Minetest logo

IRC log for #minetest, 2018-01-03

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

All times shown according to UTC.

Time Nick Message
00:19 twoelk|2 joined #minetest
00:20 lanlink joined #minetest
00:47 MaybeDragon joined #minetest
00:50 luk3yx joined #minetest
00:57 whitephoenix joined #minetest
01:28 luk3yx left #minetest
01:50 pgimeno_ joined #minetest
01:53 Animetrom1 joined #minetest
01:53 Animetrom1 left #minetest
01:54 Animetrom joined #minetest
01:55 Animetrom Hello guys!
01:55 htur hi
01:55 BillyS hi
01:55 rubenwardy yo
01:55 Animetrom Happy New Year! (:
01:55 BillyS same to you
01:55 htur Happy new year :)
01:56 Animetrom Thanks!
01:58 EvergreenTree joined #minetest
01:59 Animetrom Playing too much Minetest lately?
02:00 nowhere_man joined #minetest
02:01 Animetrom left #minetest
02:15 nowhere_man joined #minetest
02:22 behalebabo joined #minetest
02:31 dax left #minetest
02:42 SonosFuer joined #minetest
03:27 Billre joined #minetest
04:11 MinetestBot [git] paramat -> minetest/minetest: Mgv7 floatlands: Add exponent parameter 2aad3be https://git.io/vbjJV (2018-01-03T04:05:55Z)
04:11 MinetestBot [git] Thomas--S -> minetest/minetest: Connected Nodeboxes: Add `disconnected` boxes f3b9d87 https://git.io/vbjJw (2018-01-03T04:05:21Z)
04:11 MinetestBot [git] HybridDog -> minetest/minetest: Tool.cpp/.h, lua_api/l_util.cpp: Tidy up code and remove dead code 345e104 https://git.io/vbjJr (2018-01-03T04:01:15Z)
04:11 MinetestBot [git] HybridDog -> minetest/minetest: Tool getDigParams: Fix selecting the best fitting time d7c1f6c https://git.io/vbjJo (2018-01-03T03:58:11Z)
04:17 Hawk777 joined #minetest
04:43 illwieckz joined #minetest
05:00 est31 joined #minetest
05:20 Billre left #minetest
05:28 est31 joined #minetest
05:35 Fritigern joined #minetest
05:57 Cornelia joined #minetest
05:58 StaceySerafin joined #minetest
05:58 StaceySerafin was kicked by MinetestBot: StaceySerafin
05:59 StaceySerafin joined #minetest
05:59 StaceySerafin was kicked by MinetestBot: StaceySerafin
06:02 sofarstacey joined #minetest
06:02 sofarstacey was kicked by MinetestBot: sofarstacey
06:06 Emerald2 o_O
06:15 SonosFuer joined #minetest
06:18 Akberid joined #minetest
06:41 swift110-phone Hey Emerald2
06:44 Emerald2 What's up swift?
06:47 Emerald2 Hey swift110-phone
06:47 red-F00 have to say the new bot update is nice
06:47 red-F00 way better then freenode anti spambot stuff
06:48 swift110-phone How are you Emerald2
06:49 swift110-phone What new bot update red-F00
06:49 Emerald2 I'm good. You pinged me. :P
06:49 red-F00 it kicks spambots/humans
06:49 Emerald2 Oh you mean the MinetestBot, red?
06:49 red-F00 spam(bots/humans)
06:49 swift110-phone Good Emerald2 and I know
06:49 red-F00 Emerald2, yes
06:50 swift110-phone red-F00: is that just for here
06:50 red-F00 yes
06:50 swift110-phone Ok
06:50 red-F00 -dev blocks users that aren't logged in
06:58 VargaD joined #minetest
07:08 GreenDimond joined #minetest
07:31 CWz joined #minetest
08:02 htur joined #minetest
08:12 swift110-phone Oh
08:16 VargaD joined #minetest
08:23 Emerald2 So what's up swift? Just saying hi to people? Looking for me for something? Looking for the other Emerald? lol. :)
08:55 swift110-phone I do like to say hello to folks it's a bit of a habit
08:55 swift110-phone I'm just relaxing at the moment
09:06 est31 joined #minetest
09:17 lanlink_ joined #minetest
11:00 Krock joined #minetest
11:01 lumberJ joined #minetest
11:26 Telesight joined #minetest
11:28 aristotle_IRC joined #minetest
11:28 lumberJ hey Krock, the server is backing up at the moment
11:28 lumberJ and in general just taking a break from the annoyances of being on the public list
11:28 lumberJ but it is alive and well. should be back up in 30-40 minutes
11:29 Krock hi. was asking as I've looked a few times on the server list and it wasn't there
11:29 lumberJ yeah, i just took it off for a little break. too many annoying little kids with the holiday break :P
11:37 pgimeno red-F00: re your question in #6863, no, Minetest doesn't include LuaJIT headers; when LuaJIT is requested, it's assumed to be installed by the user
11:41 swift110-phone Sup Krock long time no see
11:45 Krock hi swift110-phone
11:45 swift110-phone How are you krock
11:46 Krock just fine. how about you, swift110-phone ?
