Minetest logo

IRC log for #minetest-dev, 2015-01-06

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

All times shown according to UTC.

Time Nick Message
00:00 celeron55 wait what, i failed to actually remove these mods
00:02 celeron55 what i described was still the combination of homedecor+moreblocks+paragenv7
00:02 celeron55 if that can barely make it, i'll have no issues whatsoever
00:03 celeron55 yes, minetest_game+paragenv7 starts at under 90MB
00:03 paramat lol
00:04 celeron55 and i start on a typical mini island with nowhere to go, thank you mgv5
00:06 celeron55 having travelled some distance, it's 187M
00:07 celeron55 the generation speed is quite slow
00:07 celeron55 how much more speed do you think you can get out of this?
00:09 sapier none map generation is primary bottleneck on android devices
00:09 paramat a little more, not sure how significant
00:09 sapier once map is there it's ok but I've never seen a device beeing capable to generate it as fast as you can walk ... maybe latest quad core phones
00:10 paramat generating the trees is slow but you can control the density of those
00:11 paramat except ... jungles don't look right unless they're thick, it's fun to jump from tree to tree
00:11 celeron55 well, v6 generates quite fast enough
00:12 celeron55 lol good luck jumping from tree to tree on a touchscreen, or getting on a tree in the first place
00:12 paramat otherwise i would suggest ethereal or BFD that use the core biome API
00:14 paramat well, paragenv7 needs updating anyway so i'll do that, give me a day and you could try it again with the slight performance increases
00:16 kilbith ShadowNinja: tabs instead of the checkboxes, and description below the list would be OK for you ?
00:16 kilbith checkbox*
00:18 kilbith lemme know here, will read the logs tomorrow. 'night.
00:20 paramat an alternative would be me squeezing paragenv7 into the biome API and trees into schematics, that would speed things up a lot. need a few days for that
00:20 paramat but not difficult. all trees would be identical though apart from rotation
00:22 sapier left #minetest-dev
00:22 celeron55 i do know the speed of the biome api and schematic trees and it would be very beneficial to this
00:23 celeron55 as i said, i don't care for map generator changes for this so it's fine to use an unstable api
00:24 paramat okay well i'm happy to do that it's not much work
00:24 celeron55 can you recall what parameter of v5 changes the height of the terrain
00:26 celeron55 i mean, like, the steepness of the hills
00:27 celeron55 it's np_factor i think
00:31 paramat yeah factor, but that's range limited, so icrease 'scale' of np_ground instead
00:31 paramat *increase
00:36 celeron55 with np_ground i can also decrease the general amount of water too
00:36 celeron55 let's see
00:37 celeron55 wait, how can these even be set in the first place
00:39 celeron55 if i try to copy mgv5_np_ground from map_meta.txt to the game's minetest.conf, it fails saying "terminate called after throwing an instance of 'SettingNotFoundException'" "Setting [mgv5_np_ground] is a group."
00:40 celeron55 oh well, i'll just modify the engine source
00:40 celeron55 #yolo
00:41 paramat the params need to be in group format because eased noise is specified, see docs
00:42 hmmmmm paramat, you could use my tree schematic, but it doesn't have y slice probabilities
00:42 paramat or rather see .coonf.example
00:42 celeron55 this page mentions nothing about that http://dev.minetest.net/Mapgen_Parameters
00:43 hmmmmm lol that's outdated
00:43 celeron55 paramat: i wrote them in group format because that's the format they were in in map_meta.txt
00:43 paramat thanks hmmmm but i'll save schematics from paragenv7 since those appletrees are different
00:44 paramat yeah sorry i just realised you probably got it right
00:44 paramat another way to increase performance is to use non-eased 3D noise, but the cave width will have to be retuned
00:45 paramat i understand original v5 was non eased and spikier
00:47 hmmmmm hrmmm so you're running into memory problems
00:47 hmmmmm lots of stuff in minetest sacrifice low memory usage for execution speed
00:48 hmmmmm i know it'd probably make an insiginificant difference, but i bet some memory could be saved by not allocating like 3-4 buffers for each noise object
00:48 celeron55 no i'm not
00:48 paramat to un-ease 3D noise delete the noise 'flags' parameter
00:48 celeron55 it happened only because i attempted to run homedecor on a mobile phone
00:48 hmmmmm oh.. vanessae's mods seem to target power users
00:49 celeron55 which is frankly ludicrous, but i tried it
00:49 hmmmmm [07:39 PM] <celeron55> if i try to copy mgv5_np_ground from map_meta.txt to the game's minetest.conf, it fails saying "terminate called after throwing an instance of 'SettingNotFoundException'" "Setting [mgv5_np_ground] is a group."
00:49 hmmmmm that shouldn't happen...
00:51 hmmmmm somebody tried using get() directly on mapgen params which is bad, Settings::getNoiseParams() should always be used instead which reads it as a table just fine
00:51 hmmmmm if i recall correctly, i fixed all instances of that, so it a mod must be the culprit
00:51 hmmmmm s/it//
00:52 celeron55 why would a mod read mgv5 parameters
00:52 celeron55 that isn't happening
00:56 hmmmmm shrug
00:56 hmmmmm i just tried doing exactly what you said and i don't get the error
00:57 hmmmmm it's either a mod, or it was a bug that got fixed between now and when the binaries you're using were built
00:58 celeron55 that would be between 0.4.11 and master; can you confirm the possibility?
00:58 hmmmmm why not just try disabling all mods
00:59 celeron55 i have rather lot going on right now, this issue is not my priority
00:59 hmmmmm i hadn't made any changes to settings between 0.4.11's release and now
00:59 celeron55 i'm moving on already
01:04 Player-2 joined #minetest-dev
01:05 hmmmmm hrmmm
01:15 paramat left #minetest-dev
01:17 Robby joined #minetest-dev
01:18 shadowzone joined #minetest-dev
01:38 Robby joined #minetest-dev
01:43 Megaf joined #minetest-dev
01:45 ShadowNinja celeron55: I feal like releasing to the play store now would just be a good way to accumulate a lot of unfavorable reviews.
01:46 ShadowNinja kilbith:No, keep the checkbox, and the description is good at the side.
01:46 celeron55 ShadowNinja: i will not give any value to this comment unless you can tell me WHEN to release it instead
01:47 hmmmmm once client-side scripting has been established and most of the lingering issues have been solved
01:47 celeron55 what, that's like 2 years of development
01:47 hmmmmm not so
01:47 ShadowNinja celeron55: When it's faster and less hacky and buggy.  I can't really specify exactly when though.
01:47 celeron55 no that's not how minetest's releases have ever been made
01:47 hmmmmm did you see my client side scripting requirement list
01:48 celeron55 no
01:48 ShadowNinja There are definitely a lot of bugs, I have like two lists, and I've heard of a few more today.
01:48 hmmmmm also, i have greater confidence in minetest stability with the new releasing procedures
01:48 hmmmmm celeron55:  the tl;dr is that I have a solid plan and now all there's left is to execute it
01:48 celeron55 what's the timeline
01:49 hmmmmm that i don't know unless i make this more than a one man thing
01:49 ShadowNinja celeron55: Usually it's just "release when enough stuf's accumulated, usually a few months."
01:49 hmmmmm the only reason i've had this burst of activity is because of the long christmas vacation
01:50 hmmmmm after all this development i'd still have to say things aren't in a 'happy place' yet
01:51 celeron55 waiting for longer for what i'm doing (longer than maybe letting paramat do a quick port of paragenv7 on the old biome api) makes no sense; you aren't focusing on the things that people will not like about this anyway
01:51 celeron55 the largest issue with this is that the UI is shit
01:51 celeron55 but that's not on your list
01:51 hmmmmm well for what it's worth, I am working on an options menu right now
01:52 celeron55 and the main issue with the UI is that it's not scaled to the full screen and the scrollbars don't function like you'd expect and stuff like that
01:52 hmmmmm that part is all sapier
01:52 celeron55 which is quite deeply rooted to the nature of irrlicht and formspecs
01:52 celeron55 yes, this is a reason i don't have to wait for you
01:52 hmmmmm kahrl and sapier do formspec, i don't
01:53 hmmmmm formspec and irrlicht gui related things are a totally different brand of nasty that i'm not willing to step foot into
01:53 hmmmmm so i don't get it
01:53 celeron55 the other issue is performance, but you aren't doing anything to that either; you're just making things that are now not even considerable to run on this thing a bit more considerable; it won't affect the speed of this set-up that i'm going to ship on the play store
01:54 hmmmmm in order to make the GUI good you basically need to create your own widget toolkit
01:54 celeron55 i'm mostly replying to shadowninja
01:54 hmmmmm i don't think we'll be successful hacking on more irrlicht stuff
01:55 hmmmmm mesh making I could work on a bit, but i'm not overflowing with graphics knowledge
01:58 ShadowNinja celeron55: There were a number of significant performance improvements recently.  The UI code is big and sapier's the only one that really knows it.  I don't know much about anything Irrlicht either.
01:58 hmmmmm those were all cpu bound though
01:59 hmmmmm and consisted of removing C++-isms that add slowness
01:59 ShadowNinja RBA, kahrl, and you probably know the audiovisual stuff best.
01:59 celeron55 yes, the fact that there were improvements means further improvements are likely to be much harder and now it's a good time to publish it
01:59 hmmmmm I'd rather not publish it until the outstanding mobile formspec issues get fixed
01:59 * ShadowNinja pulls up one of his bug lists.
02:01 celeron55 i've solely worked on making the current android port work in a reasonable manner focusing on the general user experience for 12 hours now
02:01 celeron55 i can tell you that this was totally unusable when i started and now it's much less so
02:01 ShadowNinja celeron55: http://pastebin.ubuntu.com/9679921/
02:01 celeron55 and i can tell you that i'm not going to spend much more than this on it for now
02:02 ShadowNinja celeron55: Other users are probably going to have scaling issues like you, and they aren't going to accept "edit this text file" as a solution.
02:03 celeron55 i didn't edit any text file
02:04 celeron55 you don't have to jump now that the player climbs 1-node steps by just walking
02:04 celeron55 users don't care about logcat or logs whatsoever
02:04 celeron55 crouch is... questionably useful in any case
02:04 ShadowNinja celeron55: Dodn't you have to edit minetest.conf to set gui_scaling?
