Minetest logo

IRC log for #minetest-dev, 2015-08-18

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

All times shown according to UTC.

Time Nick Message
01:21 sgtbigman joined #minetest-dev
03:36 Hunterz joined #minetest-dev
04:48 Sketch2 https://www.grc.com
04:55 hmmmm uhm... yes?
05:26 hmmmm https://github.com/kwolekr/minetest/commit​/c2d23ff9ccf04dd0e4c2ba8a8026857c034bc947   will push in 30 minutes
05:26 hmmmm ptal if you'd like
05:46 Hunterz joined #minetest-dev
06:22 nrzkt joined #minetest-dev
06:39 nore joined #minetest-dev
06:44 OldCoder I have been debugging for hours. Is anybody able to explain why mts files placed even in an empty world with just minetest_game might not appear?
06:46 hmmmm if the schematic file path is absolute, it will use that path.  otherwise, the schematic file path is assumed to be relative to the current mod path.
06:47 OldCoder There are no error messages. The mtschemplace command says placed the mts. However, it does not appear.
06:48 OldCoder Worked for a 1GB mts file the other day
06:48 OldCoder However, multiple slightly smaller mts files behave as indicated above
06:49 OldCoder Issue has actually occurred for a year, with different git-current sources along the way
06:49 OldCoder Current minetest_game, no other mods but worldedit, and empty world
06:50 hmmmm dunno, it seems to work for everyone else, maybe they can help debug your problem
06:50 OldCoder I suspect it does not work for everyone else; i.e., not that many people load large mts files. However, thank you.
06:51 OldCoder I will confirm this suspicion by seeing if anybody else will try to load specific mts files from the forums. My guess is that nobody has tried to do so for a year but me.
06:51 hmmmm well, file a bug report and upload the failing schematic somewhere
06:51 OldCoder Yes; once I have confirmed that the problem occurs for others.
06:55 nrzkt hmmm: why https://github.com/minetest/minetest/commi​t/c2d23ff9ccf04dd0e4c2ba8a8026857c034bc947 ?
06:56 hmmmm because it hasn't been used by anybody for years and it's silently sucking about 2.5 microseconds of execution time
06:56 hmmmm probably more on other peoples' slower processors
06:57 nrzkt good :)
06:57 hmmmm it's not a lot but it adds up when people have these calls inside of loops
06:58 nrzkt it's right, profiler make game a little bit slow sometimers and should be enabled on debug builds, not release i think
07:00 leat joined #minetest-dev
07:00 leat joined #minetest-dev
07:11 leat joined #minetest-dev
07:21 nanepiwo joined #minetest-dev
07:28 nanepiwo joined #minetest-dev
07:35 nrzkt joined #minetest-dev
07:54 tuy joined #minetest-dev
08:05 Yepoleb joined #minetest-dev
08:14 johnwayne1986 joined #minetest-dev
08:35 deltib joined #minetest-dev
08:36 Calinou joined #minetest-dev
08:37 Ardonel joined #minetest-dev
09:01 Gael-de-Sailly joined #minetest-dev
09:12 leat joined #minetest-dev
09:14 johnwayne1986 joined #minetest-dev
09:27 johnwayne1986 Hi all!
09:30 johnwayne1986 When I try the unit tests with the latest commit, the test readF1000(is) == 53.534f at test_serialization.cpp:299 fails. Can somebody confirm this? Is this meaningful? (Just wanted to tell you that I noticed that.)
09:41 err404 joined #minetest-dev
09:58 johnwayne1986_ joined #minetest-dev
10:01 johnwayne1986 joined #minetest-dev
10:04 kahrl johnwayne1986: the tests pass on my machine
10:04 kahrl it might depend on the FPU rounding mode or whatever though
10:04 johnwayne1986 interesting...
10:05 kahrl the original version of that test was UASSERT(fabs(readF1000(is) - 53.534f) < 0.005);
10:06 kahrl I wonder why it was changed
10:08 kahrl the commit message of 42cf5e972d says it should be safe to do an exact comparison, but I don't understand exactly why
10:08 kahrl hmmmm: ^
10:15 Darcidride joined #minetest-dev
10:16 johnwayne1986 so, I get: f: 53.534, diff: 3.8147e-06
10:29 nanepiwo could someone pls reopen #2985 ?
10:29 ShadowBot https://github.com/minetest/minetest/issues/2985 -- Player's direction of view pitches up slightly when clicking
10:32 wischi joined #minetest-dev
10:40 nore nanepiwo: see my comment on #2985
10:40 ShadowBot https://github.com/minetest/minetest/issues/2985 -- Player's direction of view pitches up slightly when clicking
10:45 ElectronLibre joined #minetest-dev
10:48 Zeitgeist_ joined #minetest-dev
10:49 err404 joined #minetest-dev
10:50 AnotherBrick joined #minetest-dev
10:51 kahrl johnwayne1986: sorry for the delay, had to shower & eat and have to leave now, but just so hmmmm has the info when he comes back: what is your OS/compiler?