11:49 swift110-phone Im good Krock what are you up to
11:49 Krock just browsing
11:53 swift110-phone Oh cool
12:07 YuGiOhJCJ joined #minetest
12:12 YuGiOhJCJ joined #minetest
12:16 Fixer joined #minetest
12:19 YuGiOhJCJ joined #minetest
12:21 YuGiOhJCJ joined #minetest
13:09 eddie joined #minetest
13:10 Guest5468 is it possible to create a map through cli
13:10 Guest5468 im trying to setup a minetest server on a headless 'server'
13:12 rafalcpp joined #minetest
13:17 rafalcpp joined #minetest
13:27 VargaD joined #minetest
13:28 htur joined #minetest
13:37 Gael-de-Sailly joined #minetest
13:40 RobbieF joined #minetest
13:41 RobbieF left #minetest
13:42 Tux[Qyou] joined #minetest
13:45 lanlink joined #minetest
13:47 lanlink joined #minetest
13:52 Beginer joined #minetest
13:54 Gael-de-Sailly joined #minetest
13:56 Beginer Hi, how do i install already generated levelDB map? I have no problems with sqlite3, but levelDB map is corrupted by mapgen (map is re-generated).
13:56 rubenwardy you need to change the backend type
13:56 rubenwardy in world.mt
13:56 rubenwardy before opening the world
13:56 Beginer yes, but still not working
13:58 lanlink joined #minetest
14:00 lanlink joined #minetest
14:08 EvergreenTree joined #minetest
14:10 MaybeDragon joined #minetest
14:10 Darcidride joined #minetest
14:16 DI3HARD139 joined #minetest
14:29 niko joined #minetest
14:29 lRC joined #minetest
14:37 niko joined #minetest
14:39 MinetestBot [git] rubenwardy -> minetest/minetest: Fix crash on can_bypass_userlimit returning non-boolean 2af47e6 https://git.io/vbjDK (2018-01-03T14:18:13Z)
14:41 benrob0329 left #minetest
14:43 red-F00 pgimeno,
14:43 red-F00 https://github.com/minetest/minetest/blob​/master/src/script/cpp_api/s_base.cpp#L38
15:02 Copenhagen_Bram joined #minetest
15:09 illwieckz joined #minetest
15:34 Beginer looks like map is corrupted: Corruption: corrupted compressed block contents :(
15:36 red-F00 thats some bad luck
15:39 pgimeno red-F00: yeah, that's detected by https://github.com/minetest/minetest/blo​b/master/cmake/Modules/FindLuaJIT.cmake as an external dependency
15:40 pgimeno maybe you meant "include" as in "#include" and not as in "package"
15:41 red-F00 how about we refer to #include as "included", and having it in our code tree as "bundled" from now on to avoid confusion?
15:41 pgimeno yes, sorry
15:42 red-F00 so yes we don't bundle it but I don't see why you can't include the header as long as everything is inside the right preprocessor cases
15:43 pgimeno I meant it's not bundled. Anyway, adding cdata support requires LuaJIT internal includes that are, to my knowledge, not bundled with normal LuaJIT headers.
15:44 pgimeno That's why caused my confusion, because unless LuaJIT is bundled, those headers are not available.
15:46 pgimeno see e.g. https://gist.github.com/rtsisyk/6110709
15:49 pgimeno or https://github.com/AltraMayor/gatekeep​er/blob/master/lib/luajit-ffi-cdata.c
15:50 pgimeno (the latter is based on the above, but the includes point to ../dependencies/luajit-2.0/src/lj_* which should make it clearer that they are internal LJ headers)
16:13 t20kdc joined #minetest
16:18 Beginer left #minetest
16:25 Copenhagen_Bram joined #minetest
16:31 quater joined #minetest
16:34 Akberid joined #minetest
16:41 riff-IRC joined #minetest
16:42 Jousway joined #minetest
16:43 quater hello, can I use GMP in lua mod ?