02:05 celeron55 i don't have multi-second lag in singleplayer once the map has generated, and you generally end up in a situation like that after a while
02:05 celeron55 ShadowNinja: no, i just restarted the game; it fails to save the settings which is handy
02:05 twoelk|2 joined #minetest-dev
02:06 celeron55 and the default scale works kind of okay for tablets; it's not optimal but usable
02:06 ShadowNinja celeron55: Um, well that's a significant bug.
02:07 celeron55 yeah but doesn't matter if you just hit "singleplayer" and build a house to kill some time
02:08 ShadowNinja celeron55: You should remove the settings menu if it doesn't actually work persistently.  It worked before.
02:08 celeron55 lol
02:08 celeron55 you should work on it instead of arguing with me about things i already know better
02:08 kahrl why not add a line like this in GUIFormSpecMenu somewhere: if (formspec size > screen size) { reduce the scaling factor until this is no longer the case }
02:09 kahrl very simplistically speaking of course
02:09 celeron55 it should go the other way too on android so that the screen would always be fully used
02:09 kahrl true
02:10 celeron55 anyway, my point is, we can't get a polished android experience with this technology
02:11 celeron55 we just can't; we can only declare a point where it's sufficiently non-horrible to be given to users without wasting their time
02:11 hmmmmm i agree... wtf is the point of having unused space on the mainmenu on a mobile device
02:11 hmmmmm note that i don't even have a mobile device so i can't go testing or working on that
02:12 celeron55 be glad you don't, you couldn't tolerate the total non-quality MT is on such 8)))
02:12 kahrl about the performance issues, how much could a texture atlas help on mobile?
02:12 hmmmmm from what celeron says, a lot
02:13 celeron55 in this case, dunno
02:13 kahrl we already had some code for doing that so it shouldn't be completely out of reach to add that
02:13 hmmmmm he wants us to readd the texture atlas
02:13 celeron55 my device feels like being limited by vertex count, in which case texture atlases and opengl object counts do not matter
02:14 celeron55 and neither VBOs
02:14 celeron55 but i'm not sure
02:14 hmmmmm do we have an overdraw problem, I wonder?
02:15 celeron55 i mean, this is running at a view range of 20
02:15 celeron55 i can't open the profiler to check what the meshbuffer counts are though
02:15 ShadowNinja hmmmmm: Yes, because the culling is no good.
02:15 celeron55 probably <50
02:16 hmmmmm supposedly the hardware occlusion culling blows horribly
02:16 hmmmmm that's the reason why I put that entire thing on pause
02:17 celeron55 there's no hardware occlusion culling
02:17 hmmmmm occlusion query rather
02:17 celeron55 and that's a heavy operation anyway
02:18 celeron55 in buildat i have the option to use urho3d's generic occlusion culling capability (it's just a flag you set for each mesh); i'm not sure how it works but it kind of varies between improving and degrading performance, depending on what one is looking at
02:19 hmmmmm soo much work needs to be done on the client side :/
02:19 celeron55 i think i ended up just turning it off
02:19 hmmmmm celeron, you should work on minetest with us!
02:19 hmmmmm it'll be great i promise
02:21 kahrl something I've just noticed while looking at the game in wireframe mode:
02:21 kahrl the maximum number of identical-material faces that are ever combined into one is 2
02:21 hmmmmm caves get drawn, right? :/
02:21 celeron55 i won't spend a lot of my time on a thing that i personally won't use; that's just a matter of fact
02:21 kahrl that might explain the vertex count problem
02:21 hmmmmm ahh
02:22 hmmmmm celeron, why don't you use it?
02:22 celeron55 kahrl: even with allowing chains of 16, with smooth lighting the most benefit you get is generally like 10%
02:22 hmmmmm because of the problems or you just don't care for it anymore
02:23 celeron55 kahrl: because you need the vertices for setting the lighting
02:23 celeron55 in caves you don't get even that
02:23 celeron55 not even with non-smooth lighting
02:23 kahrl is there any downside to allowing chains of 16?
02:23 celeron55 it just is an almost useless optimization, buildat doesn't do it at all for this reason
02:24 kahrl I mean if you make the terrain less hilly on android, the benefit will be greater
02:24 kahrl and in a cave there is not that much to render anyway (assuming the occlusion culling doesn't completely fail)
02:24 celeron55 texture atlases become a problem when you need to connect vertices of rotated nodes and have a repeated texture on them 8)
02:25 celeron55 minetest originally had stripes of 16 textures in the atlas for this purpose
02:25 kahrl ah I see
02:25 celeron55 well, it still would work as long as grass nodes don't get facedir set on them
02:26 celeron55 and if they would, it wouldn't be tileable anyway
02:26 celeron55 trying to maintain that 10% optimization with this texture-based style is a losing game 8)
02:26 celeron55 unless you actually generate combined textures for mapblocks on the fly
02:27 celeron55 that's what the one guy did that has the blog with 100 posts with the voxel engine that isn't in actual use
02:27 kahrl heh
02:28 kahrl that sounds incredibly expensive if you use HD textures though
02:28 celeron55 if you generated them like that, you'd probably get a 2x or more optimization on regular generated grass
02:28 celeron55 it probably is
02:29 celeron55 the modern solution to all of this would probably be some shader code which can rotate textures or something
02:30 celeron55 this all is weird because other games don't deal with these kinds of issues
02:35 celeron55 this will finish uploading in 15 minutes: http://packages.8dromeda.net/minetest/M​inetest-0.4.11-8dromeda-1-rc1.apk.part
02:35 celeron55 i'll make a poll on the forum tomorrow asking whether people think it's a fine enough initial official google play release
02:36 celeron55 (there's 8dromeda all over that url because my google play account is with that name too; it's under what i develop my solo projects)
02:37 celeron55 oops, i mean, http://packages.8dromeda.net/minetes​t/Minetest-0.4.11-8dromeda-1-rc1.apk
02:37 kahrl ok, judging by http://tomcc.github.io/images/cull_after.jpg, MCPE doesn't need face combining
02:38 kahrl afaik they also have multiple textures for some common nodes that are chosen pseudo-randomly so it wouldn't work anyway
02:40 celeron55 hmmmmm: after 1 year of development, i was interested in a different game than the community and it's gone to that direction for all the time
02:40 celeron55 since then, i mean
02:41 hmmmmm you were looking to make an actual game kind of game
02:41 hmmmmm what do you think about voxellands/minetest classic
02:41 celeron55 yes; not a house builder
02:41 hmmmmm but programming is fun
02:41 hmmmmm the game part of minetest is writing mods 8)
02:41 celeron55 i have better projects for programming
02:41 VanessaE +
02:41 hmmmmm this is your fault
02:41 hmmmmm you did it with the addition of lua api
02:42 celeron55 well frankly it was the best the project could have had; you all are here because of it
02:42 hmmmmm i have a lot more work to do on it thanks to modding
02:43 hmmmmm it's apparently not enough to do task X
02:43 hmmmmm it needs to be completely configurable, have an excellent interface, have every single edge case working 100%, documented, etc.
02:43 hmmmmm honestly i came here because i wanted to figure out how to make the terrain flatter so i could actually build buildings
02:44 celeron55 you don't have to make the content thanks to modding
02:44 hmmmmm hahaha yes you do
02:44 hmmmmm the assets suck ass
02:44 celeron55 if there weren't modding you'd have to yourself imagine the biomes that the users want, and the mobs and everything
02:44 hmmmmm nobody's made half decent sounds
02:44 celeron55 and chances are because they would be yours, nobody would help you
02:44 celeron55 and everyone would complain
02:45 hmmmmm they seem to lap up minecraft though
02:45 hmmmmm they love that shit
02:45 hmmmmm they EAT the slop!
02:45 celeron55 frankly, i might work on minetest if i could make money of it
02:46 hmmmmm haha
02:46 celeron55 but that's not going to happen
02:46 hmmmmm hey have you updated noise.cpp in buildat lately?
02:46 twoelk|2 from the mc players I know it is more a combination of love and hate
02:46 celeron55 i'm not working on buildat, so no
02:46 hmmmmm oh, you should, it's gotten better
02:46 hmmmmm more organized
02:49 celeron55 anyway, the charm of minetest originally was that i could slap together any kind of crap i wanted with the amount effort i felt like and publish it, and people had fun with it
02:49 celeron55 it's pretty difficult to do that now
02:50 hmmmmm why do you suppose that is?
02:50 celeron55 there are these pesky other people around that think they're developers or something
02:51 hmmmmm haha
02:53 celeron55 i mean
02:53 celeron55 after all these concerns that were now said, i don't feel like putting the apk up on google play at all anymore
02:53 celeron55 i feel like moving to work on my solo projects where i don't need to listen to naysayers
02:54 celeron55 and put anything i want on google play
02:54 celeron55 and elsewhere
02:55 ShadowNinja Forum poll sounds good.  I'd like to see your fixes merged though.
02:55 celeron55 most of them are hacks, not fixes
02:56 ShadowNinja Well, made into fixes and merged them.  :-)
03:19 hmmmmm i am getting so upset with irrlicht
03:19 kaeza joined #minetest-dev
03:19 hmmmmm it's not calling the correct (useful) constructor
03:55 hmmmmm we desperately need to fix config files
03:56 hmmmmm prefix options that affect the client with cl_, server options with sv_
03:56 hmmmmm or perhaps wrap them in a settings group
04:03 * kaeza mumbles something about [sections]
04:04 VanessaE or just split them into separate files.
04:04 VanessaE we can already do that now
04:05 VanessaE (after a fashion)
04:10 hmmmmm true
04:11 hmmmmm i'm just thinking how easily it can get jumbled up if you don't use minetest.conf.example
04:11 VanessaE indeed.
04:11 hmmmmm anyway
04:11 hmmmmm i'm not so sure how i'm going to do this
04:12 hmmmmm right now i have the options menu taking up the entire screen
04:12 hmmmmm sort of ugly with the text in the background though and the transparency
04:12 hmmmmm what do you guys think?
04:12 hmmmmm really?  big menus like that deserve the entire screen
04:26 selat joined #minetest-dev
04:47 ShadowNinja hmmmmm: There should be several pages of settings.  Eg, xonotic has audio, video, effects, user, misc, player...
