict.  Clearly what works for a flat torch won't work for a 3d torch in most cases.
14:38 BlockMen sfan5, yes it does. pushing it then
14:48 nore joined #minetest-dev
14:50 asl joined #minetest-dev
14:57 BlockMen VanessaE, ok. new drawtype and other texture name
14:57 VanessaE good :)
14:58 BlockMen more comments on #1328 ?
14:58 VanessaE hyave you decided how you're gonna handle the flames/animation?
14:58 VanessaE have*
14:59 BlockMen by texture. you can set a not animated texture when someone finds a nice way for other kind of flames, e.g. particles
15:00 VanessaE ok
15:00 VanessaE gonna do four separate sides for the texture?
15:01 VanessaE (well four parts of one image, UV map)
15:03 BlockMen i prefer a simple image instead uv nodes. that could be done if we get meshnodes IMO
15:04 BlockMen *uv nodes = uv maps ^^
15:04 VanessaE well I mean mapped like a fencepost is
15:04 VanessaE you know, 1/4 of the image on one side, another 1/4 on the next side, retc.
15:04 VanessaE etc*
15:05 VanessaE so that the burning flame isn't copied around all four sides + top/bottom ;)
15:05 VanessaE it'd be rather nice to make all four sides of the torch look different.
15:06 BlockMen no. doing that you cant make that for example http://i.imgur.com/e5CGz.png
15:06 BlockMen (the metal on the top of the torch)
15:06 BlockMen for fences it might be nice, but for the torch its done better this way
15:06 VanessaE sure you can
15:07 BlockMen you only have 1/4 for that. with 16px you cant
15:07 VanessaE make the default texture 32px then
15:07 VanessaE there's no law that says it HAS to be 16px is there?
15:07 BlockMen so everything is 16 px and just torches 32px? no
15:07 VanessaE especially in a case like this one
15:08 nore joined #minetest-dev
15:08 BlockMen let the uv maps for meshnodes, not this
15:08 rubenwardy joined #minetest-dev
15:08 VanessaE I dunno...
15:09 VanessaE then give multiple images per side
15:09 VanessaE you know, like any other node
15:09 VanessaE er that reads wrong.
15:09 VanessaE give 6 images, like any other node I mean
15:09 VanessaE it'll keep the 16px resolution for default and give maximum flexibility for texture pack artists.
15:09 VanessaE won't add more than 1k or 2k or so to the game
15:10 BlockMen 3 dirs would make sense
15:10 BlockMen top bottom sides
15:10 VanessaE four.  top, bottom, N/S, E/W.
15:11 BlockMen different textures on side?
15:11 BlockMen we are talking aabout a torch drawtype ;)
15:11 VanessaE yeah, because at least the flames could be mirrored
15:11 BlockMen they are already mirrored
15:11 VanessaE oh
15:11 VanessaE hm.
15:12 VanessaE well at least make it so that the drawtype can be given a full 6-texture treatment
15:12 VanessaE even if the GAME doesn't do it
15:12 VanessaE then we can decide later if more than 3 textures are needed.
15:24 troller joined #minetest-dev
15:34 RealBadAngel VanessaE, here?
15:34 troller joined #minetest-dev
15:34 VanessaE yes
15:37 Piggybear87 joined #minetest-dev
15:37 Piggybear87 left #minetest-dev
15:37 casimir Blockmen, does this replace 2D torches? I don't understand the code.
15:38 sfan5 casimir: no, that code just adds a 2d torch drawtype IIRC
15:38 BlockMen currently yes, but since that wont go in i will change it if other agree to merge (in general)
15:38 BlockMen sfan5, no
15:39 BlockMen casimir, since you use fm, look there
15:39 BlockMen they took my code and gave credit someone else
15:41 casimir I don't use fm, but I'll take a look.
15:44 troller joined #minetest-dev
15:46 Calinou joined #minetest-dev
15:48 BlockMen casimir, oh..ok. had in mind your game supports fm
15:49 Eater4 joined #minetest-dev
15:55 sapier joined #minetest-dev
16:00 grrk-bzzt joined #minetest-dev
16:03 sfan5 wtf
16:04 sfan5 wtf
16:04 sfan5 each Database class basically duplicated the code from ServerMap::loadBlock(std::string *blob, v3s16 p3d, MapSector *sector, bool save_after_load)
16:04 sfan5 who designed this shit?