16:44 Krock only if there are extern C functions you can call with FFI
16:44 AndDT joined #minetest
16:45 quater minetest need GMP to compile, is it available for mod ?
16:47 pgimeno not directly; if you've compiled your minetest with LuaJIT enabled, then what Krock said
16:49 quater without LuaJIT, can I access it ?
16:50 Krock minetest.check_password_entry and minetest.get_password_hash  can be accessed from the Lua API. Always.
16:50 Krock These use some GMP functions, if you meant that
16:50 Krock if you want anything else, it can't be done without FFI
16:50 Krock or an engine change, of course
16:53 quater something like: local ffi = require("ffi")
16:54 Krock something like that. if there was no mod security
16:54 Krock with mod security, you'll either have to disable that or add the mod to your trusted_mods setting
16:55 Krock then request the insecure env like this: local ie = minetest.request_insecure_environment() local ffi = ie.require("fii")
16:57 quater ok
16:58 Krock however, after scrolling through the bundled mini-gmp.c I think there's no function you could call from Lua, even with FFI, as they're not marked as extern ones
16:59 pgimeno are they marked as static? if not, I don't see why they can't be called with FFI
17:01 pgimeno using C functions through FFI is not that easy though; you need to basically replicate the C header in the Lua side, sometimes with a few changes
17:03 htur joined #minetest
17:03 quater I would like to use/write some cryptographic functions in a mod
17:03 htur brb
17:06 htur joined #minetest
17:07 pgimeno public key?
17:08 pgimeno quater: I mean, public key cryptographic functions?
17:09 red-F00 what for?
17:09 red-F00 pgimeno, can you call c++ functions if you handle the name mangling yourself?
17:11 pgimeno ah sorry, not marked extern "C"
17:11 pgimeno red-F00: yes, if you know the compiler and platform used
17:12 red-F00 but I assume they also have to be extern right?
17:12 red-F00 unless this does some next level magic
17:13 pgimeno at this point I'm not sure, I assumed that the executable had the symbols not marked as static (unless stripped)
17:13 whitephoenix joined #minetest
17:13 red-F00 pgimeno, can ffi set variable values>
17:13 red-F00 *?
17:14 pgimeno red-F00: not sure, I haven't tried to do that
17:15 pgimeno if you can get a pointer to a variable, certainly
17:17 pgimeno e.g. int var; int *var_ptr() { return &var; } then in the LuaJIT side: ffi.cdef 'int *var_ptr();' local var_ptr = ffi.C.var_ptr(); var_ptr[0] = 5
17:18 pgimeno that sets var to 5
17:18 red-F00 I mean obviously it can if you know the address of the variable I was wondering if it also could get the address of extern variables by name
17:20 pgimeno it seems you can, I haven't ever tried it though
17:20 Krock I did some insecure stuff: https://pastebin.com/WXAydxT3 creating an uninitialized pointer array, which then gets modified by C++ to the MapNode data in the Vmanip
17:20 pgimeno "ffi.cdef(def) Adds multiple C declarations for types or external symbols (named variables or functions)."
17:20 pgimeno note "variables"
17:22 pgimeno Krock: cool
17:23 quater yes, to secure the exchanges between players
17:24 rubenwardy it'd probably be easier to find a lua library for this
17:24 red-F00 oh are you the guy that wrote some modifed xor-cryptr for csm>
17:24 red-F00 *?
17:25 sofar rot13
17:25 sofar kek ;-)
17:25 Krock XOR with 0xFF
17:25 quater no, but I am writing something more generale
17:25 Krock * 0x7F
17:25 red-F00 pmr 24 gq uyw zcrrcp
17:26 pgimeno Krock: I'm not sure the double indirection is necessary. Can't the function just return it?
17:27 Krock pgimeno, I wasn't sure whether LuaJIT/FFI would accept that as CDATA
17:27 Krock or do you mean as #2 argument in ffi.new() ?
17:28 pgimeno I mean in the case of MapNode **data, I think the function can return MapNode *VoxelManip_get(void **lvmp)
17:28 * red-F00 waits for mapgens to be leaky
17:29 Krock red-F00, that will result in amazing fast mapgens if done right
17:29 Krock using a struct (as in using the entire MapNode data) seems to slow down LuaJIT quite much
17:30 pgimeno maybe due to cache locality?
17:30 pgimeno I don't know
17:32 Krock array[index1][index2] is the same as array[index1].index2 (used in the script), which is an additional step to get the right value for each coordinate
17:35 pgimeno hm, I expected that to be compiled to very plain code
17:36 red-F00 quater, I don't see the point...