11:01 johnwayne1986 I have to leave as well in a few minutes or so...
11:03 johnwayne1986 I use LMDE 1, which is outdated and supported with security fixes only. Compiler is g++ 4.8.2.
11:06 johnwayne1986 If I run 'file minetest' it tells me: minetest: ELF 64-bit LSB  executable, x86-64, version 1 (GNU/Linux), dynamically linked (uses shared libs), for GNU/Linux 2.6.32
11:07 johnwayne1986 I am running 3.11-2-amd64 though. Is the output 2.6.32 normal?
11:29 Fritigern joined #minetest-dev
11:44 sgtbigman joined #minetest-dev
11:50 kahrl johnwayne1986: the "for GNU/Linux 2.6.32" is from the PT_NOTE section of the executable and the value depends on what --enable-kernel option was given when glibc was built
11:51 err404 left #minetest-dev
11:54 TenPlus1 joined #minetest-dev
11:54 TenPlus1 Hi folks
11:56 TenPlus1 Question. is gui_scaling in minetest.conf meant to be server specific ??
11:57 Gael-de-Sailly joined #minetest-dev
12:00 kahrl TenPlus1: it's a client setting
12:00 nanepiwo @tenplus GUI_scaling is client side, for any formspecs (like main menu and in multiplayer servers)
12:01 TenPlus1 I read the gui_scaling setting and it always reads the server's setting, never the client
12:01 TenPlus1 so personal scaling for hud always appears at 1.0
12:03 proller joined #minetest-dev
12:03 TenPlus1 am using: local gui_scale = tonumber(minetest.setting_get("gui_scaling"))  .. am I doing something wrong ?
12:05 proller joined #minetest-dev
12:05 kahrl minetest.setting_get reads serverside settings, yes
12:05 kahrl client settings are generally not sent to the server
12:05 Fritigern joined #minetest-dev
12:05 TenPlus1 so the gui scaling is useless unless you are playing singleplayer
12:06 kahrl why useless? it's applied to all formspecs automatically
12:06 TenPlus1 formspec yes, hud display no...
12:07 TenPlus1 there is no way for a mob to read player gui_scaling to assist in hud output...
12:07 TenPlus1 *mod
12:07 kahrl it's not applied to hud? that might be a bug
12:07 TenPlus1 that's the point though, I want to be able to scale up HUD graphics for HD and 4k screens but cannot access player scaling setting...
12:09 kahrl well the proper solution would be to apply it to hud in the client, not to do it serverside
12:10 TenPlus1 not possible if conencting to server...
12:10 H-H-H joined #minetest-dev
12:10 kahrl I mean in the C++ code
12:11 TenPlus1 oh, gotcha...
12:12 TenPlus1 maybe I could put in a github issue to read it always from client
12:13 TenPlus1 or have a personal minetest variable for the machine it's run on with scaling number itself
12:14 proller joined #minetest-dev
12:17 RealBadAngel agree, only way is to do that client side
12:17 RealBadAngel server dont need to know actual clients resolution at all
12:17 TenPlus1 agreed...  will post github issue for that one
12:17 TenPlus1 thanks
12:36 Calinou joined #minetest-dev
12:37 SopaXorzTaker joined #minetest-dev
13:06 srifqi joined #minetest-dev
13:12 johnwayne1986 joined #minetest-dev
13:22 Niebieski joined #minetest-dev
13:37 zupoman joined #minetest-dev
13:37 zupoman joined #minetest-dev
14:15 H-H-H joined #minetest-dev
14:35 Sketch2 joined #minetest-dev
14:37 VanessaE joined #minetest-dev
14:50 hmmmm joined #minetest-dev
14:53 Sketch2 joined #minetest-dev
14:59 Yepoleb_ joined #minetest-dev
15:06 Fritigern joined #minetest-dev
15:07 SopaXorzTaker joined #minetest-dev
15:08 nrzkt joined #minetest-dev
15:12 thatgraemeguy joined #minetest-dev
15:30 damiel joined #minetest-dev
15:37 paramat joined #minetest-dev
15:41 paramat RealBadAngel #3073
15:41 ShadowBot https://github.com/minetest/minetest/issues/3073 -- tileable_{horizontal,vertical} flags cause fps decrease
15:43 paramat nore sfan5 game#640
15:43 ShadowBot https://github.com/minetes​t/minetest_game/issues/640 -- Fire: Slow down spread and reduce lua load by paramat
15:45 paramat perhaps i'll merge game#639 later unless anyone knows a better implementation
15:45 ShadowBot https://github.com/minetes​t/minetest_game/issues/639 -- Boats: Fix sinking through boat when detaching, by CProgrammerRU by paramat
15:51 kahrl why is there special code in the shader to handle tileable_* flags at all?