16:04 sapier I don't know who designed it but I know who committed ;-)
16:05 sfan5 I commited it, and?
16:05 sapier Megaf have a look at top of android makefile
16:05 sapier ok then I meant who wrote it
16:06 sapier VanessaE Waiting for texture xxxxxxxx timed out should actually be a assert cause
16:06 sapier if this happens memory may or may not be corrupted
16:06 sfan5 sapier: don't you dare to make that an assert(0), it happens when using valgrind too
16:07 sapier If I'd wanted to do that I already would've done still if you see this anything happening after it might be result of it
16:07 VanessaE sapier: see what RBA said about it earlier.
16:07 zat joined #minetest-dev
16:08 VanessaE it's partly due to normalmap file access while in the middle of a mesh update.  it'll be moved out of mesh generation to somewhere else.
16:08 sapier it's just crap to abort an action having passed a stack pointer to some other thread before that pointer returns
16:09 sapier http://i.imgur.com/GVMSMDC.png imho that's extremely ugly
16:10 VanessaE ew.
16:10 sfan5 <RealBadAngel> icons are quick find, just examples
16:10 VanessaE the icons have to fit the theme, too
16:10 VanessaE those don't.
16:11 sapier then it's a quite stupid example
16:11 sapier as this featuee comes and falls with icon quality
16:11 VanessaE yes
16:12 VanessaE but if the icons are clear, simple, and they fit the grey/black/green theme, then they'll work fine
16:12 sapier and increase main menu height
16:12 VanessaE those icons look like the belong in something like LinCity or something.
16:13 sapier ok they don't increase
16:13 sapier problem with bitmap icons is they always have to be scaled
16:13 VanessaE use SVG ;)
16:13 sapier tell irrlicht
16:14 sapier and usually you need tooltips
16:14 sapier no offence RealBadAngel but your inventory is a perferct proof for this
16:14 VanessaE really though, if scaling is an issue, there's only one solution and that's oversized icons.  No, I think it may be better to stick to buttons with text.
16:16 Hunterz joined #minetest-dev
16:18 sapier https://github.com/minetest/minetest/issues/1455 is because of mainmenu set to fixed size. For what I remember that "workaround" was only there because of formspecs didn't scale correct by that time. So what about dropping that strange fixed size formspec feature and just let mainmenu scale as it's supposed to ?
16:18 sapier ok to be precise as all other formspecs do
16:26 LemonLake joined #minetest-dev
16:30 RealBadAngel sapier, those icons were about to show it can be readable, using less space and gaining more space
16:31 RealBadAngel it was not about my f**ing taste ;)
16:31 RealBadAngel i just asked google for some icons
16:31 sapier well I wouldn't understand the "play" button if I didn't know there was one before
16:32 sapier I've never seen the triangle being used for starting a game
16:32 RealBadAngel think about another one then
16:32 sapier to me the triangle is playing audio or video only
16:33 Megaf sapier; Not really, Thanks
16:33 sapier but of course that's personal as any image associations are
16:33 RealBadAngel all can think about it
16:33 celeron55 i think that text-based menu buttons are the best we can have; it's stupid to consciously make things harder to understand
16:33 celeron55 most games have text-based buttons
16:33 RealBadAngel point is that text on buttons are wrong
16:33 celeron55 why? well maybe because they work
16:33 celeron55 in their menu
16:33 RealBadAngel most of translations doesnt fit the provided size anyway :P
16:34 celeron55 why are people discussing these stupid icons instead of my best ever menu modification?
16:34 sapier hmm maybe because that one wasn't even noticed ;-)
16:35 VanessaE celeron55: screenshot of said menu? :)
16:35 celeron55 http://irc.minetest.ru/minet​est-dev/2014-07-08#i_3802944
16:35 celeron55 VanessaE: only description
16:35 VanessaE hm
16:36 VanessaE you know, that's not such a bad idea
16:36 sapier sounds like apples workaround single thread pardigma and microsofts metro gui
16:36 celeron55 i think MT needs that before it can bundle multiple games in releases
16:36 RealBadAngel btw, tellin that folks wont use image_buttons because they wont know if theyre clickable is most stupid answer ever
16:36 VanessaE RealBadAngel: people are afraid to click.