17:37 EvergreenTree joined #minetest
17:41 quater to be able to exchange a top secret between players or a mod
17:45 AndDT joined #minetest
17:46 Animetrom joined #minetest
17:46 * VanessaE sighs
17:46 * Krock looks at VanessaE
17:46 VanessaE just got a call from the neurologist
17:46 Animetrom left #minetest
17:46 Animetrom joined #minetest
17:47 Animetrom left #minetest
17:47 Krock "madam', sadly we must come and take you to a safe place where you can't harm anyone"
17:47 VanessaE that stroke I had?  lacunar infarct.  very common, caused by a physiological defect in small blood vessels.
17:48 rubenwardy interesting
17:48 rubenwardy what's the chance of it happening again?
17:48 rubenwardy very small I hope
17:49 VanessaE the infarct part, along with the MRI results, suggests I suffered some brain cell death :(
17:49 VanessaE not a huge chance of a relapse, as long as I keep taking my aspirin and statin med.
17:50 VanessaE so, now I'm brain damaged.  I am NOT happy.
17:51 VanessaE they said it damaged a ~15mm section of tissue
17:51 VanessaE they consider that especially small though
17:53 VanessaE one good bit of news though: they said, in no uncertain terms, that it was not caused by diet or lifestyle.
17:54 Darcidride_ joined #minetest
17:56 VanessaE at least I'll recover.  at this point, my left hand is 90% functional, it's just sloppy on the keyboard and a few very fine actions.
18:00 mitrom 06:06 <sofarstacey> Tell sofar that I logged the New Years Day chat and will send it to Stacey Serafin. And to his neighbors. And to Intel. This is moving to real life, folks. Mock that. He knows what I mean. I can't imagine what he was thinking; he's fucked.
18:00 mitrom ...
18:01 red-F00 mitrom, nothing to see
18:01 VanessaE ssdd, mitq
18:01 VanessaE ssdd, mitrom
18:01 mitrom ssdd?
18:02 VanessaE same shit, different day
18:02 VanessaE (or s/shit/story/)
18:03 red-F00 Keep Calm and set umode +R
18:03 VanessaE that, too
18:03 rubenwardy if you do that, does the sender get an error message?
18:03 red-F00 don't think so
18:03 red-F00 sadly
18:04 rubenwardy I'd either want the sender to get an error message, or for the messages to be logged somewhere for records
18:04 vanessatest joined #minetest
18:05 VanessaE yes
18:05 rubenwardy cool
18:05 VanessaE there is an error msg
18:05 vanessatest left #minetest
18:05 red-F00 oh cool
18:05 red-F00 they changed it
18:05 red-F00 or I forgot idk
18:09 red-F00 https://i.imgur.com/ELWDcLQ.png
18:09 VanessaE I imagine that about now, He Who Has A Child Porn Obsession is getting pissed off beyond all previous measure ;)
18:11 red-F00 The One That We Are Discussing seems to Also enjoy Random capital Letters in the middle of Sentences
18:11 VanessaE rubenwardy: regarding the stroke, I am so thankful that I could spot the symptoms before it was too late.
18:12 VanessaE red-F00:  maybe he thinks he's a writer of the caliber of those who composed the US Constitution  ;)
18:13 red-F00 maybe he is a German immigrant
18:13 VanessaE heh
18:13 Krock Not All GermAn WorDs sTarT wItH CaPiTaLs
18:13 BillyS YaY
18:14 red-F00 #NotAllGermans
18:15 rubenwardy that's good to hear, VanessaE
18:15 red-F00 yeah
18:17 pgimeno rubenwardy: sorry for neglecting to search before submitting
18:17 rubenwardy np
18:17 rubenwardy github search is a pain anyway
18:18 sofar the new sourceforge
18:19 pgimeno I have the Perlin noise LuaJIT functions almost ready (pending testing). Should I submit them to the same PR or open a new one?
18:19 red-F00 same one
18:19 pgimeno thanks, I wondered because there was already one approval
18:20 red-F00 kinda funny how LuaJIT claims it means you don't have to write bindings to use FFI
18:20 red-F00 it just means it has to be reapproved by that dev
18:20 Krock how's that going to work? perlin noise is ~1 ... ~(-1). FFI only makes it faster when dealing with integers
18:21 red-F00 is no-one able to review #6087 ?