15:51 kahrl wouldn't it be enough to set TextureWrapU/TextureWrapV in the material settings?
15:51 paramat to fix a visual glitch i think?
15:51 kahrl most likely with no fps drop
15:52 kahrl the glitch is the same glitch that happened with wield meshes
15:52 kahrl I think
15:52 kahrl I fixed that glitch by setting TextureWrapU/TextureWrapV
15:52 kahrl to video::ETC_CLAMP_TO_EDGE
15:57 RealBadAngel paramat, i do have fixes for that rdy
15:58 paramat ok
15:58 RealBadAngel when we made fsaa fixes it looks like setting uv wrapping was also enough to fix tiling issues
15:59 RealBadAngel i have already removed that code from shaders
15:59 Hunterz joined #minetest-dev
15:59 RealBadAngel sidenote, its weird he noticed slowdown with it
15:59 RealBadAngel the mechanism is here all the time
16:00 RealBadAngel anyway, the code in question will be just gone
16:01 RealBadAngel i will help my wife with the dinner and then prepare PRs with latest stuff
16:01 RealBadAngel paramat: https://imgrush.com/vxfrNtQojkvy.png
16:02 RealBadAngel now, tell me what do you think about it when its complete :)
16:03 Zeitgeist_ joined #minetest-dev
16:03 paramat crystals look ok themselves but it's certainly for an eye-candy fork
16:09 RealBadAngel partially youre right, but those are named crystals already
16:10 RealBadAngel and other hand having quite big number of ores that do actually look the same its not a candy, its eye boring ;)
16:12 RealBadAngel anyway im lookin on ideas (and conditions to fullfill) to make crystals grow
16:12 RealBadAngel if any of you have some please let me know
16:24 paramat in minetest the crystals are (i think) fragments of mese blocks, so crystals growing in these natural-forming shapes is something very different
16:28 nrzkt joined #minetest-dev
16:30 Robert_Zenz joined #minetest-dev
16:34 Krock joined #minetest-dev
16:36 hmmmm hmm whatever happened to est?
16:36 hmmmm he dropped off the earth with perfect timing
16:36 hmmmm the release is basically waiting on him alone :p
16:42 paramat i have 2 more PRs for game, need perhaps 1 more day
16:46 paramat also there's the tiling bug mentioned above, RBA possibly has a PR for engine.
16:47 hmmmm great
16:48 hmmmm hmm est is active on github but not.. here
16:48 hmmmm wonder how the inventory replace bug is coming along
16:49 RealBadAngel im preparing my fixes atm, gimme an hour or two
16:55 Krock \o/ MSVCRT linking errors
16:56 MinetestForFun joined #minetest-dev
17:12 paramat left #minetest-dev
17:22 RealBadAngel hmmmm, some1 said that est31 will be off for a few days, reason unknown
17:22 crazyR joined #minetest-dev
17:23 FR^2 joined #minetest-dev
18:00 est31 joined #minetest-dev
18:01 est31 hmmmm, I have said I'll be away for some days
18:02 est31 but now im back!
18:02 est31 I'll fix that inventory bug, then we wait for paramat to do the release,
18:02 est31 then lets hope blockmen does a build
18:04 * Krock tries to compile
18:09 * nrzkt thinks est31 should go in holidays another time :D
18:09 Krock a library uses the /MD comiler flag instad of /MT .. I think a clean build is neccessary :(
18:10 est31 nrzkt, if we went after original schedule, there would have been no problem
18:10 est31 now there is neither
18:11 * nrzkt thinks germans works too much and should work like mexicans
18:18 hmmmm lol
18:18 hmmmm isn't there a labor shortage in germany
18:19 est31 depends on the job I guess
18:20 nrzkt if you work with a contract without a minimum salary like germans like to decrease their unemployment, you could earn 100$ per month for 40 hours a weak :D
18:20 est31 also part of it is self caused
18:20 nrzkt week*
18:20 est31 we do have minimum wage nrzkt
18:21 est31 since this year
18:21 est31 or the year before
18:21 est31 but the great coalition that took over gvt in 2014 promised minimum salaries
18:21 est31 was demand of the SPD
18:22 est31 germandy does have doctor shortage on the countryside
18:23 est31 many german doctors go to norway or switzerland where they are better paid
18:23 est31 or just stay inside the cities
18:23 * Krock misses the swiss german in hospitals...
18:24 est31 lol
18:26 Krock est31, win32 build is coming in ~1 min, if interested
18:27 * Krock slaps himself after noticing the actual release was meant
18:31 Hunterz joined #minetest-dev
18:48 Calinou joined #minetest-dev
19:18 tuy joined #minetest-dev
19:26 tuy joined #minetest-dev
19:27 proller joined #minetest-dev
19:49 rubenwardy joined #minetest-dev
20:03 Player_2 joined #minetest-dev
20:05 Gael-de-Sailly joined #minetest-dev
20:09 Icedream joined #minetest-dev
20:09 Icedream joined #minetest-dev
20:37 paramat joined #minetest-dev
20:38 nrzkt joined #minetest-dev
20:40 paramat nore sfan5 i'd like to merge game#639 very soon, any objections? also any comments on game#640 ?