16:37 VanessaE believe it or not, as stupid as that sounds.
16:37 RealBadAngel sure
16:37 RealBadAngel all the modern OSes are not clicklable then
16:37 RealBadAngel because using of icons
16:37 VanessaE my mom, G*d rest her, used to be afraid of confusing the computer from clicking stuff she didn't understand
16:37 sapier basicaly celerons suggestion is increasing menu depth in order to reduce width (I don't talk about physical dimensions)
16:37 RealBadAngel ROTFL
16:37 celeron55 sapier: yes, you can describe it like that
16:38 sapier as it's formspec there's no reason not to do it
16:38 celeron55 the current menu has zero depth and that makes it kind of convoluted
16:38 VanessaE c55's idea is sound.
16:38 VanessaE too much shit on one screen can confuse people
16:39 celeron55 and i'm not thinking about this only for new users; i would prefer it myself too
16:40 sapier on the other hand switching from one mode to another is less comfortable for experienced users
16:40 sapier but I guess this depends on how good it is made
16:41 celeron55 changing the game should be a matter of hitting esc and clicking another one
16:41 VanessaE now the question is, who gets to refactor the main menu? ;)
16:41 sapier twice escape to get to multiplayer mode
16:42 VanessaE sapier: unintuitive.
16:42 celeron55 that doesn't make sense
16:42 sapier I wont :-) I'm still recovering from last time
16:42 celeron55 but for example "m" in the first screen could go to multiplayer
16:42 celeron55 and 1...9 could be games
16:42 sapier then you need more then mainmenu refactoring
16:42 celeron55 then you can just hit those quickly when you get used to it
16:42 celeron55 well it's something to consider in the future, not a priority
16:43 sapier well someone already asked for having access to all keys from formspec
16:43 RealBadAngel since its lua, and VanessaE has good taste, imho she shall do that
16:43 VanessaE fuck no
16:43 celeron55 lol
16:43 VanessaE I suck at formspecs.
16:43 sapier well if noone wants to do it it's not worth discussing ;-)
16:44 sfan5 VanessaE: no, formspecs suck
16:44 VanessaE sfan5: +1
16:44 sapier that complaint is as old as minetest
16:44 sapier almost
16:44 celeron55 i might try it, but i probably have to prioritize it under a million other things so in practice i won't in the near future
16:44 sfan5 sapier: before you say anything, no I'm currenctly making the Database class not suck; I don't have time to develop better formspecs
16:45 celeron55 could we recruit some non-coredev modder to do it? :3
16:45 sapier I wont tell anyone to replace formspecs because imho they're by far not most important point
16:45 RealBadAngel so maybe different way, VanessaE, could you propably make a set icons we could use? just as proposal, we could see if it could look a bit better
16:45 sapier they do what they're supposed to do ... that's not true for other components
16:47 RealBadAngel its not about formscpecs by now, we already know how to use them
16:47 RealBadAngel now its point how to make clear and nice lookin menu
16:47 RealBadAngel without making players being too scared to click something :)
16:48 sapier why do ppl always wanna do the "visible" stuff only (exept sfan5 right now) ?
16:48 sfan5 RealBadAngel: players are too scared to click things because double clicking somewhere would exit the menu
16:48 * sfan5 looks at sapier
16:48 sfan5 that damn gesture is not needed for PC
16:49 sapier ok next questions why do ppl complain after release if they coudl've complained 6 weeks before?
16:49 sfan5 it only annoyed me today
16:50 celeron55 releases come and go, who cares
16:50 sapier it's in there for about 6 weeks if someone would've told adding a setting would've been a matter of 5 min
16:50 sfan5 <celeron55> [...] who cares
16:50 sapier yes once a year celeron55 ;-)
16:50 celeron55 well... okay, they *should* come and go
16:51 sapier maybe we get twice this year ... anyone against scheduling next release for end of novemer?
16:51 sapier (hoping to get it done till christmas)
16:53 celeron55 i would suggest a month or two earlier
16:54 sapier end of october?