18:21 Krock https://github.com/minetest/minetest/pull/6087
18:22 red-F00 sorry I'm just kinda annoyed since it was ~half a year since I finished that and all I did in the meantime was rebase it like once
18:22 pgimeno Krock: it's not so much for accelerating it, as it is for being able to do the allocation in memory not restricted by LuaJIT's limits
18:23 Krock if so, then it's an ugly fix. can't we modify a Lua variable pointer (if existent) to our fresh generated array/data?
18:24 red-F00 and moved a list from one part of the code to another a few months ago
18:25 pgimeno Krock: it also uses memcpy or even directly the array, hence a bit faster to access the data (not that that should make much of a difference). I'm not aware of Lua variable pointers, besides LuaJIT cdata.
18:26 Shara left #minetest
18:26 Shara joined #minetest
18:28 pgimeno also, the array is in floats, not doubles
18:39 nepugia hey, asked this some days ago and got no response.. is there some way to tell minetest to crop textures instead of resizing them? i made some bigger textures intended to be placed in formspecs as a progress indicator, it looks rather ugly when the texture gets resized, the correct way would be to crop them imo, i dont see any reason why it schould resize them, ever
18:41 rubenwardy I believe you may be able to do that with texture modifiers
18:41 Krock like the furnace progress arrow?
18:41 rubenwardy it makes sense to resize them as the element is to show an image at that position and size
18:41 rubenwardy although stretching is ew
18:42 nepugia well, i have no other way currently to show textures in a formspec :3
18:43 rubenwardy say you have an 64x16 image foo.png
18:43 rubenwardy [combine:8x16:0,0=foo.png
18:43 rubenwardy should be to crop, I'm not sure
18:43 rubenwardy formspecs such generally
18:43 rubenwardy *suck
18:43 nepugia @krock a bar that increases in size from left to right (i attempted to code right to left aswell, but the width and height parameters or picture[ seem to operate with another scale compared to the x and y position)
18:44 nepugia do i need to calculate the pixels for the image then? formspecs operate on inventory sizes though afaik, how would i get the correct conversion rate
18:45 rubenwardy 1 slot is 64 pixels on a standard screen
18:45 rubenwardy a standard screen is 96 DPI at 800x600 I believe
18:45 nepugia always?
18:45 Krock nepugia, https://github.com/minetest/minetest_game/b​lob/master/mods/default/furnace.lua#L15-L16
18:45 rubenwardy if the DPI increases, the pixel increases
18:45 Krock use lowpart and then rotate
18:45 rubenwardy ah there we are
18:45 rubenwardy Ah!
18:45 rubenwardy I saw lowpart but didn't think to rotate it
18:46 Krock use it after rotating again if you want to chop it from two sides
18:48 nepugia where is this documented? i dont really understand what lowerpart is supposed to do
18:48 rubenwardy https://rubenwardy.com/minetest_moddin​g_book/lua_api.html#lowpartpercentfile
18:49 nepugia http://dev.minetest.net/formspec#image
18:49 nepugia doesnt seem to give any info on that
18:49 rubenwardy or  https://github.com/minetest/minetes​t/blob/master/doc/lua_api.txt#L478
18:49 rubenwardy you want texture modifiers
18:49 rubenwardy they're not specific to formspecs
18:49 rubenwardy Basically, lowpart is lower part
18:49 nepugia hmm, my browser sais the certificate expired 13 minutes ago :3
18:49 rubenwardy and it copies the lower % of an image to the texture
18:50 rubenwardy ehat
18:51 nepugia then i would have to calculate the percentage, for which i would need to know the inventory<->pixel thingy
18:51 rubenwardy nepugia, looks like my auto renew failed
18:51 rubenwardy doing it now
18:51 rubenwardy I thought it's a progress bar?
18:52 rubenwardy nepugia, done
18:52 nepugia yes
18:52 rubenwardy surely you'd know the progress if it's a progress bar?
18:54 nepugia my texture is 1080pixels wide, my progress is 30%, now how would i calculate the percentage to cut the image if the area i need to fill is much smaller than 1080px
18:54 nepugia i already know how to tell minetest where to put it, if i now cut it in the wrong place it will be stretched or squezzed
18:54 nepugia and both will look shitty
18:54 rubenwardy use 30% and size the image element so it's 30% of the max (in units
18:54 nepugia (still, beeing able to tell it to just use the image 1:1 and crop off the rest it doesnt need would be a fairly elegant solution)
18:55 nepugia again, i have no idea how many units me image is
18:55 rubenwardy how big is the window?
18:55 nepugia they dont appear to have a fixed conversion rate
18:55 rubenwardy images resize their content
18:55 nepugia how would i know?