20:40 ShadowBot https://github.com/minetes​t/minetest_game/issues/639 -- Boats: Fix sinking through boat when detaching, by CProgrammerRU by paramat
20:40 ShadowBot https://github.com/minetes​t/minetest_game/issues/640 -- Fire: Slow down spread and reduce lua load by paramat
20:40 sfan5 639 is fine
20:43 paramat apart from those there are also 2 PRs coming from RBA then that's everything
20:57 tuy joined #minetest-dev
21:00 est31 joined #minetest-dev
21:05 tuy joined #minetest-dev
21:06 tuy joined #minetest-dev
21:06 proller joined #minetest-dev
21:32 paramat now merging game 639
21:37 paramat merge complete
21:38 nrzkt paramat: you add many empty lines....
21:40 paramat yes separating functions with 2 newlines for clarity
22:00 rom1504 why not 5 newlines
22:01 RealBadAngel paramat, #3079
22:01 ShadowBot https://github.com/minetest/minetest/issues/3079 -- Remove use of engine sent texture tiling flags - theyre no longer needed by RealBadAngel
22:01 RealBadAngel thats for the engine
22:01 est31 RealBadAngel, why are they no longer needed now?
22:01 est31 also is the bug from cheapie still fixed?
22:01 est31 s/from/reported by/
22:02 RealBadAngel theyre no longer needed becaue setting UV clamping in the material is enough
22:03 est31 is that done by the pr?
22:03 RealBadAngel kahrl solution for fsaa bug solved also problems with tiling
22:03 est31 ok
22:03 RealBadAngel in this pr i removed code that did the same in shaders
22:03 est31 so are the tiling flags totally unneccessary?
22:04 est31 or do they serve some purpose?
22:04 RealBadAngel absolutely no
22:04 RealBadAngel theyre needed
22:04 RealBadAngel based on them material is configured
22:05 RealBadAngel what was not needed is shaders doing it again
22:05 est31 ok
22:06 RealBadAngel theres also the code that generates the textures flag - it should also be changed
22:06 est31 how
22:06 est31 ah
22:06 est31 you mean the code that sends the #defines?
22:06 est31 err
22:06 est31 the code that generates the texture flags texture?*
22:06 RealBadAngel no, i mean the code that generates the texture for shaders with flags
22:07 RealBadAngel yes, but i do believe its harmless now and i would like to use it for another purposes
22:07 est31 ok
22:08 RealBadAngel thats why i left it unchanged
22:08 RealBadAngel i wil make now two PRs for the game
22:09 paramat ok
22:27 RealBadAngel paramat, https://github.com/minetest/minetest_game/pull/641
22:27 RealBadAngel first one
22:31 RealBadAngel https://github.com/minetest/minetest_game/pull/642
22:31 RealBadAngel and the second
22:33 wilkgr joined #minetest-dev
22:47 RealBadAngel ouch, minetest doesnt have marble and granite, fixing 642 now
22:49 VanessaE so I don't recall seeing where you said what happens if a user switches to some other texture pack which doesn't supply all the normalmaps.  do these get used then?  or only the ones their texture pack supplies?
22:51 RealBadAngel if some1 wants to use displacement mapping he needs to get tp pack that supports it
22:52 VanessaE let me rephrase that
22:52 RealBadAngel i know what you mean
22:52 VanessaE if some texture pack has NO normals at all, what gets used for normals?  nothing?  these defaults?  something pulled from the aether?
22:52 RealBadAngel normal maps are default textures
22:52 RealBadAngel tp that does not support them is not complete
22:53 VanessaE FAIL
22:53 RealBadAngel displacement mapping is core feature
22:53 VanessaE in effect you break every texture pack in existence except for Haven
22:53 RealBadAngel this is not texture pack
22:53 est31 grrr, there seem to be other bugs with inventory replacements too
22:54 est31 e.g. if the destination inventory is full
22:54 est31 and you take from a creative inventory
22:54 VanessaE special handling of normals is needed - only pull normalmaps from the user's texture pack if there is a texture pack in use.
22:54 est31 it deletes one element from the destination inventory
22:54 est31 without asking
22:54 hmmmm can players crouch in minetest?
22:54 VanessaE hmmmm: no
22:54 hmmmm ahh okay
22:54 est31 why
22:54 hmmmm is that wanted
22:54 VanessaE possibly.
22:55 RealBadAngel could be useful
22:55 VanessaE there is "sneak" but that really means "walk slowly and more carefully".
22:55 est31 you can sit on chairs
22:55 est31 but not very reliably
22:56 VanessaE RealBadAngel: my point being, just about every actual texture pack in existence is going to be devoid of normalmaps.  They were never "incomplete" before, you can't go retroactively making them incomplete now.