04:48 hmmmmm i have it:
04:48 hmmmmm Game Options, Video Options, Sound Options, Control Options, Multiplayer Options
04:49 hmmmmm things like view bobbing amount and fall bobbing amount should probably go into a Player Options
04:49 hmmmmm Multiplayer Options should be changed to Server Options or something
04:57 Miner_48er joined #minetest-dev
04:58 chchjesus__ joined #minetest-dev
05:04 hmmmmm hmm
05:04 hmmmmm android doesn't have tooltips so i need another way to explain the options for mobile
05:06 VanessaE press-and-hold for a popup?
05:06 hmmmmm no idea how to do this or even attempt to
05:06 VanessaE or maybe a widget that you can drag around that gives help as it crosses over an item?
05:07 hmmmmm i like the second idea
05:07 VanessaE if screens that could detect hovering a finger (like on Galaxy S4) were more common, I'd just suggest that
05:07 ShadowNinja hmmmmm: There are a lot of visual options, said game splits it into video and effects, you hay have to do something similat.
05:08 hmmmmm yeah
05:08 hmmmmm shaders and related things should count as being an effect
05:09 ShadowNinja hmmmmm: Also, the current setup if pretty horrible, with 5+ lines to handle just setting every indifidual option.  That should be cleaned.
05:09 hmmmmm then there's shit like screen dpi that i have no idea aboy
05:09 hmmmmm about*
05:09 ShadowNinja That would be video.
05:09 hmmmmm my way takes like 5 lines total
05:10 hmmmmm whoever coded most of the current shit is dum
05:10 ShadowNinja aniso, mip-map, etc are also effects.
05:10 hmmmmm nope
05:10 hmmmmm that's definitely video options tab material
05:10 ShadowNinja video is just basic stuff like resolution.
05:11 ShadowNinja At least that's how xonotic does it.
05:11 hmmmmm they're dum
05:11 ShadowNinja It works well.
05:12 hmmmmm i could poop on a piece of paper that has 'options' written on it and that'd work well
05:13 hmmmmm we just need to have *an* options page
05:13 hmmmmm as it exists it's quite pathetic
05:14 hmmmmm so I'm thinking things such as dpi, hud options, tooltip delay amount, etc... they need to go into some sort of "2d gui" settings section
05:14 hmmmmm tooltip show delay shouldn't even be user-configurable.  that's just configuration insanity
05:16 Wayward_One joined #minetest-dev
05:16 hmmmmm so fahnny quit message
05:16 Wayward_One i couldn't resist xD
05:45 chchjesus joined #minetest-dev
06:14 prozacgod joined #minetest-dev
06:22 paramat joined #minetest-dev
06:23 paramat hmmmmm in 0.4.11 stable the z component of 2D perlinmap size is actually necessary to avoid crashes, z=1 is required
06:24 paramat otherwise 'Error in `./minetest': corrupted double-linked list: 0x00007f89b82170d0'
06:24 paramat either that or segfault
06:29 OldCoder joined #minetest-dev
06:37 Hunterz joined #minetest-dev
06:55 paramat left #minetest-dev
06:56 hmmmmm hrmmm
06:56 hmmmmm well that's interesting
06:56 hmmmmm i checked it out, and it shouldn't be necessary
06:57 hmmmmm reason being that when you pass the Z component into the ctor it compares <= 1 for deciding whether or not it's 3d
06:57 hmmmmm rather > 1
06:58 hmmmmm so if z is 0 (which it is as the default value) or a negative number, then is3d == false
06:58 hmmmmm ahhhhh
06:58 hmmmmm that only works for the noise buffer, not the interpolation buffer
06:58 hmmmmm i need to fix this
07:00 hmmmmm when i initially looked at it, i was pleasantly surprised at how the problem of not having to specify a z param works itself out but i guess i was wrong
07:24 paramat joined #minetest-dev
07:27 paramat hmmmmm, i was just looking at your lvm mapgen example, should update_liquids be before write_to_map or after? i have always placed it after, in your example it is before
07:27 hmmmmm actually it doesn't matter
07:27 paramat okay good
07:28 hmmmmm hrmm
07:28 hmmmmm wondering if I should add a "Show Advanced Options" button for the Video tab
07:40 paramat left #minetest-dev
07:46 VanessaE hmmmmm: +1
07:47 hmmmmm what's an advanced option is up for debate however
07:47 hmmmmm i'm going to leave that up to you people to decide
07:47 VanessaE hmmmmm: anything your average dumb user won't understand :PO
07:47 VanessaE :P
07:47 VanessaE really though, all the shader stuff belongs there probably
07:47 hmmmmm well I have the basic options as:
07:48 hmmmmm video driver, resolution, fullscreen, vsync, mipmaps, anisotropic filtering, bilinear/trilinear filtering, fsaa, smooth lighting
07:48 VanessaE hrm
07:49 hmmmmm advanced.. oh man, that's a clusterfuck
07:49 VanessaE smooth lighting is a basic thing, video driver and fullscreen too, resolution maybe also.  the rest?  advanced.
07:49 hmmmmm enable_particles, enable_waving_*, enable_fog, directional_colored_fog, etc.
07:49 hmmmmm well
07:49 hmmmmm on the scale from basic to really freaking advanced
07:49 hmmmmm those are fairly basic compared to the other things
07:49 VanessaE true
07:50 hmmmmm like... desynchronized mapblock texture animation?
07:50 hmmmmm parallax occlusion bias?
07:50 VanessaE but the average user won't know what FSAA even *does*  or that ^^ one either
07:50 hmmmmm yes they do
07:50 hmmmmm they see it in like every video game ever
07:50 VanessaE (and desync, you almost never want to expose, btw)
07:51 VanessaE ok I'll agree with you on FSAA I guess
07:51 VanessaE I'm too used to dealing with kids that don't even know how to google something :-/
07:51 hmmmmm i'm sure the average PC gamer will notice their game looks all pixelly and say to themselves, "i'm going to look in the graphics options menu and see if i can find "antialiasing" or "aa" or "fsaa" "
07:55 jluc joined #minetest-dev
07:56 hmmmmm what do you guys think about biomes overriding settings such as water_wave_height/water_wav​e_length/water_wave_speed?
07:56 hmmmmm or enable_clouds
07:56 hmmmmm if you're in hell you probably don't... want clouds
07:56 hmmmmm or an alien planet
07:57 VanessaE actually that's not a bad idea
07:57 VanessaE I've wanted, for a long while, to be able to specify the cloud height from the server at least.
07:58 VanessaE and disabling clouds would be surely wanted on e.g. moontest
07:58 hmmmmm on minecraft i think clouds are decided server-side
07:59 VanessaE (I forgot/wasn't ware that the wave height/speed/length was configurable)
07:59 VanessaE aware*
08:00 leat /bye
08:00 leat whoops
08:27 leat joined #minetest-dev
08:45 jin_xi joined #minetest-dev
09:01 kilbith joined #minetest-dev
09:15 leat joined #minetest-dev
09:27 kilbith sfan5, the local map saving ought 1) save in its own dir 2) save the worldmods
09:27 sfan5 worldmods?
09:28 sfan5 how exactly is it supposed to do that
09:28 sfan5 it could theortically generate lua code with the node/item definition received from the server, but I'm not going to implement that
09:28 kilbith i mean download the respective mods for each world and put them in /worldmods
09:28 sfan5 download from where?
09:30 kilbith from the server ofc
09:30 sfan5 the server only sends itemdefs, nodedefs and textures to the cleint
09:30 sfan5 client*
09:30 sfan5 not the full mods
09:31 sfan5 there is zero lua code being transferred between client and server
09:32 kilbith ok and it's not technically possible to fetch the mods dir to local ?
09:32 sfan5 no
09:32 sfan5 not technically possible
09:33 kilbith hmm, ok
09:33 kilbith that's a bit embarrassing
09:33 sfan5 <sfan5> it could theortically generate lua code with the node/item definition received from the server, but I'm not going to implement that
09:43 kilbith well, i added your feature in the Client tab: https://cdn.mediacru.sh/z/zuEcbEJK9prO.jpg
10:01 ImQ009 joined #minetest-dev
10:08 celeron55 hmmmmm: it's already possible to disable clouds independently for each player by setting a non-default sky with the set_sky function; but i guess biomes could control how the default sky looks
10:08 celeron55 so then you'd end up using that functionality only if you want to do something really special or don't use native biomes
10:08 celeron55 it's a parallel implementation though, which will cause problems when people want something in between but can't have it
10:10 celeron55 it should probably be something like override_sky() which could affect any individual part of the sky and not just override everything
10:22 selat joined #minetest-dev
10:24 leat joined #minetest-dev
10:47 ImQ009 joined #minetest-dev
11:02 kaeza joined #minetest-dev
11:08 celeron55 has somebody broken profiler_print_interval?
11:08 celeron55 i'm setting it to 10 and enabling info-level logging, but nothing comes up in logs
11:17 celeron55 no, i'm giving up
11:17 celeron55 what the fuck
11:17 celeron55 this "improved" configuration phase at start-up is ridiculously convoluted
11:18 celeron55 fix this shit, i'm out
11:21 celeron55 all i wanted to do is find out what is taking so much processing time per frame on my phone, but apparently i can't do it
11:27 proller joined #minetest-dev
11:29 leat joined #minetest-dev
12:03 leat joined #minetest-dev
12:14 cib0 joined #minetest-dev
12:20 sapier joined #minetest-dev
12:24 sapier ShadowNinja: why don't you just go an help fixing the android bugs on your list instead of insisting on "I told there's a bug" ... there are some bugs on it which don't even require major knowledge about android to be fixed
12:25 sapier e.g. your shader config "BUG"
12:25 sapier or the log "bug"
12:25 sapier or "inventory fields debug"
12:27 sapier hmmmm celeron55 android ain't only phones but tablets too and tablets may have screensizes up to small pc's so there unused space in menu might be better than having large distances between controls
12:27 sapier just to be back at main android issue ... the big range of device types
12:29 celeron55 well, true
12:30 sapier "ShadowNinja  hmmmmm: There should be several pages of settings.  Eg, xonotic has audio, video, effects, user, misc, player..." well someone promised to do this change about 6 months ago
12:30 celeron55 the size should be set in such a way that it tries to be physically a certain size
12:30 celeron55 and if that doesn't fit on the screen, then whatever fits
12:31 sapier actually I try to do this, yet it's very simplistic and my have bugs in combinationwith new font scaling code
12:31 celeron55 on my phone's screen the menus are way smaller than what fits on the screen, and physically they are sized for insects
12:31 sapier android evaluates dpi as well as screen size and trys to set a sane gui_scaling default ... well what I found on my devices to be sane .. I don't put my hand into fire this is true for other devices too
12:32 sapier yes that's been same for me after gui scaling too
12:32 sapier wait
12:32 sapier ShadowNinja: did merge the two menus
12:32 sapier and simple menu didn't have fixed menu size as complex has
12:32 sapier maybe that difference was lost
12:33 sapier let me check
12:33 sapier no he did it correct
12:37 sapier I guess I found the issue with our font's they seem to handle dpi different then formspec
12:38 sapier I gues they're adjusted to dpi twice
12:40 sapier grr they're not adjusted at all because of that silly half disabling font size hack ... ok back to beginning to find the real issue
12:43 chchjesus joined #minetest-dev
12:44 sapier reenabling the font size scaling as it's been meant to behave fixes the issue ... yet there have been complaints about the font size scaling
12:44 sapier let's see if now menu size is correct on my phone too
12:46 sapier Zeno for what I remember you did the workaround can you tell me what exactly you tried to fix?