18:55 rubenwardy size[w,h]
18:56 rubenwardy sorry, window as in formspec dialog
18:57 nepugia having to querry information about dpi window size to calculate a conversion rate just to be able to put an image into a formspec seems rather overkill to me
18:57 rubenwardy you don't
18:57 sfan5 news flash: formspecs suck ass and should be replaced with a better system at the next possible point
18:57 rubenwardy that too
18:57 nepugia i dont know how big the formspec window is either, i made a utility function to append the progress bars
18:57 rubenwardy but you just need to decide how big the progress bar should be in formspec sizes
18:58 rubenwardy so if the window is size[8;6] you may decide to make the progress bar 7,1
18:58 rubenwardy in which case the image should be of a ratio 7:1
18:58 rubenwardy then when making the formspec, make the progress bar size  7*progress,1
18:58 rubenwardy and crop using lowpart
18:59 nepugia you dont understand me, the image has to be cropped
18:59 nepugia it will look shtity otherwise
18:59 rubenwardy I don't understand you
18:59 rubenwardy what is the image?
18:59 rubenwardy the way I say won't have any stretching, and will crop to the right
18:59 nepugia a texture i made
19:00 nac joined #minetest
19:00 nepugia >news flash: formspecs suck ass and should be replaced with a better system at the next possible point
19:00 nepugia yep, so ill retire modding mintest and come back in three years?
19:00 sfan5 no that's not an option ;)
19:01 nepugia @rubenwardy if i specify a percentage that is too big it will still crop the rest of the image, otherwise there is really no point in specifying a percentage
19:01 rubenwardy well, cap at 100$
19:01 rubenwardy *100%
19:01 rubenwardy then the entire image will be shown in the entire image element
19:01 rubenwardy at 50%, the image element will be half as wide
19:01 rubenwardy but the image will also be cropped to be half as wide
19:01 nepugia i dont want that, i want it to fill the area 1:1
19:02 nepugia and crop the rest
19:02 rubenwardy no possible
19:02 Krock do each step in GIMP and switch the texture each time 8)
19:03 nepugia possible surely, with way too much coding involved (at the lua side anyway).... unless someone made an image[ alternative that cropped
19:03 rubenwardy doesn't help him though, Krock
19:03 Krock s/texture/texture name/
19:03 nepugia @krock well, that would be the same approcah as animating nodes with node_swap, i dont want to spend all my time writing around stuff that the game really schould implemennt itself :/
19:04 VanessaE rubenwardy: no apparent cell death in that stroke.  whew, thank G*d
19:06 pgimeno yay for acting fast :)
19:11 whitephoenix joined #minetest
19:31 YuGiOhJCJ joined #minetest
19:38 YuGiOhJCJ joined #minetest
19:46 nepugia is there any way to get a callback for when a player wants to acces a blocks formspec?
19:46 nepugia (so i can modify the formspec beforehand)
19:47 sfan5 if that formspec is contained in the metadata, no
19:47 nepugia .-.
19:48 rubenwardy you can use on_rightclick to show the formspec instead
19:48 rubenwardy but there will be latency
19:48 rubenwardy but for your usecase that's fine
19:48 nepugia could i then show the formspec to the players upon them rightclicking the block?
19:48 rubenwardy yes
19:48 soloojos joined #minetest
19:48 nepugia there will be latency... noticeable?
19:49 nepugia i can still use all the formspec functions then?
19:50 rubenwardy depends
19:50 rubenwardy in most cases no
19:50 rubenwardy but it does depend on the player's internet
19:50 nepugia ah, so bugs that definetely wont show up in singleplayer testing, nice
19:50 rubenwardy also, when you use Facebook to monitor certificate registrations on your site: https://i.rubenwardy.com/62ZoeHaP6m.png
19:51 nepugia well, the other option i have is to continiously update the formspec in the background, which seems like a huge waste of ressources
19:53 nepugia what functions would i loose if i use it in on_rightclick then?