22:57 VanessaE not when there are 345 textures in minetest_game alone.
22:57 VanessaE that's too much breakage.
22:57 paramat ok
22:57 RealBadAngel enabling displacement mapping doesnt mean selecting new texture pack at all
22:58 VanessaE I'm not talking about that, RealBadAngel
22:58 RealBadAngel yes you want that
22:58 RealBadAngel for normal maps that are for core feature
22:58 VanessaE I'm talking about if it's *already* enabled for say, several days/weeks, and the user *then* selects a texture pack that has partial support for normalmaps
22:59 hmmmm what events are missing here that you would think is useful for a client mod?  http://fpaste.org/256506/39938718/
22:59 VanessaE are you saying that normalmaps included with minetest_game will get used when a texture pack does not supply them?
22:59 RealBadAngel yes, hightlight the word "partial" - it means its incomplete
22:59 VanessaE then it's a fail
22:59 VanessaE you can't do that,.
22:59 VanessaE you HAVE to make sure normalmaps are always pulled from the same dir as their corresponding color texture
23:00 VanessaE (regardless whether it's a texture pack or a mod or a game)
23:00 RealBadAngel cant do that im afraid
23:00 VanessaE why not?
23:00 RealBadAngel that would break even more things
23:00 RealBadAngel and even the features
23:01 VanessaE for normalmaps?
23:01 RealBadAngel ability to override normals for example
23:01 VanessaE it would break less than what you're proposing to do now.
23:01 VanessaE wat
23:01 VanessaE foo/bar/blahblah.png --> foo/bar/blahblah_normal.png
23:01 VanessaE that's all.
23:01 RealBadAngel override_normal.png
23:01 VanessaE anything more complicated than that breaks shit that should not break.
23:02 est31 or alternative solution, just rename all textures
23:02 VanessaE hmmmm: what about if the player is merely punched (perhaps by a MOB) with no HP change?
23:02 est31 then there wont be breakage
23:03 hmmmm yep got it
23:03 est31 you'll just see that texture packs wont work
23:03 VanessaE ...
23:04 est31 but VanessaE's point is valid
23:04 RealBadAngel lemme show you another situation
23:04 est31 until we get to the point where texture packs authors only want to change the normals, but not the original texture
23:04 est31 or only the original texture, but not the normals
23:04 RealBadAngel you decide to use texture pack, that has partial support for textures
23:05 paramat fallback to auto-generated normalmaps?
23:05 RealBadAngel if you enable it, you will loose all the normal maps
23:05 VanessaE paramat: doesn't work out so well for some textures.
23:05 est31 hmmmm, what should CSEVT_NETWORK do
23:05 paramat yeah
23:05 est31 also, we shouldnt have keyboard and mouse input
23:05 est31 imagine android users
23:05 hmmmm that event is triggered with the message payload any time a TOCLIENT_SCRIPT_MESSAGE is received
23:06 hmmmm yes about that
23:06 est31 should perhaps be renamed
23:06 VanessaE est31: call it generic HID input then?
23:06 RealBadAngel texturing nodes in minetest engine does require 2 files, diffuse map and a normal map
23:06 hmmmm if a mod doesn't want any inputs it's ignored
23:06 VanessaE RealBadAngel: in other words, "so, the proposal is to make twice as much work for texture authors".
23:06 est31 we shouldnt offer apis this low level
23:06 RealBadAngel any texture pack that wants to alter existing textures and be full, have to provide both files
23:06 hmmmm mods have to "subscribe" to events first or else they're not even queued, saving CPU from doing useless allocations/semaphore signals
23:06 hmmmm why not?
23:07 est31 otherwise we get to the use key not availiable on android situation
23:07 est31 just even worse
23:07 RealBadAngel its not a proposal, its requirement
23:07 VanessaE RealBadAngel: then the requirement needs to be re-evaluated
23:07 hmmmm well then
23:07 VanessaE this is NOT 0.4.13 material.
23:07 hmmmm mod writers will need to adapt
23:07 est31 hmmmm, its not about adapting
23:07 RealBadAngel this is core feature for the very long time
23:07 est31 its about giving mod writers a safe sandbox
23:08 est31 and not letting mod writers getting into our way
23:08 VanessaE the vast majority of texture authors just want to change the textures, not any normalmaps, and you're requiring a skill few graphics artists possess.
23:09 VanessaE you need to pull normals from the same dir as the texture came from or generate a "zero" image for those textures.
23:09 RealBadAngel if texture pack creator havent made maps he should clearly state and warn user that tp DOESNT SUPPORT bumpmapping and displacement mapping
23:09 est31 hmmmm, you dont design good apis by adding such specific features
23:09 VanessaE but what if they've created SOME but not ALL?
23:09 est31 right now formspec is very semantic
23:09 VanessaE like they don't think e.g. gravel needs a normalmap?