13:41 twoelk joined #minetest-dev
13:55 crazyR joined #minetest-dev
13:56 shadowzone joined #minetest-dev
14:08 sapier celeron55 I found out why menu is so small ... it's related to in game controls
14:08 sapier the scaling factor is same and it's adjusted to make controls in game work
14:08 sapier increasing the scaling makes controls overlap
14:16 celeron55 yes, they will overlap a bit when the menu is at a reasonable size on the small screen
14:16 celeron55 also, some menus are much larger than others and they should be automatically limited to fit the screen
14:16 celeron55 so i guess the in-game UI should be limited similarly to them
14:17 sapier I agree that they should but I fear we can't do this without causing sideeffects like font scaling did
14:17 sapier and there someone just disabled the font scaling to get rid of them ;-)
14:18 sapier btw if you reenable it android menu fonts match the formspec size and look fine
14:19 celeron55 what do you mean someone disabled font scaling
14:19 sapier guiFormSpecMenu.cpp about line 78
14:19 sapier select_font_by_line_height selects correct font size for formspec size
14:19 sapier but it's just been disabled and provides default font size
14:20 celeron55 why
14:20 leat joined #minetest-dev
14:20 sapier because some ppl haven't been happy with fonts doing correct scaling. I haven't found out till now WHAT exactly they didin't like
14:20 celeron55 just change it back, it's obviously the right thing to do
14:21 celeron55 scaling a pixel-positioned UI like this doesn't make sense if fonts don't scale along with it
14:21 celeron55 or, makes very little sense
14:21 sapier I guess some ppl did design their oversized formspecs with small fontsize in mind so those formspecs might now have fonts overlapping gui elements
14:22 celeron55 what do they expect? development to completely stop?
14:22 sapier but I don't really know everytime I did ask what exactly is wrong I didn't get a answer
14:22 celeron55 they can stay at 0.4.11 if they don't want development
14:23 sapier I guess we could add automatic scale down if a formspec exceeds display size
14:24 sapier not sure how much has to be changed but I guess it's limited
14:26 sapier ok fixed the progress bar jumping on asset copy
14:27 sapier guess I'm gonna push there changes in a few minutest just testing them on different devices
14:27 kilbith joined #minetest-dev
14:29 cib0 joined #minetest-dev
14:52 ImQ009_ joined #minetest-dev
14:52 leat joined #minetest-dev
14:57 leat joined #minetest-dev
15:02 leat joined #minetest-dev
15:04 sapier https://github.com/minetest/minetest/pull/2057 merging in a few minutes, only non android change is re-enabling of font scaling
15:08 leat1 joined #minetest-dev
15:12 sapier I made ShadowNinjas buglist to issues ... well I don't understand why he doesn't do this himself, bugs can't be fixed if they're only known to him as he refuses to fix them too
15:16 sapier celeron55 I just pushed a bunch of fixes for android, not all we did talk about yesterday
15:20 paramat joined #minetest-dev
15:21 shadowzone joined #minetest-dev
15:26 sapier Ok after checking all of shadows list I could only fix a single of his complaints, everything else either "works for me" or is a matter of taste/design. Performance issues depend on particular devices, I don't know what device he's using.
15:27 hmmmm joined #minetest-dev
15:28 sapier e.g. the chat window .. it's a formspec so it behaves like one. And in formspecs there could be more then one text field so closing it immediatly ain't a good idea in general
15:28 sapier We could replace the formspec chat window by a native android one, yet if there are plans to remove it in favour of console why do this?
15:32 paramat hmmmm i need a more general way to switch off clouds (while keeping the day-night cycle/blue sky/sun/moon/stars), not a way only usable within biomes. something better than this https://github.com/paramat/flex​realm/blob/master/init.lua#L187 which fails to work properly when playing singleplayer, it leaves my .conf modified so i always have to edit it to re-enable clouds
15:33 paramat must go, exhausted
15:33 paramat left #minetest-dev
15:35 hmmmm I agree, a lot of audio-visual effects are left as config options that have an impact on gameplay
15:35 hmmmm what I think would be smarter is to leave clouds as configurable with a client-side
15:35 hmmmm client-side mod
15:36 hmmmm you know I really don't like adding more server/client packets for asstarded things like enabling and disabling clouds
15:38 hmmmm leaving all of these things for client-side modding is good because it builds up more pressure for getting the client side modding done
15:38 sapier why? ;-)
15:38 sapier there's no developer who really wants it by now
15:39 sapier I never heared anyone (except you) demand it who actually does work
15:40 hmmmm because the people who really want it are modders, not developers
15:42 sapier do those modders know that these mods would be 100% independent of the mods they do now?
15:42 hmmmm they wouldn't
15:42 sapier of course they would
15:42 hmmmm a server-side mod is able to create client-side mod dependencies
15:42 Wayward_One joined #minetest-dev
15:42 hmmmm the user is directed to download the client-side mods from a trusted repository like mmdb
15:43 sapier no because if there was a ways for them to communicate you'd immediatly get huge security implications
15:43 hmmmm I realize that
15:43 hmmmm as long as we strap down the GPU access as well we should be fine
15:43 sapier btw there's two different things
15:43 hmmmm having them communicate is pretty much necessary
15:43 sapier client side lua AND client side mods
15:43 sapier both are quite different
15:44 kahrl I thought they were called OTTO and ZEUS
15:44 sapier yes kahrl they are but I don't know if hmmmm alread knows the names
15:45 PilzAdam joined #minetest-dev
15:45 sapier ;-)
15:45 sapier seems the names are silly enough to get remembered
15:45 sapier so OTTO is security critical ZEUS ain't ;-)
15:45 sapier as long as ZEUS doesn't have communication capability to server
15:46 hmmmm how is this any different from running a website's javascript
15:46 leat1 joined #minetest-dev
15:47 kahrl I could only remember them because there was a play with Zeus at the Hans-Otto in Potsdam ;)
15:47 sapier OTTO ain't different (except of testing effort)
15:47 sapier ZEUS is more like a browser style theme
15:48 sapier kahrl captial letters are required ;-)
15:48 sapier hmmmm I assume you actually want OTTO and not ZEUS
15:49 ImQ009 joined #minetest-dev
15:50 hmmmm i have no idea what otto and zeus are
15:50 sapier OTTO is lua code sent from server mod to client
15:50 sapier automatically no manual install by client
15:50 hmmmm i thought we decided against OTTO
15:51 sapier I don't know about a decision like that
15:51 sapier OTTO is basically identical to javascript
15:51 sapier except of the testing of course
15:51 hmmmm right
15:52 hmmmm what i meant to say is that this is no more dangerous than javascript
15:52 sapier it is ... not by design but by confidence in code correctness
15:52 sapier even javascript does have issues every now and then it's quite likely we'll have more
15:53 sapier and ZEUS with client server communication added is no less dangerous then OTTO but way more inconvenient
15:55 hmmmm ZEUS is less likely to be malicious though
15:55 sapier no it ain't if there's client server communication ... at least if zeus aint a limited home brew language
15:56 hmmmm oh, because they can execute instructions from the server effectively making it OTTO?
15:56 hmmmm yeah, that's essential for making things work
15:56 sapier exactly
15:57 sapier so only real usecase for ZEUS is modifying the look and feel of client ... well nice but nothing I'd spend time on ;-)
15:58 sapier while OTTO is valuable and I may even implement it soon
15:58 sapier especially as most parts are already done
15:59 hmmmm woah, one thing at a time here :/
15:59 hmmmm we need to get the gui working on android well
15:59 hmmmm from what i understand, the menus don't take up all the available screen space on android... what's up with that?  it's a mobile device
15:59 sapier hmmmm write issues for what you feel not to work! don't do it like shadow
16:00 sapier menu taking all available space on my tablet would look crazy
16:00 sapier there's no one solution for all ;-)
16:01 sapier we're gonna have to do some tuning and find heuristics to decide best
16:02 PilzAdam sapier, https://github.com/minetest/minetest/commi​t/083d19b3fc8f60468e124c801296c13b66c41abc​#diff-65f34680878a6bd86f3a59ebc0c06c6dL93 wtf?
16:02 sapier but that's not gonna work the "i change it this way because it looks best on my device"-way ;-)
16:02 PilzAdam have you fixed the issues with the font scaling?
16:02 sapier what exactly?
16:02 sapier of course
16:02 sapier works for me
16:02 sapier at least all "issues" anyone told me
16:03 PilzAdam nothing changed for me
16:03 PilzAdam the text is like 200% bigger now
16:03 cib0 joined #minetest-dev
16:03 sapier did you fix your gui scaling?
16:04 PilzAdam I don't want to change the size of inventory images
16:04 PilzAdam I want to change the size of the font
16:04 sapier you can't
16:04 PilzAdam yes, thats the issue
16:04 sapier you couldn't really do this before either
16:04 sapier it's not an issue
16:04 PilzAdam it worked perfectly fine
16:05 celeron55 can't there be a setting that by default is disabled but if it is set to a value, the value is directly used as the font size with no scaling whatsoever?
16:05 sapier vanessaE has a nice cartoon there ... about some spacebar heating processor
16:05 celeron55 then pilzadam and other old-timey desktop users can use that setting
16:05 VanessaE http://xkcd.com/1172/
16:06 PilzAdam celeron55, that only fixes a part of the problem
16:06 sapier well he can set the font size smaller so it's gonna behave same but I won't add a if then else else else cascade everywhere wher dpi screen size and font size is used
16:06 PilzAdam another thing is that I want a bigger hotbar, but not a bigger inventory
16:06 sapier noone ever will be able to debug this
16:06 VanessaE sapier: xkcd always has a relevant cartoon :)
16:06 PilzAdam one setting for everything GUI related is stupid
16:06 sapier why?