16:55 sapier ok 1st november is saturday, I set milestone 0.4.11 to this date
16:56 Calinou that's about 4 months from now, sounds OK
16:56 Calinou shouldn't be too long to keep community alive
16:57 Calinou I think the ideal is every 3-8 weeks (what Minecraft Alpha/Beta did as well as earlier Minetest versions), but then you need to feature freeze more often
16:57 sapier sfan5 if you want doubleclick changed plz write a issue
16:57 sfan5 sapier: later
16:57 sapier we can't compete to minecraft they have way more manpower
16:58 Calinou current versions of Minecraft have releases every 3-6 months…
16:58 Calinou yes we can.
16:59 Calinou we don't need to pack 15 large new features with each release
16:59 Calinou even having one medium-size feature and a bunch of fixes and tweaks at every release is good enough
16:59 sapier our problem isn't writing features but testing them
17:00 sapier especially the background changes are hard to get done
17:02 BlockMen Calinou, nodeboxes are not usable for wallmounted 3d torches
17:05 us`0gb joined #minetest-dev
17:06 Calinou why not? it's predicted client-side and is not any heavier
17:06 Calinou this is fully Lua, no C++ change required
17:08 troller joined #minetest-dev
17:10 celeron55 yes the issue with fast releases is feature freezes when software grows and it needs more time for testing and fixing
17:10 celeron55 things like web browsers have multiple versions being developed at the same time so that they can have a steady flow of releases despite the huge workload and delay
17:11 celeron55 no feature freeze in the development branch
17:11 celeron55 just a snapshot and then fixes in a separate to-be-stable branch
17:11 troller joined #minetest-dev
17:11 sapier that'd require ppl to do fixes too
17:12 RealBadAngel BlockMen, think about extruding torch and use it as model
17:12 celeron55 we don't have so many developers that we could be making new features and fixing things in an another branch at the same time
17:12 RealBadAngel thist would work for any texture given
17:13 rubenwardy joined #minetest-dev
17:14 BlockMen RealBadAngel, IMO it doesnt look that good. furthermore a blocky torch makes sense in a block world ;)
17:14 sapier and to be honest I'm a little bit disapointed about last release, there's beeon only hmmmm and celeron who did help finding the database issue all other bugs have been left untouched
17:14 sapier and there are some other core devs who could've fixed in between
17:15 sapier +them
17:15 celeron55 well i also fixed the lighting issue that paramat reported
17:15 celeron55 but yes, people weren't very active
17:15 sapier yes and you did help for the other bug too
17:15 RealBadAngel BlockMen, you pick a texture pack and drop whole stack of torches on the ground
17:15 RealBadAngel you will see 3d model
17:16 celeron55 bug fixing is idel for large teams to do
17:16 BlockMen RealBadAngel, ik what you mean
17:16 celeron55 features not
17:16 celeron55 ideal*
17:16 sapier as of 30th june there's been no bugfix form team (except docs)
17:16 sapier oh 2nd july
17:16 BlockMen but IMO torches should be blocky node, not a texture with deepth
17:17 RealBadAngel but that approach makes 3d model out of any texture
17:17 RealBadAngel if you want better one just make a stick 3d and use irrlicht particles to make fire
17:17 celeron55 "this device makes nuclear waste from absolutely anything you feed in!"
17:17 celeron55 it's not always such a great idea 8)
17:17 sapier and with this support I really fear doing a release branch could result in "me or noone" fixing rrelease branch bugs
17:19 BlockMen RealBadAngel, my torches pull dont hinder anyone from adding irrlicht particles (and use them as flames)
17:19 BlockMen it will be still useable/extendable with that
17:20 RealBadAngel if done 3d, should come with particles
17:20 RealBadAngel do it right or dont touch
17:20 BlockMen ^ y? are we minecraft?
17:20 RealBadAngel no, they dont have such nice particles
17:20 BlockMen they use particles for flaes
17:20 BlockMen *flames
17:20 RealBadAngel have you seen irrlicht ones?
17:21 BlockMen yes
17:21 BlockMen implement it if you want it
17:21 BlockMen also those can be added later aswell
17:21 RealBadAngel ok
17:22 RealBadAngel btw, i posted yesterday some pics, but not here
17:22 RealBadAngel http://i.imgur.com/T8wf2tp.png
17:22 RealBadAngel http://i.imgur.com/b3usu4B.png
17:24 sapier sfan5 do you have experience with sqlite3?