23:09 est31 and thats good
23:10 wilkgr left #minetest-dev
23:10 hmmmm nobody said we were going to expect mod writers to handle keys themselves
23:10 est31 e.g. you can set up an interface where you say the name of the button
23:10 hmmmm and mouse moves
23:10 VanessaE think cartoonish textures.  usually those won't want normalmaps at all, but the user wants normalmaps enabled for all those other mods that have them
23:10 est31 hmmmm, so why offer it to them
23:10 hmmmm in case they need it
23:10 est31 they should give us their usecase
23:11 est31 and we build them then a custom plugin for that usecase
23:11 est31 that also works on touch
23:11 est31 otherwise we can forget porting minetest to game consoles
23:11 hmmmm this can all be abstracted
23:11 VanessaE est31: what about just registering *actions* rather than keys, sorta like what hmmmm already specified
23:11 est31 sooner or later every mod will require the custom input events
23:11 RealBadAngel http://i.imgur.com/RX2BUjh.png
23:12 est31 because to work around engine bugs
23:12 est31 or whatever
23:12 RealBadAngel are you sure that cartoonish tps doesnt need normal maps?
23:12 est31 even more people will cry out "lets do this in lua"
23:12 VanessaE RealBadAngel: I wasn't talking about sphax..
23:12 est31 we'll get to a point where there wont be difference between hud or formspecs
23:12 est31 where its all just a big cesspit
23:13 hmmmm there is a difference between hud and formspecs
23:13 est31 of mod incompatibilities
23:13 hmmmm a formspec is a hud element that displays the mouse when open and accepts input
23:13 hmmmm also formspecs are heirarchical
23:13 hmmmm well
23:14 est31 no mouse
23:14 hmmmm in the future that's the way it'll be
23:14 RealBadAngel VanessaE, but anywyay even if i will agree with you, theres no way to make it working the way youre describing now
23:14 est31 thats not part of the definition
23:14 est31 mouse isnt available on touch devices
23:14 hmmmm god dammit est
23:14 VanessaE RealBadAngel: then the normals can't be included with minetest_game.
23:14 est31 <hmmmm> in the future that's the way it'll be
23:14 hmmmm i understand this
23:14 est31 thats very arrogant to sa
23:14 est31 y
23:14 hmmmm when I say "shows the mouse cursor" obviously it's going to do the correct thing on android
23:15 est31 its like if RBA said "in the future we will have that tune by the legend"
23:15 hmmmm on android that means that touching the screen will behave like a mouse click on a gui element instead of punching a node
23:15 hmmmm well I have a vision
23:15 RealBadAngel VanessaE, then remove displacement mapping from the core
23:15 hmmmm I want Hud and Formspecs to be abstracted to the point where there isn't much of a difference between them
23:15 VanessaE ...
23:15 RealBadAngel without those files its not needed
23:16 VanessaE RealBadAngel: this is the realm of texture packs, not default game.
23:16 est31 hmmmm, what you propose isnt abstraction
23:16 RealBadAngel displacement mapping is not texture pack feature
23:16 VanessaE if you want to propose your normals as an alternate texture pack that the user can turn on, that is itself included with...something... I'm fine with thaty.
23:16 VanessaE -y
23:16 hmmmm from what I understand, you're afraid that if we give modders the power to detect raw clicks or whatever, they'll misuse it by hacking together their own GUI system?
23:16 est31 you propose to require mod makers to work for support specific devices
23:16 est31 supporting*
23:16 RealBadAngel users are selecting it and reporting - that doesnt work!
23:16 est31 hmmmm, yes
23:16 hmmmm I am saying that if they want to use low-level events, they need to be aware of the risks involved
23:17 hmmmm if they want to be safe, they should just stick to the high-level events
23:17 RealBadAngel because the game without those files is incomplete
23:17 RealBadAngel not the texture pack
23:17 est31 hmmmm, current attitude has been to offer a safe sandbox for modders
23:17 hmmmm in every GUI toolkit, the user has the ability to trigger events on mouse movement
23:17 VanessaE the game will *always* be "incomplete", and lack of normalmaps does not make it any worse.
23:17 est31 its also the reason why we dont have version string displays
23:17 hmmmm do you ever see people reimplementing their own gui?
23:18 est31 we are so extreme, we dont even say mods which version we run
23:18 est31 because of fear they could use it for feature detection
23:18 est31 and not detecting the features themselves
23:18 est31 what you propose is much more intrusive
23:19 hmmmm erm, 'intrusive'?
23:20 hmmmm I'm just making sure the power is there in case modders need it
23:20 hmmmm I am not going to arrogantly pretend that we know better than they do
23:20 est31 client scripting is for allowing to control the camera or so, not to make minetest an interface between lua and irrlicht
23:20 RealBadAngel also VanessaE if texture pack maker would want to change only normal map for the texture
23:20 RealBadAngel your proposal would make it impossible
23:21 VanessaE better to require a texture to go with a normalmap than to require a normalmap to go with a texture.