16:07 sapier do you wanna have a different setting per formspec element name?
16:07 PilzAdam the inventory scales with my window size while the hotbar doesn't
16:07 sapier or maybe even per formspec id
16:07 PilzAdam I don't want to change the GUI scale factor when resizing my window
16:07 sapier of course that could be added you'd just have to write your own client config per server/mod
16:07 celeron55 we should just rip out all UI and implement it properly from scratch
16:07 celeron55 and obsolete every mod that has any UI
16:08 VanessaE celeron55: define "properly"
16:08 celeron55 it's the only way to win
16:08 celeron55 taking all learned concerns into account
16:08 sapier celeron55 we just mad it consistent causing all those complains ... do you expect a new one beeing consistend wouldn't cause same complains?
16:09 PilzAdam also why is the status text (e.g. when chaning noclip) moved to the center of the screen?
16:09 PilzAdam it used to be in the bottom left corner
16:09 sapier the only way to consider all those complains would be adding some sort of html + css mechanism where each user could specify his own css ;-)
16:09 celeron55 i was wondering that too, it was changed by someone recently (i saw the commit)
16:09 celeron55 i think it's a stupid change
16:10 sapier I haven't changed that PilzAdam
16:10 sapier at least not intentionally
16:10 celeron55 PilzAdam: can you make a list of what UI things you want to configure separately, given that everything must fall under some setting
16:11 celeron55 i think it is very unclear until you do that
16:11 celeron55 the scaling of them, i mean
16:11 sapier well I guess we need two different scale factors for hud and formspec anyway so maybe PilzAdams issues would be solved this way
16:11 PilzAdam celeron55, what I personally want now or what would be good to have generally?
16:12 celeron55 PilzAdam: something that can be implemeted and that also is usable on everywhere where minetest is used
16:12 VanessaE sapier: font size in a formspec, font size in the main menu (and related), HUD scaling ... those are the three I can see an immediate need for, if you were to fully re-enable your font patches
16:12 VanessaE (c55 ^^)
16:13 celeron55 when every person decides to require one part of the UI to be exactly the same as before, with enough people no parts of the UI can change at all and we will be left with no way whatsoever to make this work on different screen sizes
16:13 sapier formspec is formspec I'll not implement different scale factors for them ... best I'd provide is implement per formspec configurable form size
16:13 celeron55 VanessaE: relative to what? pixels? window size? screen size? dpi?
16:13 sapier font size
16:13 celeron55 VanessaE: each otherr?
16:13 celeron55 -r
16:13 sapier meaning you can specify a different then default font size for each text in a formspec
16:14 VanessaE celeron55: relative to physical screen size, I suppose.
16:14 VanessaE I dunno, I'm just going from my own experience when his font patches were enabled.
16:15 celeron55 well experience is what we need, not guesswork
16:15 VanessaE the HUD would be relative to some sane "standard size" (perhaps the usual height as seen in an 800x600 window from several versions ago)
16:16 sapier right now font size keeps aspect to form size
16:16 sapier if form spec is shown at twice size the fonts are shown twice as high too
16:17 VanessaE sapier: which looks like total ass, frankly.
16:17 celeron55 well, it's exactly what one wants on the 400x300...1920x720 2.5"...15" androids
16:17 VanessaE but that takes a back seat to "gazonga fonts in the inventory, dinky fonts in the main menu"
16:18 celeron55 i mean, 400x300 15" ... 1920x1080 2.5"
16:18 VanessaE celeron55: sure -- on an Android display, but you do NOT want it on a desktop usually
16:18 sapier so how do you think it should behave VanessaE
16:18 VanessaE right, I got that.
16:18 VanessaE sapier: a consistent size on the desktop, regardless of window size.
16:18 celeron55 that's what the problem seems to be
16:19 sapier always keep your 10 pixels font even if screen has 100dpi?
16:19 PilzAdam the problem is (I guess), that formspecs are not good at defining a GUI that would work with different font sizes
16:19 VanessaE like EVERY OTHER DAMNED APPLICATION.
16:19 sfan5 ^
16:19 sapier even my tablet doesn't have enough physical size to show my desktop sized menu!
16:19 VanessaE sapier: no, keep my font at 13 point (note, points -- not pixels) regardless of 96 or 200 DPI screens, or 800x600 window or 1600x1200 window
16:19 celeron55 sapier: what if you just scale the formspecs and UI according to *display* size, not *window* size? does that even make any sense?
16:20 celeron55 it might not make
16:20 VanessaE celeron55: then big formspecs - probably even the standard inventory - will exceed the default window bounds.
16:20 celeron55 VanessaE: wut?
16:20 sapier hmm would be worth a try but would break for almost anyone
16:20 celeron55 that's what already happened previously
16:21 VanessaE celeron55: nevermind - misread.
16:21 celeron55 and what happens if you have a consistent size like you wanted
16:21 VanessaE I don't get what's so fucking hard about reading the screen DPI from the OS's windowing system?
16:22 VanessaE celeron55: what happens is what happens now - fonts remain readable at any window size and there isn't any of this mishegas about fonts suddenly becoming monster-sized just because the window is bigger.
16:22 celeron55 has anyone even tried getting the DPI from the OS
16:22 sapier yes I did
16:22 sapier it's not possible on pc right now
16:22 celeron55 what happened
16:22 VanessaE bullshit
16:23 sapier there are some discussions about implementing it in future irrlicht versions
16:23 VanessaE then don't use irrlicht to query it.
16:23 sapier well it's the only consistent way unless you wanna implement a separate way for each and every os
16:24 sapier but I have to leave no I'm gonna read the logs later
16:24 sapier left #minetest-dev
16:24 VanessaE "each and every OS" meanwhile there are only three windowing systems you need to worry about.
16:24 VanessaE X11, Windows, and OS-X.
16:24 celeron55 true
16:25 VanessaE and if you can't get DPI sanely, assume a standard default that the user can easily change.
16:25 Peon501 joined #minetest-dev
16:25 PilzAdam celeron55, I thought about that list with UI settings, and it turns out that missing settings aren't really the problem
16:25 VanessaE or get it from...*gasp* the user's window manager
16:26 PilzAdam there are 2 problems that currently annoy me: 1) some things scale with the window size (formspecs, except mainmenu) and others don't (version string in the top left; hotbar) 2) when scaling up, everything is scaled up instead of realigned
16:26 PilzAdam 2) causes these gigantic fonts on larger screens
16:27 VanessaE PilzAdam: 2. is precisely the problem I'm arguing against as well
16:27 PilzAdam if we had a proper definition of our UI's, then we could scale things up properly instead of just "zooming in"
16:28 VanessaE a quick google says Windows from XP on up, and X11 all have API calls to query the screen DPI
16:29 PilzAdam we basically just need 2  settings if these problems are fixed: 1) a gui scale factor 2) how windows size scales the GUI
16:29 roniz joined #minetest-dev
16:30 PilzAdam so we need layout managers (like in Qt or swing) to define formspecs
16:37 Calinou joined #minetest-dev
16:38 PilzAdam or we allow Lua to define a function that gets the window size as parameter to layout the formspec
16:38 PilzAdam this would only work with client size Lua, though
16:38 Calinou client can send its window size to server.
16:38 Calinou then server can work with it
16:38 Calinou of course, there should be a reasonable min/max to avoid fakes
16:38 VanessaE Calinou: s/window size/screen size/
16:38 VanessaE or send both
16:39 Calinou window or screen size? the Minetest window size is probably more relevant
16:39 VanessaE what's a reasonable max?
16:39 Calinou eg. don't allow width > 4096
16:39 Calinou but that's not really necessary
16:39 Calinou in case a client would send something like 50000 to try crashing server
16:39 VanessaE there are some screens beyond that rez :)
16:39 VanessaE oh yeah
16:40 Peon501_ joined #minetest-dev
16:54 VanessaE bottom line is, no other standard app does this whole "scale every damn thing just because we can" routine.
16:54 VanessaE not if they're trying to stick to your standard buttons/checkboxes/lists widgets paradigm anyway
16:59 est31 joined #minetest-dev
17:00 est31 hmmmm there is a srp implementation not requiring libgmp
17:00 est31 only openssl
17:22 Hunterz joined #minetest-dev
17:44 rubenwardy joined #minetest-dev
18:14 Krock joined #minetest-dev
18:36 cib0 joined #minetest-dev
18:45 sfan5 pushing in 5 minutes: http://sprunge.us/GbBd?diff
18:46 Krock looks like a hack. isn't it poss to chack if EXISTING_MINETEST_DIR is empty?
18:46 Krock *check
18:46 sfan5 this is what I'm doing
18:46 Krock but it looks like the old way failed somehow
18:47 Krock and now you're just adding dummy text and checking the result of it
18:47 sfan5 *sigh*
18:47 sfan5 -d "" is probably true
18:47 sfan5 and "x$var" = "x" is a perfectly fine way to check whether the var is empty
18:47 Krock mhm okay
18:48 sfan5 Krock: http://stackoverflow.com/a/9097530
19:00 chchjesus joined #minetest-dev
19:21 Miner_48er joined #minetest-dev
19:30 hmmmm est31:  cool, but I'm not the one interested in strengthening the authentication.  i just recommended using SRP as an alternative to challenge response if you were going to strengthen it
19:31 hmmmm like I said before... auth is FUBAR but it's just a crappy game, not important things such as SSH or online banking or whatever
19:31 hmmmm people who use their "real" passwords for minetest are fools
19:31 VanessaE but...but...players might lose tens of thousands of minegeld! :P
19:45 russia_nekto joined #minetest-dev
19:45 russia_nekto hi every1
19:45 russia_nekto want to ask, how minetest rendres settings menu ?
19:46 VanessaE it
19:46 VanessaE it's an Lua formspec
19:46 russia_nekto but where?
19:46 hmmmm it uses IGUIEnvironment
19:46 russia_nekto therey is only  gettest
19:46 leat joined #minetest-dev
19:46 russia_nekto core.show_keys_menu()
19:46 hmmmm oh
19:47 hmmmm that's the script api for showing the key menu
19:47 russia_nekto where is it placed?