19:56 rubenwardy nepugia, nothing
19:57 rubenwardy the only difference between formspecs in meta and show_formspec is the time difference (meta formspecs are shown instantly as the client has them locally) and which callbacks are ran (meta formspecs use a node definition, normal formspecs use a global callback)
20:00 nepugia minetest uses a lua compiler though? if i have a specific static formspec and either store it in node meta or show it as formspec on_rightclick i wouldnt really expect any difference in time
20:00 rubenwardy in singleplayer, no
20:01 rubenwardy only in multiplayer / on a remote server
20:01 rubenwardy as the server needs to send a message with the formspec's content
20:05 nepugia hmm, is there a way to see if a player still has the formspec open (as to update the content while the player still has it open)
20:06 sy joined #minetest
20:07 rubenwardy there's events when they close it
20:07 rubenwardy but that's subject to network latency
20:07 rubenwardy ie: you may sent an update just after they closed it, but before you receive a notification that it was closed
20:07 nepugia i think i can use register fields and on_rightclick for this
20:07 nepugia though, it could lead to some cases where it would reopen it if it was already closed :/
20:07 rubenwardy this is where client-side scripting will be useful, when ready
20:08 * red-F00 complains about security PR's taking a while to review
20:08 * red-F00 hides
20:08 nepugia an option to show the formspec to a player only if they already have it open would mitigate this i suppose
20:09 nepugia i might generate one formspec too much and have it discarded though that would be acceptable :3
20:09 t20kdc red-F00: in core or mods, though...
20:15 soloojos joined #minetest
20:28 riff-IRC joined #minetest
20:31 Billre joined #minetest
20:36 swift110 joined #minetest
20:38 soloojos joined #minetest
21:21 swift110 joined #minetest
21:21 soloojos joined #minetest
21:23 Darcidride_ joined #minetest
21:23 quater can I copy the lua environment ?
21:25 red-F00 quater, what do you mean?
21:25 red-F00 if you want a reference to the globals use _G
21:26 quater I would like a copy
21:26 red-F00 you can't just copy tables in lua
21:26 red-F00 quater, how about you tell me what you want to do, and not a solution
21:28 red-F00 you can just do a deep table copy, but there is a good chance thats not what you want to do
21:29 quater setfenv('t = 2', env); with env equals the current environment
21:30 quater and in env, t = 2
21:30 red-F00 then us _G
21:30 red-F00 quater, you trying to sandbox something?
21:30 swift110 joined #minetest
21:31 quater no
21:31 red-F00 oh is this os.setenv?
21:32 quater I can create a environment like env = { _G = _G, pairs = pairs, ... } but I would like to avoid the pairs = pairs and so on
21:33 red-F00 why can't you just use the current enviroment?
21:34 red-F00 quater, just loop over _G and copy stuff
21:35 quater I would like to clean the environment if necessary for a new restart like: setfenv('t = 2', env); env = new_env() and t = nil again
21:35 red-F00 https://gist.github.com/red-001/​493937b94dcdcea867b309872325fe61
21:35 Tux[Qyou] joined #minetest
21:36 quater ok
21:36 red-F00 oh you want to store a full backup?
21:36 quater yes
21:36 Pie-jacker875 joined #minetest
21:37 red-F00 well you can use table.copy
21:37 red-F00 but really there shouldn't be any reason to do this
21:38 red-F00 env_bck = table.copy(_G)
21:40 red-F00 oh and make sure to change what _G points to in that table
21:40 red-F00 env_bck._G = env_bck
21:40 EvergreenTree joined #minetest
21:40 red-F00 wait
21:42 red-F00 I'm not sure if table.copy handles that
21:42 rubenwardy table.copy looks for recursion
21:42 rubenwardy and will reuse the copied value
21:42 red-F00 ah
21:42 red-F00 so no need to do that then
21:43 quater the gist seems to work fine, I will do more tests
21:43 red-F00 quater, why are you restoring the globals from a backup?
21:44 red-F00 why not just stop your code from messing them up in the first place?
21:44 red-F00 huh
21:44 Fusl joined #minetest
21:44 Fusl joined #minetest
21:45 red-F00 Fusl, seems you got accidently k-lined
21:45 quater the code must be as close as possible to the mod environment, which will allow the user to write a mod in a mod
21:46 Fusl red-F00: oh noes
21:46 red-F00 so you do want to sandbox then, lol
21:46 Fusl just kidding, i /quot
21:46 Fusl s/quot/quit/
21:46 red-F00 oh yeah
21:46 red-F00 silly me
21:46 t20kdc I'll admit it fooled me
21:47 red-F00 forgot about the quit: prefix
21:47 Fusl happens
21:47 red-F00 quater, http://lua-users.org/wiki/SandBoxes
21:49 pgimeno this is fun, btw: 1.0.0.127.in-addr.arpa.li
21:50 rubenwardy nothing there
21:50 rubenwardy \o/
21:50 rubenwardy oh wait, it's Fusl's IP
21:50 pgimeno as in: * Fusl (fusl@1.0.0.127.in-addr.arpa.li) has joined
21:51 Fusl yes
21:51 Fusl i'm your local horst
21:51 sfan5 in-addr.ar.pa when
21:51 JackFrost Needs more SASL, Fusl.