17:25 sfan5 RealBadAngel: add shadows, then everyone will love you
17:25 sfan5 sapier: no
17:25 sapier sad ... I'd really wanna know why REPLACE fails on android
17:25 RealBadAngel btw, irrlicht lights can be of any color
17:25 sfan5 ask the sqlite3 devs
17:25 sfan5 19:25:44: ERROR[EmergeThread0]: Invalid block data in database (-122,1,-32) (SerializationError): MapBlock::deSerialize: no enough input data
17:25 sfan5 lol
17:25 sfan5 my rewrite is coming along good
17:27 sapier didn't someone else do a rewrite of sqlite db backend too?
17:28 sfan5 note to self: don't do sqlite3_reset before data is copied somewhere
17:29 sfan5 oh god what
17:29 sfan5 I think it's broken
17:30 sapier fix it :-Ü
17:30 sfan5 I think I should backup my worlds before letting this code touch them...
17:32 sfan5 ok, yes it is broken
17:34 sfan5 19:34:00: INFO[EmergeThread0]: WARNING: Could not bind block position for load: library routine called out of sequence
17:34 sfan5 wat
17:34 sfan5 why does this go to infostream
17:34 sapier I'd guess you didn't reset the prepared statement before
17:34 sfan5 it's a fatal error
17:35 sfan5 maybe I deleted that code
17:35 sapier good morning sfan
17:35 * sfan5 looks at git digg
17:35 sfan5 diff*
17:35 sapier I've been searching for two days because of that sort of error beeing warning only ;-Ü
17:35 sapier -Ü+P
17:37 sfan5 I didn't delete a single sqlite3_ statement
17:38 tomreyn joined #minetest-dev
17:39 sfan5 do we have a warnstream?
17:39 sfan5 I don't think we do..
17:39 sapier no
17:42 sfan5 doing s/infostream/errorstream/ seems to have fixed it
17:42 sfan5 I think database-sqlite3.cpp is cursed..
17:43 sapier then don't touch it
17:44 sfan5 I can't not touch it if I change the Database class methods
17:44 sapier well you just started ... I already spent 2 full days at db code so you've got some time left to finish ;-)
17:45 sfan5 I'm already finished with code
17:45 sfan5 I just need to test it now
17:45 sapier please don't break the android fix or merge it
17:46 sapier hmm of course I meant you should merge it if you can't stop from breaking ;-)
17:48 sapier btw I don't know how you changed but >begin< .... a lot of operations ... >commit< is way more fast then >begin< one operation >commit<  ... or even doing the ops without begin
17:48 sfan5 I changed something?
17:48 sapier I don't know I haven't seen your code by now
17:49 sfan5 I only changed functions args and deduplicated code
17:49 sapier I just mention because shadowninja suggested using begin commit
17:49 sfan5 we use that
17:49 sfan5 beginSave() and endSave() do that
17:50 sapier yes he suggested using it within save block .. which doesn't work because of beeing a nested transaction in this case
17:50 sfan5 why within saveBlock() ?
17:50 sapier https://github.com/minetest/minetest/pull/1457 this fix
17:51 sapier didn't you have a look to pull requests prior changing code beeing worked on by others the last days?
17:51 sfan5 ShadowNinja's suggestion is not about speed
17:51 sapier no but mine was
17:51 sfan5 no
17:51 sapier that's something you should do in cases like that
17:52 BlockMen left #minetest-dev
17:53 sapier as long as noone finds out why REPLACE doesn't work I need that fix for android
18:08 celeron55 joined #minetest-dev
18:11 sfan5 Database_SQLite3::listAllLoadableBlocks is broken
18:11 sfan5 noooo
18:11 sfan5 and I didn't even touch it
18:11 sapier lol did anyone use it before?