19:47 hmmmm the real work is done in src/guiKeyChangeMenu.cpp
19:48 russia_nekto so , render code is there?
19:49 hmmmm it doesn't "render" the GUI elements, but it places them and handles events from them
19:49 hmmmm the rendering is done inside of irrlicht
19:49 russia_nekto ok. i got it.
19:49 russia_nekto wait i take a look in this cpp
19:50 russia_nekto i have one more Q
19:50 crazyR joined #minetest-dev
19:51 hmmmm yes..?
19:54 russia_nekto i have seen this cpp
19:54 russia_nekto is it true , that keys placed in enum?
19:55 russia_nekto another Q: how to unbind key ? or clear the key ?
19:56 est31 hmmmm, I could implement srp, but I need to get insight in how minetest contributing works. any small issue I can fix?
19:58 russia_nekto hmmmm: i want to add invert_mouse options to GUI but it seems to be not so fast to do
19:58 hmmmm hmmm
19:58 Calinou russia_nekto, look at how other booleans are added in the Lua code
19:58 hmmmm russia_nekto:  you accomplish this by writing to the g_settings entry by that name
19:58 Calinou you will have to add some space though
19:59 Calinou else, your checkbox will fly over the clouds
19:59 Calinou you could add mouse_sensitivity slider too…
19:59 hmmmm coincidentally I'm already working on an all-encompassing options menu which includes that
19:59 hmmmm est31, I have no idea :/
19:59 russia_nekto hmmmm: nice to hear
20:00 russia_nekto one more Q: is is possible to render some texture under invetory menu? want to light some beauty to this grey boxes
20:02 russia_nekto and Chest menu too
20:02 est31 http://irc.minetest.ru/min​etest/2014-09-22#i_3940514
20:02 est31 I'll write a PR
20:03 russia_nekto what is PR?
20:03 sfan5 pull request
20:03 est31 pull request
20:04 russia_nekto :)
20:04 sapier joined #minetest-dev
20:04 russia_nekto PR for what u`ll write?
20:05 est31 wget 'url' && mv here there → git clone
20:05 est31 simpler and better
20:05 proller joined #minetest-dev
20:05 sapier VanessaE we've got already 3 windowing systems on linux
20:05 VanessaE sapier: we have one.  X11.
20:05 sapier x11 wayland and the ubuntu stuff
20:06 VanessaE oh you mean those upstarts?
20:06 VanessaE fuck them.
20:06 VanessaE (the ubuntu one is Mir or some such I think)
20:06 sapier doesn't help they're gonna be used in near future
20:06 Player_2 joined #minetest-dev
20:06 russia_nekto one more Q: is is possible to render some texture under invetory menu? want to light some beauty to this grey boxes
20:07 VanessaE russia_nekto: there's a formspec command for that, yes
20:07 VanessaE sapier: my point is, getting the DPI of the user's screen is a basic function.
20:07 russia_nekto can u explain more detail?
20:08 VanessaE russia_nekto: take a look at the lua_api.txt file in doc/
20:08 russia_nekto i need to call formspec with texture arg?
20:08 VanessaE I don't remember the command, but it's just an extra field you add to the default inventory formspec is all
20:08 sapier well it should and I'm not against it but the problems we're argueing about are not caused by dpi
20:09 VanessaE sapier: they're caused *exactly* by not knowing the screen DPI
20:09 Warr1024 joined #minetest-dev
20:09 sapier our problem is based on two historicyl issues
20:09 VanessaE sapier: because then you're measuring everything in pixels instead of points or mm or something else physical
20:09 sapier 1) requirement to make menu exactly same (pixel wise) size as on 800x600
20:09 sapier and 2) requirement do to do some other adjustment
20:09 sapier obviously those requirements don't match
20:10 VanessaE no one said the *menu* has to be exactly the same.  the argument is that the fonts have to be the same - you can't go blowing a font that's sized to 13 points in the main menu, up to 50 point sized in-game just because the window size changed.  no application I know of does that.
20:10 sapier well this HAS ben told when the old one was replaced by formspec
20:10 VanessaE that's because too much changed all at once
20:11 sapier I'm the first one to drop this if there's an agreement to it
20:11 sapier actually on android it IS dropped
20:11 VanessaE besides, our current menu isn't really the same as it used to be.
20:11 sapier it is
20:11 VanessaE it didn't used to have the game selector on the bottom did it?
20:12 VanessaE or the Texture Packs tab
20:12 sapier last version did have
20:12 VanessaE I'm comparing to the C++ menu
20:12 sapier texture packs tab doesn't change the overall size
20:12 VanessaE and we don't have those vertical labels along the left anymore either
20:12 VanessaE so it's not "exactly the same" as before
20:12 sapier we do
20:12 ImQ009 joined #minetest-dev
20:12 VanessaE um, no we don't.  I'm looking at it :)
20:13 sapier at least I have those vertical labels don't they work for you
20:13 VanessaE I don't even see them, I thought kahrl removed them
20:13 VanessaE hell, "Client" tab clearly lacks the room for a vert label
20:14 sapier hmm you're right we don't have them .... who did remove them?
20:14 sapier I didn't do this for sure
20:14 VanessaE kahrl did
20:14 VanessaE 581efea60e8fad18b9a2fc9d544f014e2ac693f8
20:14 russia_nekto VanessaE: background[<X>,<Y>;<W>,<H>;<texture name>] this assigns bg?
20:14 sapier I see
20:14 VanessaE russia_nekto: yep, I think that's the one.
20:15 sapier well as I said I always preferred scaling main menu it's just not been accepted by that time
20:15 VanessaE just add that to the formspec code that sets the inventory
20:15 sapier we didn't have proper formspec scaling by that time maybe it's better now
20:15 russia_nekto VanessaE: can u answer where is a call for formspec for inventory?
20:16 VanessaE russia_nekto: I don't know, try digging around in builtin/
20:16 russia_nekto ok.
20:16 russia_nekto anybody knows how to find formspec call for inventory ?
20:18 sapier well I'm gonna implement the x11 screen dpi code for testing purposes ... but I don't think that's gonna solve everything
20:18 sapier I think we need more pieces to get it into a shape everyone can live with
20:21 VanessaE according to MSDN, sapier: there are also calls for WinXP through 8.1
20:21 VanessaE s/sapier: //
20:21 sapier I still don't think this solves the issues
20:21 VanessaE well, if you know the screen DPI, you can compute font size as a function of *points* rather than pixels
20:21 sapier you can try it set screen_dpi variable manually
20:22 VanessaE then it doesn't matter if you have 800x600 or 16000x12000
20:22 sapier no because font size is not relative to screen dpi but to formspec
20:22 VanessaE I mean absolute size
20:22 sapier later one is relative to screen dpi
20:22 VanessaE it should NEVER be relative to formspec.  ever.
20:22 sapier EVER
20:23 VanessaE EVER. :)
20:23 sapier because that's the only consistent way of scaling the forspec as a whole
20:23 Sokomine i'm testing celeron's new android build. as the topic seems to be the formspecs...have you considered doing one for android where the main actions are more like icons, and lists to select from fullscreen?
20:23 sapier because some of our formspec elements depend on font size
20:23 VanessaE why do they depend on font size at all?
20:23 sapier that's silly I know but that's the way it is
20:24 VanessaE you're talking about the height of a button for example?
20:24 sapier tell this to the guys who initially implemented buttons and text fields
20:24 sapier even labels depend on font size
20:24 Sokomine :-/
20:24 VanessaE then change the code so that the button height depends on the *rendered* font size.
20:24 sapier I can't
20:24 VanessaE yes you can.
20:24 Sokomine guess people don't line up in order to rewrite formspecs...
20:24 sapier Well it's not I can't as I can't write the code it's I can't because YOU don't accept it
20:25 sapier remember our discussion with zefram?
20:25 russia_nekto repeat: anybody knows how to find formspec call for inventory menu in lua scripts ?
20:25 VanessaE sapier: look, no UI in existence on a desktop tolerates the kind of font scaling you're trying to push through
20:25 sapier I already fixed it but YOU (and others) didn't want that fix
20:25 VanessaE it just does. not. work.
20:25 Sokomine isn't that something based on irrlicht?
20:26 sapier It's not MY font scaling this is YOUR style of font scaling vanessaE I just accepted it for peace reason and now I have to defend YOUR requirement AGAINST you? ironic
20:26 * VanessaE sighs
20:27 sapier I tried to fix the formspec alignment and font dependencys but that'd have broken any single existing formspec so you didn't want it
20:27 VanessaE it's not my style of font scaling, sapier.  it's just how desktops work. you don't go fucking around with peoples' font settings (or looking like you do) just because those settings seem "wrong" in certain cases.
20:27 exio4 joined #minetest-dev
20:28 sapier well I can't change the fact that I can't travel back in time and fix the bug in current formspec positioning ... and if I fix it now it's gonna break existing things
20:28 sapier you don't want existing things to be broken so I can't fix the bug ... quite simple
20:28 hmmmm I suggest we break things and move forward
20:28 hmmmm this next release is going to be quite a doozy so it's justified
20:29 VanessaE hmmmm: break things in what way?  so that every last mod that has some kind of formspec has to be rewritten and ends up incompatible with 0.4.11?
20:29 VanessaE (and prior)_
20:29 hmmmm possibly
20:29 hmmmm it depends on how difficult it is to do the rewriting
20:29 hmmmm maybe we could do an automatic conversion or add a compatibility mode
20:29 sapier hmmmmm I'm not gonna do this if there ain't a common agreement on it. I'm not gonna spend another week of develompent for trash
20:29 hmmmm I definitely won't trash it
20:30 n4x joined #minetest-dev
20:30 hmmmm again... spacebar CPU heater
20:30 hmmmm there's always going to be some people upset with changes
20:30 russia_nekto devs, is there a predefined name of file for background texture for inventory menu ?
20:30 hmmmm i don't know
20:30 VanessaE this goes WAY beyond just breaking someone's workflow
20:31 VanessaE russia_nekto: no, there isn't.
20:31 sapier https://github.com/minetest/minetest/pull/1561 that's the prototype
20:31 sapier most is already done yet ppl didn't want it
20:31 VanessaE http://digitalaudioconcepts.com/vanessa/h​obbies/minetest/screenshots/Screenshot%20​-%2001062015%20-%2003%3a32%3a24%20PM.png
20:31 VanessaE explain this, hmmmm
20:31 VanessaE this is HEAD, right now.