23:21 est31 hmmmm, we dont know better than they do, but thats what why the c++ level isn't final
23:21 VanessaE if they want to only change the normals, they can copy the default textures that go with them and leave them unchanged,.
23:21 VanessaE come on, really, that's about as far into the corner as a corner case gets.
23:22 est31 hmmmm, minetest is good because modding is easy, making it more extendible doesn't help
23:22 hmmmm alright, I'll leave those out for now
23:22 est31 there are tons of super extendible engines with super easy languages out there
23:22 est31 take unity
23:22 paramat RealBadAngel, is 641 tested? you know what to look for so it's best tested by you
23:23 est31 note, I'm not against an actually customizeable "gui canvas" where you can e.g. make client side terminals
23:23 est31 there you can process mouse output etc
23:23 est31 just it should be clearly separated
23:23 RealBadAngel paramat, just enable displacement mapping with this commit and without. and look at grass
23:23 est31 as its own formspec element
23:24 est31 or "new" generic gui
23:24 RealBadAngel without 641 grass will be messed up
23:25 hmmmm btw we're calling the new gui "formspec" but it's definitely not going to share any resemblence to formspec
23:25 hmmmm formspec is bad and it needs to go
23:25 est31 dunno share the opinion with you that its especially bad
23:25 est31 but shrug, if you want to replace it, go on
23:25 paramat sure but please test it yourself as i'm unfamiliar with the glitch
23:26 est31 as long as we will have ways to maintain compatibility with old mods
23:26 est31 and the new api will be as easy to use as formspecs
23:26 hmmmm three times, I've went to add a feature using formspec and I've been unsatisfied to the point where it made me want to scrap what it was i was working on
23:26 paramat we'll add 642 but then please can we remove the 'tree top' normalmaps because these are cut logs and need a flat surface?
23:27 hmmmm formspec is junk that was created to fill a very niche type of GUI form and then people expanded upon it until it became what it is present day
23:27 RealBadAngel removing the normal map wont make it flat at all
23:27 RealBadAngel autogen will make it bumped too
23:27 hmmmm also formspec is not easy
23:27 hmmmm it's confusing as all hell
23:27 paramat ah
23:28 RealBadAngel hmmmm, but each and every mod out there is using it
23:28 hmmmm if we want a text-based GUI description language, let's use XML instead
23:28 hmmmm this is one of the things i wanted to do
23:28 est31 XML is super ugly
23:28 hmmmm there should be several frontends
23:28 est31 you need end tags
23:28 hmmmm formspec front end, xml front end, etc.
23:28 est31 end tags bear no information
23:28 est31 ok they do
23:28 RealBadAngel removing formspec will efectively make all mods not working
23:29 est31 but still, formspecs are straightforward
23:29 est31 why xml is shiny dynamic
23:29 paramat well we'll sort tree tops out later, 642 can go in for relesae
23:29 est31 formspecs are the lua way
23:29 paramat (release)
23:29 RealBadAngel paramat, i can make that one more flat on request
23:30 hmmmm XML was made for this exact type of bullshit
23:30 paramat yes please if you have time, completely flat
23:30 RealBadAngel have you actually checked how tree tops look like with it?
23:30 hmmmm it might not be the most efficient design but there is loads of support for it
23:30 est31 hmmmm, as things to add to your list, I'd propose CSEVT_PLAYER_MOVE, CSEVT_NODE_PUNCH, CSEVT_NODE_DIG, CSEVT_NODE_PLACE
23:30 paramat well i know they need to be flat
23:31 hmmmm est, I have added more
23:31 RealBadAngel paramat, https://imgrush.com/mXL-Y3TCcykB.png
23:31 est31 hmmmm, whats your new list
23:31 hmmmm http://fpaste.org/256522/40683143/
23:31 hmmmm so
23:31 hmmmm just to be clear:
23:31 paramat i'll merge 641 once you've tested it
23:31 hmmmm client side scripting runs in a separate thread
23:31 est31 yes
23:31 hmmmm events are posted to it
23:31 est31 yes
23:31 hmmmm everything is asynchronous here
23:31 est31 yes
23:31 hmmmm no laggy bullshit
23:32 hmmmm we learned our lesson from the server modding
23:32 est31 well yeah
23:32 est31 separate thread is a good idea
23:32 RealBadAngel paramat, 641 is made weeks ago
23:32 hmmmm every time one of these events happen an event gets posed to the client m_script
23:32 hmmmm the script thread executes the lua handler for this event
23:33 hmmmm s/posed/posted/
23:33 RealBadAngel so i believe its tested enough ;)
23:33 est31 I'd rename CSEVT_PLAYER_CHATMESSAGE ---> CSEVT_CHATMESSAGE
23:33 hmmmm then I have some ideas for what the lua api will look like
23:33 hmmmm by the way, do clients get individual player join/leave events?