18:11 sfan5 yes
18:11 sfan5 it is used for migrating from sqlite3
18:12 sapier then you most likely changed something
18:12 sfan5 I did not
18:13 sfan5 sapier: https://github.com/sfan5/minetest/commit/​537e63005a5f1da1907bcb0d5b1a1b73fa49c5be
18:13 sfan5 I did not even touch that function
18:17 sfan5 WAT
18:17 sfan5 oh
18:17 sfan5 --world did not do what I expected it to do
18:19 sfan5 ok, migrating works
18:23 sfan5 please test: https://github.com/minetest/minetest/pull/1461
19:28 Miner_48er joined #minetest-dev
20:02 LemonLake joined #minetest-dev
20:11 us`0gb joined #minetest-dev
20:22 lemonlake_ joined #minetest-dev
20:23 alexxss joined #minetest-dev
20:24 proller joined #minetest-dev
20:35 alexxs joined #minetest-dev
20:37 Mimilus joined #minetest-dev
20:57 us`0gb joined #minetest-dev
21:03 daswort joined #minetest-dev
21:19 alexxss joined #minetest-dev
21:23 ShadowNinja There are a lot of assert(0)s in Minetest.  These may cause issues in release builds if we remove our custom assert.
21:26 ShadowNinja BlockMen, sfan5, nore: Why only you maintaining _game?  Why isn't it under a seperate _game team or under a seperate organization/user if only you can change it?
21:29 alexxs joined #minetest-dev
21:32 celeron55 ShadowNinja: why are you rambling about that
21:33 celeron55 ShadowNinja: this team proved by _next that they can handle it better than it was handled previously
21:33 celeron55 ShadowNinja: you have no basis for your whining at all
21:35 celeron55 ShadowNinja: show some respect
21:35 ShadowNinja celeron55: As far as I can tell there's no difference other than that it isn't considered frozen and the approval isn't nearly as thorough.
21:35 ShadowNinja celeron55: And I'm not whining.
21:36 celeron55 and why is that automatically considered worse?
21:36 ShadowNinja hmmmm: My server has about 8171422 blocks in it's DB.
21:36 ShadowNinja celeron55: Bad code is let in.
21:36 celeron55 the approval is different but works in practice; i'm going to whine if i start seeing bad results
21:38 alexxss joined #minetest-dev
21:39 ShadowNinja Also, TNT shouldn't be enabled by default.  It's a huge greifing tool.  And BlockMen, sfan5, nore: check this please http://ix.io/dhK/diff It cleans up the TNT mod a lot and disables it by default.
21:39 ShadowNinja My server's running that now.
21:41 ShadowNinja Some of the code is lifted from my unpublished nuke mod.  It can make 256-wide holes in a minute, but this will be much slower because of drops.
21:41 ShadowNinja I might be able to speed it up by using the VM more, but I'd need content IDs for all nodes.
21:43 ShadowNinja I also overwrote the entity physics code without looking at it too closely.  It should be better, but the damage calculation is different, which you may or may not like.
21:46 ShadowNinja Converting from set_node to the VM will break rollback's rollback scope though, like with trees (hence why I removed treegen from my server).
21:48 proller ... and you kill main idea of freeminer's modifications..
21:51 hmmmm celeron55, I asked the other day if there was any rationale behind the design decision to place mapblocks into the map before the map has actually been generated
21:52 ShadowNinja proller: Which was?
21:53 hmmmm which was to break out of calculations every X amount of loops to "improve" performance
21:53 hmmmm :\
21:53 proller ShadowNinja, you like rewriting anything without understanding what it does
22:35 Eater4 joined #minetest-dev
22:45 MichaelRpdx joined #minetest-dev
22:45 sapier left #minetest-dev
23:05 ShadowNinja http://i.imgur.com/UTGmV1y.png This is what the tnt mod's explosions look like now.  It might be a bit big though...
23:06 crazyR not big enough lol
23:06 ShadowNinja Explosion time is roughly 0.03 seconds for a filled area like that.  Airy areas are faster.
23:08 ShadowNinja crazyR: http://shadowninja.minetest.net/mods/nuke.tgz A bit old, but if you modify it you can get explosions with a radius of 128 to work in about a minute.  For comparison the above shot is with a radius of 4.
23:09 crazyR lol, was joking. i kinda had a play with tnt in sp world, they are quite fun, especially when you stack hundreds fo them
23:11 ImQ009 joined #minetest-dev
23:50 werwerwer joined #minetest-dev
23:53 casimir joined #minetest-dev

| Channels | #minetest-dev index | Today | | Google Search | Plaintext