21:52 red-F00 quater, try this https://gist.github.com/red-001/​292a2b52ee18b4a92e02599b4b245674
21:52 Fusl sfan5: meo.ws already costs 1k€, i don't want to imagine what ar.pa costs
21:52 sfan5 10k€ :D?
21:52 Fusl more
21:53 Fusl i'd say around 50k
21:53 red-F00 you might have to redirect some minetest tables
21:53 red-F00 huh you had a decent idea to start off with
21:54 red-F00 if this is not meant just for running code, but for testing a mod
21:54 Fusl a standard .pa dom already costs ~500€
21:54 red-F00 there really isn't a good way to test a mod at runtime
22:06 soloojos joined #minetest
22:07 quater joined #minetest
22:10 YuGiOhJCJ joined #minetest
22:18 swift110 joined #minetest
22:31 Fusl joined #minetest
22:41 swift110 joined #minetest
22:46 Akberid joined #minetest
22:54 Akberid joined #minetest
22:55 swift110 joined #minetest
23:09 quater is it possible to read a text file ?
23:10 Calinou in Lua? of course, yes
23:11 red-F00 trivial even
23:11 quater I can't use io
23:12 rubenwardy why not?
23:12 quater attempt to index global 'io' (a nil value)
23:12 rubenwardy is this on the server API?
23:13 rubenwardy is this in the normal environment, or a custom one?
23:13 quater local env = {}
23:13 quater for k, v in pairs(_G) do
23:13 quater env[k] = v
23:13 quater end
23:13 quater env._G = env
23:13 sofar mod_security?
23:14 sofar or reading from a file outside the world folder?
23:14 sofar oh, io is unknown
23:14 red-F00 wouldn't be a can't index error
23:16 quater I am trying:
23:16 quater local file = io.open(core.get_modpath('csm_lua') .. '/TODO')
23:16 quater local text = io.read(file);
23:16 quater core.display_chat_message(text)
23:17 quater it is a CSM
23:17 red-F00 no io for CSM
23:17 rubenwardy you'll need to use mod_storage instead
23:18 rubenwardy to store data on the client
23:18 quater_ joined #minetest
23:18 red-F00 see logs
23:19 quater_ attempt to index global 'io' (a nil value)
23:19 swift110 joined #minetest
23:19 red-F00 <quater> it is a CSM
23:19 red-F00 <red-F00> no io for CSM
23:19 red-F00 <rubenwardy> you'll need to use mod_storage instead
23:19 red-F00 <rubenwardy> to store data on the client
23:19 red-F00 * quater_ (~lRC@120-164-41-78.ripe.phibee-telecom.net) has joined
23:19 red-F00 was kicked by ShadowBot: Paste flood detected. Use a pastebin like pastebin.ubuntu.com or gist.github.com.
23:19 red-F00 joined #minetest
23:19 sofar lol
23:19 red-F00 it's 4 lines shadowbot
23:19 lumberJ :D
23:20 sofar where does csm mod_storage live???
23:20 sofar clientmods/mod_storage ?
23:20 rubenwardy no idea
23:20 sofar kek
23:20 red-F00 client/??/modname.??
23:21 red-F00 client/mod_storage/modname
23:21 sofar wait, csm mod_storage is persistent?
23:21 red-F00 yes
23:21 sofar awesome
23:21 soloojos joined #minetest
23:21 Gael-de-Sailly joined #minetest
23:21 sofar can't wait to... uhh, uuuuuuuse that
23:21 sofar ab
23:21 red-F00 well harddrive space is cheap
23:22 red-F00 so eh don't use minetest on an ssd
23:23 red-F00 in all seriousness we should just let the user set a storeage size limit
23:23 sofar csm storage should be per server
23:23 sofar or transient
23:23 rubenwardy no, timeout
23:23 red-F00 you would know if you got kicked
23:24 red-F00 sofar, not worth implementing that till we get around to having the server send the mods but yes it should be
23:25 red-F00 also we should use a database for all modstorage
23:26 red-F00 if we do make it per-server it would be nice to also give server guid's
23:26 red-F00 servers*
23:29 Fusl joined #minetest
23:29 Fusl joined #minetest
23:29 red-F00 Fusl, dammit stop trying to fool us all
23:29 Fusl sorry
23:30 red-F00 jk
23:30 Fusl had to apt install weechat-devel-perl :D
23:59 Cornelia joined #minetest

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