20:32 VanessaE before the font changes, that chat font you see behind the formspec just happens to be about the size that everything inside the formspec also was.
20:33 sapier looks like an exact scale up of 800x600 formspec
20:33 sapier no disorted font locations, nothing hidden
20:33 VanessaE http://digitalaudioconcepts.com/vanessa/h​obbies/minetest/screenshots/Screenshot%20​-%2001062015%20-%2003%3a34%3a38%20PM.png
20:33 sapier if modders don't want this they can tell a formspec to not to be scaled
20:33 VanessaE how about this one then?
20:34 sapier how is it gonna look like on 800x600?
20:35 hmmmm christ
20:35 hmmmm this is so FUBAR
20:35 VanessaE before, the fonts would be correct at any reasonable window size.  Now, they're just as broken at any window size.
20:35 hmmmm erm
20:35 sapier no the fonts havent been correct
20:35 sapier the fonts have been SAME size
20:35 VanessaE they were correct as in they didn't overflow their fields.
20:35 sapier them beeing readable was by accident only
20:35 sapier they haven't been readable on high dpi devices
20:37 sapier btw any modder can make his formspecs use fixed size
20:37 sapier then it's gonna stay exactly the way it was by now
20:37 sapier with all positive and negative effects
20:37 sapier hmm
20:37 sapier if we did change the default to fixed sizes would this make everyone happy?
20:38 VanessaE define "fixed size".
20:38 sapier SAME as now
20:38 VanessaE fixed to a certain pixel height, or fixed to a certain POINT size?
20:38 sapier everything as broken as now
20:38 VanessaE yeah, actually that probably would solve most of the issue
20:39 VanessaE but you still need to change the calculations so that fonts are sized in points and based on DPI
20:39 sapier hmm well no it'd not be same every formspec would look like now on 800x600
20:39 VanessaE otherwise this is just gonna happen again sooner or later.
20:39 sapier exactly that way
20:40 sapier screen dpi is independent if you believe this to change anything TRY it ... it's not gonna help
20:41 VanessaE it's only independent if your calculations treat it wrongly.  if you've got a 12-point font, it should display at precisely 4.233 mm high on any screen
20:41 sapier screen_dpi is the config setting you have to change, default is 300
20:41 VanessaE regardless of DPI, but you have to know the screen DPI to calculate how tall to actually render the font
20:42 VanessaE 300?  who the hell has a DPI that high except some retina screens?
20:42 VanessaE that's laser printer resolution there
20:44 sapier wait sorry my test value is 300
20:44 sapier set is 72
20:44 VanessaE well that's better :P ... 72's a tad low, most screens are closer to 100
20:44 russia_nekto devs, i have one idea.
20:44 russia_nekto i think if u route your will to make some simple file structure for artists to allow them change minetest skin (main menu, inventory , hud and etc) what will rise players. because beuaty is why i play in games.
20:44 russia_nekto as i digged today there is no fast technique to beuatify menus in minetest.
20:44 russia_nekto so please, think of that.
20:44 VanessaE but you get what I'm saying
20:44 VanessaE russia_nekto: there's a mod for that.
20:45 russia_nekto VanessaE: where to read about that mod?
20:45 VanessaE russia_nekto: please check the minetest forums and take further questions there or to #minetest
20:45 russia_nekto no help?
20:45 VanessaE russia_nekto: this channel is for core development discussion, not general questions
20:46 russia_nekto is it general?
20:46 VanessaE #minetest is.
20:46 VanessaE please go there.
20:46 jin_xi VanessaE: does the entity detach crash happen on specific locations (any coordinate multiple of 200?)
20:46 russia_nekto i mean my idea is general?
20:47 VanessaE jin_xi: I wasn't able to determine it, since it was a production server I didn't really have much chance to try to reproduce it.
20:47 VanessaE s/determine it/determine that/
20:48 VanessaE russia_nekto: yes.
20:48 russia_nekto u thnk beuaty is not corre?
20:49 VanessaE russia_nekto: this channel is for discussion of the engine code and the like.  go to #minetest and ask your questions there, please
20:53 sapier VanessaE https://gist.github.com/sa​pier/95e7c31c69419502977e
20:54 sapier replace getDisplayDensity in porting.cpp by this version
20:54 russia_nekto the beauty gives inspiration. gives players and again gives inspiration to do more and loop here.
20:54 VanessaE sapier: the irony here is that if you set the default font scaling to fixed, then you could just as well set the main menu to auto-scale (with "fixed" turned on) to replicate the in-game behavior :P
20:54 VanessaE ok, lemme try that
20:54 russia_nekto VanessaE: you advice to ask in #minetest doesnt helps.
20:55 sapier no because I won't change font scaling
20:55 sapier only formspec scaling
20:55 VanessaE sapier: I was kidding :P
20:55 kilbith russia_nekto: we don't care, move away.
20:55 Warr1024 joined #minetest-dev
20:55 sapier hmm
20:56 russia_nekto kilbith: sad really sad
20:56 VanessaE sapier: trying that patch...
20:56 VanessaE russia_nekto: your questions are of a general nature.  this is not the place to ask them.
20:56 sapier XCloseDisplay(x11display);
20:56 sapier add this somewhere after the float lines
20:56 sapier but prior return
20:56 VanessaE sapier: ok, added.
20:57 * VanessaE re-recompiles.
20:57 sapier crap
20:57 sapier doesn't work
20:57 * VanessaE stops.
20:58 sapier hmm wtf why does it work now
20:58 sapier try it
20:58 sapier and tell me your dpi error log should be flooded
20:59 VanessaE ok, I'll try
21:00 * VanessaE waits for compile script...
21:01 VanessaE there we go....
21:01 VanessaE 16:02:08: ERROR[main]: DPIH: 96 DPIW: 96
21:01 VanessaE [     5.620] (II) fglrx(0): DPI set to (96, 96)
21:01 VanessaE so that's a match
21:01 sapier but it doesn't fix your issues does it?
21:02 VanessaE well aside from the font being REALLY BIG now,...  lemme turn the size down first
21:03 VanessaE testing...
21:04 VanessaE haha
21:04 VanessaE Maximum number of clients reachedSegmentation fault (core dumped)
21:05 VanessaE yep, it keeps dieing with that error when I try to start a world
21:05 proller joined #minetest-dev
21:05 VanessaE however, with the font size set to 15 in my config, it's approximately consistent with what I had before, but the formspec is larger on the 800x600 window than before
21:05 VanessaE (larger even, than it was before your patch)
21:06 VanessaE I mean the main menu.
21:07 VanessaE specifically, the formspec is *wider*, but the height seems about the same.
21:08 sapier does anyone know how to find out the correct screen number? hardcoding 0 doesn't work for multidisplay scenarios
21:08 sapier stop messing around with font adjustment size
21:09 sapier you cannot use it same way then before ... if you insist on messing around there I'm just gonna change the setting name to custom_font_size_adjustment
21:09 sapier which is it's actual meaning by now
21:10 VanessaE joined #minetest-dev
21:10 VanessaE er......X didn't like that,.
21:10 VanessaE that last run actually crashed it.  had to REISUB.
21:10 VanessaE what I was about to say was that in-game formspecs, while wider than they should be, had the correct font sizes now, consistent with the main menu.
21:11 russia_nekto left #minetest-dev
21:11 sapier Didn't we solve that issue a couple of days ago when I merged the patch I gave you weeks ago?
21:12 VanessaE apparently not :)
21:12 VanessaE I'd take a screenshot, but I expect it'll crash X again.
21:12 sapier well it's not gonna work on mutlihead at all
21:12 VanessaE I was also trying to say that I'm not able to see the world in the server I connected to, but I can see my HUD, inventory, and chat.
21:13 sapier x11 doesn't provide per display information
21:13 VanessaE ah lemme guess, your code picked up 3200x1200 then
21:13 VanessaE (2* 1600x1200)
21:13 sapier I don't even think the numbers are correct
21:14 sapier no 5120x1080
21:15 VanessaE I mean from mine
21:15 sapier according to x11 it's 1368x292 mm
21:15 sapier most likely it did yes
21:15 VanessaE [     5.220] (II) fglrx(0): clock: 162.0 MHz   Image Size:  367 x 275 mm
21:16 VanessaE maybe because I don't use xinerama?
21:16 sapier maybe
21:16 sapier 292 is right for first screen
21:17 sapier maybe for second too but third is 320 and only 1024 pixels
21:18 sapier it'd be nice to know what would be calculated on  1280x1024 as well as a 4k screen
21:18 VanessaE looks like you can also read it from XRandR
21:18 sapier so a fourth dpi source ;-)
21:18 VanessaE heh
21:19 VanessaE well I suppose that's available on Wayland and Mir also
21:19 VanessaE hm, no... there's a tool for that
21:19 VanessaE bleh
21:19 sapier nice to know we've got 4 for linux ... not even started with windows osX ... hopefully bsd uses same as linx
21:20 sapier +u
21:21 VanessaE meanwhile since that test patch makes things...rather unstable, I'll remove it now
21:24 ShadowNinja sapier: I found a bug.  Type a long character like 'a' or 'M' into the password field, and compare to a short character like 'i' or 'j'.
21:24 ShadowNinja A lot of them rather.
21:24 sapier and?
21:24 ShadowNinja Until it exceeds the field width.
21:24 sapier what's gonna happen?
21:24 ShadowNinja See what happens to the cursor, it's set wrong.
21:25 sapier fix it ;-)
21:25 sapier guess it's there for some time
21:25 ShadowNinja sapier: I don't know how.
21:26 ShadowNinja sapier: My guess is that it's calculating the offset for the text, but printing *s, which puts it off if it doesn't have the exact length of a *.
21:26 ShadowNinja It doesn't work well for the name field either though -- it just doesn't adjust enough for any character.
21:27 * ShadowNinja goes back to log reading...
21:27 sapier but content of the password field is done by irrlicht itself I don't think we're doing something on our own?
21:31 est31 ShadowNinja, did you get a ping when I amended my PR?
21:35 ShadowNinja est31: Erm, what do you mean?
21:35 est31 https://github.com/minetest/minetest/pull/2073
21:35 ShadowNinja sapier: Yes, it's likely an Irrlicht bug, unless it's related to FreeType.
21:36 ShadowNinja est31: Oh, no.