23:33 hmmmm probably not, right?
23:34 est31 they get it
23:34 est31 as chat messages
23:34 hmmmm they do get it
23:34 hmmmm yeah exactly
23:34 est31 also I think it might be possible they even get it more low level too
23:34 est31 for the f10 console completion
23:34 est31 s
23:34 hmmmm that's what I was asking mostly
23:34 hmmmm hmm let's see
23:34 RealBadAngel VanessaE, atm only solution to your request would be moving all the normal maps to another folder and forcing it to be texture pack
23:34 est31 but it might be horribly hacked
23:34 est31 by the client parsing server sent messages
23:35 RealBadAngel theres no other solution
23:35 VanessaE RealBadAngel: then do that.
23:35 est31 hmmmm, we should add some API for completion too
23:35 hmmmm nick completion?
23:35 est31 or command completion
23:35 hmmmm ehh
23:35 est31 in f10 console
23:36 est31 or in the t console
23:36 RealBadAngel VanessaE, but then core feature will become broken. i would like to hear other voices on that
23:36 hmmmm we'll see what we can do later
23:36 est31 ok
23:36 hmmmm hrm
23:36 hmmmm i'm not seeing a player join/leave notify packet
23:37 est31 the f10 console does maintain such a list
23:37 est31 so it seems it parses it...
23:37 hmmmm are you sure??
23:37 est31 no
23:37 hmmmm I am not seeing it
23:37 hmmmm well nevermind that
23:38 paramat okay will merge 641
23:39 hmmmm let's not discuss anything more until i get this portion done
23:39 est31 what portion
23:39 hmmmm the amount of discussion about a feature is inversely proportional to how fast it actually gets done
23:39 hmmmm just the basic
23:40 est31 we should discuss before there is an implementation
23:40 hmmmm my first goalpost is to have the server send client scripts as media, client loads them, and posts scripting events to them
23:40 est31 otherwise the implementation might be wasted
23:40 hmmmm we already agree on what i'm doing so far
23:40 est31 yes
23:41 hmmmm so right now I just need to add in the event hooks
23:41 hmmmm and then I'm pretty much done
23:41 est31 ok
23:41 hmmmm but I'd rather focus on getting 0.4.13 out
23:41 est31 agreed
23:42 hmmmm less talking more doing
23:43 est31 well, I do have PRs around that can be merged sooner or later
23:43 est31 e.g. my utf chat pr
23:43 hmmmm how about 3058
23:44 est31 #3058
23:44 ShadowBot https://github.com/minetest/minetest/issues/3058 -- Modstore: show an ui error if retrieving failed by est31
23:44 est31 I don't think we should add http_response_code to the exception object
23:45 hmmmm why
23:45 est31 a JsonFetchException is just noting that something went wrong during fetching json information
23:45 est31 it might be a wrong http code
23:45 est31 it might be dns failure
23:45 hmmmm so if it's 0 then that means it wasn't a failure at the http level
23:46 est31 or just the returned thing isnt json
23:46 est31 also what would it use us
23:46 hmmmm i just want there to be a friendlier message
23:46 hmmmm can't you make it say
23:46 est31 I'd go with what VanessaE proposed
23:46 hmmmm "Oops!  Sorry about that, mmdb seems to be having problems right now."
23:47 est31 "Sorry but there was a problem with loading the modstore. Its technical message, which you should include in error reports, is:"
23:47 hmmmm no don't do that
23:47 hmmmm you're going to encourage useless bug reports for a website 404 error
23:48 hmmmm this is why I want to know if it's an HTTP error vs. some other kind of error I guess
23:48 est31 perhaps we should point out then that the modstore might be offline because of choice?
23:48 hmmmm is it some operational problem or is it an engine problem
23:48 est31 then a bool vale suffices
23:48 est31 is_http_404
23:48 hmmmm i do NOT want our github to get spammed with operational problems
23:49 hmmmm no, is_http_error probably
23:49 est31 no
23:49 hmmmm what if the webserver gets misconfigured and it posts 503 or something
23:49 est31 is_http_error might be a problem of the wrong request too
23:50 est31 hmmmm, what if the webserver is just simply gone, because the domain isnt hosted anymore?
23:50 est31 then its no 404
23:50 hmmmm right
23:50 est31 it might be a dns problem
23:50 est31 or timeout
23:50 hmmmm do we want users to file /bug/ reports over that?
23:50 est31 or tcp error
23:50 hmmmm I want to be able to differentiate between actual bugs and imaginary bugs
23:50 est31 hmmmm, agreed
23:50 hmmmm so
23:50 hmmmm however that's best accomplished
23:50 est31 but we dont solve it by adding http codes to the exception
23:50 hmmmm but right now i need to go take a shit
23:51 hmmmm bbl
23:56 wischi2 joined #minetest-dev
23:57 paramat now merging game 641

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