21:37 sapier well ShadowNinja irrlicht doesn't support freetype at all and our freetype to irrlicht code contains bugs
21:37 sapier especially for font details like width height character distances (up/down left/right)
21:37 sapier I tried to understand it but gave up as you need a lot of font specific knowledge I don't have
21:40 ShadowNinja sapier: I didn't make issues because at the time the Android port wasn't merged.
21:42 sapier it's in there about a quarter of a year ShadowNinja that's not a good excuse ;-)
21:42 sapier but that's not really relevant can you check those issues which I couldn't verify? They might be fixed already
21:45 sapier making hud scale independent from formspec is a minor change ShadowNinja do you have time to test?
21:46 ShadowNinja sapier: I'll re-check soon, but I don't have a 0.4.11 build.
21:46 sapier It's not gonna be 0.4.11 it's based uppon latest master
21:47 sapier VanessaE I just merged the X11 dpi code ... case someone wants to provide same for windows ;-)
21:47 VanessaE sapier: that commit works, but you should check how the HUD in dreambuilder behaves with it
21:47 ShadowNinja sapier: I'm not sure what you mean.  Does the size of the displayed formspec change the size of the hud elements?  if so that's definitely broken.
21:48 sapier VanessaE that doesn't matter, dpi detection is required if we wanna do it correct ... unless we wanna wait for irrlicht to implement it we have to take the ugly way of implementing it separate for each screen compositor
21:49 sapier no
21:49 VanessaE sapier: just saying, it throws off the HUD elements' positioning at small screen sizes
21:49 sapier ShadowNinja: the gui_scaling factor affects both
21:49 ShadowNinja sapier: Can you make a recent build for me?
21:49 sapier but as I realized lately you have to change it to a quite low value to get working hud on at least some phones
21:49 * VanessaE turns gui_scaling down from 1.0 to 0.8 and removes her font size settings
21:49 sapier using that small value causes menu to be quite small
21:50 sapier yes I'm gonna upload a recent apk for oyu
21:50 VanessaE ok houston we have a problem
21:51 sapier one? :-)
21:51 VanessaE http://digitalaudioconcepts.com/vanessa/h​obbies/minetest/screenshots/Screenshot%20​-%2001062015%20-%2004%3a51%3a21%20PM.png
21:51 VanessaE look at those two signs.
21:51 sapier we've got a bunch of vanessaE
21:51 VanessaE I'm pretty sure that's not what was fed to the GPU :P
21:51 sapier what's wrong with them?
21:52 VanessaE they are defined in the code and on disk with a wood texture.
21:52 sapier and?
21:52 VanessaE this is mesh corruption here
21:52 sapier I don't see how this could be related to dpi calculation change ;-)
21:52 hmmmm looks to me like a bug with the mod that creates those signs
21:52 VanessaE hmmmm: it was fine earlier today.
21:52 hmmmm =/
21:53 VanessaE I don't see how it could be related either but *shrug*
21:53 * VanessaE restarts
21:53 VanessaE (the client)
21:53 VanessaE textures are fine on the second ruin
21:53 VanessaE run*
21:54 VanessaE that's the first time I've seen in-game corruption of textures on my system.  I also note that the game's window decorations were also corrupted
21:54 VanessaE on the first run
21:55 sapier wtf
21:55 sapier since when is android using xorg?
21:55 VanessaE since never? I thought it had its own windowing system?
21:56 sapier but it seems to set XORG_USED define
21:56 hmmmm uh oh
21:56 sapier wait
21:56 VanessaE what?
21:56 sapier my fault
21:56 hmmmm phew
21:57 sapier wrong ifndef cascade
21:58 VanessaE hm, in-game graphics seem fine on subsequent runs.  must be that roving memory corruption bug?
21:59 VanessaE anyway with gui_scaling at 0.85 and no font size settings, the UIs and fonts are usable now.  this is acceptable to me at least. :)
22:00 VanessaE (will be moreso if you set the default to fixed)
22:00 sapier well only 665 other users to satisfy
22:00 VanessaE haha
22:11 sapier didn't someone provide a file exchange server for minetest?
22:12 VanessaE I think Krock runs something like that doesn't he?
22:12 VanessaE (at least for skins)
22:12 VanessaE (and mods)
22:17 sapier http://www.fileconvoy.com/dfl.php?id=g89​7117ee8502eefc99960535676e62f7b46488772 ShadowNinja, link valid for about 2 days
22:23 cib0 joined #minetest-dev
22:27 sapier of course anyone else is free to test and give feedback too ;-)
22:28 ShadowNinja sapier: The menu is scalled too small.  Usable, but it't 7.5cm X 4cm.
22:28 ShadowNinja scaled*
22:28 sapier did you delete your minetest.conf before?
22:29 sapier it may have the old setting there
22:29 ShadowNinja Nope...
22:29 sapier please try again
22:30 ShadowNinja sapier: Still a bit small, but better.  9cm X 4.5cm.
22:32 ShadowNinja sapier: pitch limiting is fixed...
22:32 sapier ok you should be able to do finetuning by changeing the gui_scaling in settings
22:32 ShadowNinja Jumping too.
22:32 sapier ok I'm closing those issues
22:33 ShadowNinja And chat closing...
22:33 sapier well that's most likely still there, as chat is a regular formspec right now there's no reasonable fix
22:36 ShadowNinja sapier: Aaaand SIGABRT.
22:36 acerspyro joined #minetest-dev
22:36 ShadowNinja In ConnectionReceiveThread.
22:36 sapier hmm do you have more information? out of memory?
22:37 sapier can you post the full error log?
22:37 ShadowNinja Yep, one minute.
22:42 ShadowNinja PMd.
22:43 acerspyro ShadowNinja: Did you really send him the full log via PM, line by line?
22:45 sapier nope he sent me a link ;-)
22:45 ShadowNinja acerspyro: No, of course not.  It's like 200 lines.
22:45 acerspyro lol ok phew
22:47 kaeza joined #minetest-dev
22:48 prozacgod joined #minetest-dev
22:48 ShadowNinja sapier: You can remove the "Fly mode" checkbox from the main menu now, since there's an in-game button.
22:49 sapier true but those buttons are quite ugly my display isn't even big enough to show all of them
22:49 sapier resulting in multiple buttons beeing drawn on top of each other
22:49 acerspyro There's an in-game button? Since?
22:50 FR^2 joined #minetest-dev
22:50 sapier I don't know but it's not a well tested final solution ... I'd suggest providing a settings menu
22:50 kilbith sapier: can you test the PR 2054 on android for me please ? i can't actually
22:50 sapier I don't think those settings are changed each minute
22:50 ShadowNinja acerspyro: https://github.com/minetest/minetest/commi​t/62feade05da387cd6b230e7a0edf558c6fd7c099
22:50 sapier #2ß54
22:50 sapier #2054
22:50 ShadowBot https://github.com/minetest/minetest/issues/2054 -- Reorganizing client and settings tabs by kilbith2
22:51 acerspyro Oh, android?
22:53 sapier And buttons not required for gameplay shouldn't be on main screen
22:54 sapier #2054 breaks settings tab on android
22:54 ShadowBot https://github.com/minetest/minetest/issues/2054 -- Reorganizing client and settings tabs by kilbith2
22:55 sapier but looks like it could be fixed
22:55 sapier you'll have to move tha andoid specific settings block below the middle column
22:55 kilbith hmm.
22:56 acerspyro Scroll bars are a thing
22:56 sapier right now it's overlapping
22:56 sapier no scrollbars without containers and formspec doesn't support those
22:56 kilbith i'd like a screenie
22:57 kilbith or just fix that in my PR, sapier
22:57 kilbith if it's painless enough
22:58 sapier http://imgur.com/L34g1yl
22:59 kilbith oh yes, i'll move that further downward
22:59 sapier btw as you're optimizing the settings menu, could you make anisotropic bi/trilinear filtering a drop down? ;-)
23:00 kilbith yes
23:00 VanessaE no
23:00 sapier a little bit to the left too ;-)
23:00 VanessaE aniso is not mutually exclusive with bi/tri
23:01 sapier are you sure?
23:01 VanessaE the latter two can be in a drop-down, but don't put aniso in with them.
23:01 VanessaE I am positive.
23:01 sapier ok if you tell this it's most likely rigth as I wasn't sure about it
23:01 VanessaE aniso defines the filtering applied over a mipmap, bi/tri is an after-effect, or at least that's how it's supposed to work
23:01 VanessaE some drivers are busted though
23:01 sapier so it's no-filtering/bi-linearfiltering/tri-linear filtering
23:01 VanessaE yes
23:02 VanessaE now for the other part, you can put them into a dropdown also:  none -> mipmap only -> mipmap + aniso
23:02 VanessaE (you'll never use aniso without mipmap)
23:03 sapier no-mipmapping please
23:04 sapier unless you want to add a description line too ;-)
23:04 kilbith 2 dropdowns then ?
23:04 sapier yes one no-mipmap/mipmap/mipmap + aniso
23:04 sapier the other one
23:04 sapier no filtering/bi-linear filtering/tri-linear filtering
23:05 kilbith noted
23:05 sapier wait "no mipmap" of course
23:05 sapier space instead of -
23:08 zat joined #minetest-dev
23:14 sapier https://github.com/minetest/minetest/pull/2074 the hud gui scaling split
23:15 VanessaE from a quick check, lgtm
23:15 VanessaE but I didn't test :P
23:17 y joined #minetest-dev
23:17 * VanessaE tests.
23:18 VanessaE yep, works for me
23:24 loggingbot_ joined #minetest-dev
23:24 Topic for #minetest-dev is now Minetest core development and maintenance. Chit-chat goes to #minetest. Consider this instead of /msg celeron55. http://irc.minetest.ru/minetest-dev/ http://dev.minetest.net/
23:32 EvergreenTree joined #minetest-dev
23:53 VanessaE it seems direct3d8/direct3d9 are not showing up for some people as a video driver option, while opengl et al. are.
23:53 VanessaE on Windows that is
23:56 sapier well someone wanted them to be hidden I guess there's a bug
23:56 sapier for what I remember the list is provided by irrlicht
23:57 Sokomine joined #minetest-dev
23:57 kilbith i *hate* the new font rendering, the menu is totally fucked up with the french words
23:58 sapier yes the full driver list is checked against irrlicht support if irrlicht tells NO we don't show it
23:58 kilbith s/rendering/scaling
23:58 acerspyro kilbith: how so
23:58 sapier kilbith: screenshots please, did you remove manual font adjustments from your config file?
23:59 sapier do you use a custom font?

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