Minetest logo

IRC log for #minetest-dev, 2014-11-08

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

All times shown according to UTC.

Time Nick Message
00:35 ShadowBot joined #minetest-dev
00:42 paramat joined #minetest-dev
00:57 kaeza joined #minetest-dev
01:18 alexxss joined #minetest-dev
01:18 AnotherBrick joined #minetest-dev
01:21 groknsberger joined #minetest-dev
01:21 CraigyDavi` joined #minetest-dev
01:22 paramat joined #minetest-dev
01:29 mos_basik_ joined #minetest-dev
01:31 Dragonop joined #minetest-dev
01:33 OldCoder joined #minetest-dev
01:33 OldCoder joined #minetest-dev
02:23 MikeFair joined #minetest-dev
03:21 zat joined #minetest-dev
04:44 paramat left #minetest-dev
04:47 hmmmm i don't quite understand what #1791 does
04:47 ShadowBot https://github.com/minetest/minetest/issues/1791 -- Add factor of maximal player transfer distance by SmallJoker
04:47 hmmmm i guess it's to accelerate pickup-items-around-player lua code?
04:48 hmmmm agh christ... #1806 and #1567 want me to revisit the idea of using GTK+ for our widget toolkit
04:48 ShadowBot https://github.com/minetest/minetest/issues/1806 -- Double-clicking the scroll bar of the server list joins a server
04:48 ShadowBot https://github.com/minetest/minetest/issues/1567 -- Configure mods menu: scrolling by clicking arrows or bar enables/disables mods
04:49 hmmmm or that other OGL widget toolkit that didn't suck celeron was talking about
04:50 hmmmm http://sprunge.us/NLQL?diff LGTM
04:51 hmmmm what's the status of #1804?
04:51 ShadowBot https://github.com/minetest/minetest/issues/1804 -- Fix regression and make minor improvements in the_game by Zeno-
04:51 VanessaE hmmmm: 1804 works, you'll want add_profiler_display_back also
04:52 hmmmm that's a branch name in zeno's repo?
04:52 VanessaE yeah
04:52 hmmmm ko
04:52 hmmmm gh
04:52 hmmmm @paramat:  you want to add schematics to minetest game by default?
04:53 hmmmm I wanted to as well, I had the schematic equivalent of the default tree sitting around for a while now.  we need better tools for creating schematics though, I'm sorry, creating them through the game itself is a shitty idea.  it's just incredibly cumbersome.
04:54 hmmmm does your schematic include y-slice probabilities?  you need that otherwise all your trees will turn out to be the same exact height
04:56 hmmmm papyrus and cacti should be done via decorations as well... i personally find the density of cacti in particular to be difficult to control using just the simple offset + scale * noise > threshhold
04:57 hmmmm nyancats need to be made using schematics however
04:57 hmmmm rather than simple decorations
05:02 RealBadAngel joined #minetest-dev
05:03 RealBadAngel hi hmmmm
05:03 hmmmm hello RBA
05:04 VanessaE good morning.
05:04 RealBadAngel morning
05:04 RealBadAngel hmmmm, have a little speedup patch, saw it already?
05:05 hmmmm no, what
05:05 hmmmm #?
05:05 RealBadAngel #1812
05:05 ShadowBot https://github.com/minetest/minetest/issues/1812 -- Speed up removing a node (less block mesh updates). by RealBadAngel
05:07 hmmmm why is it okay to add a node without updating edges
05:08 hmmmm oh, removenode
05:09 RealBadAngel because its a complete nonsense to update 13 blocks no matter the case
05:09 hmmmm yes but it's there because it was considered necessary at one point
05:09 RealBadAngel it is not necesary
05:10 RealBadAngel if lighting will require it, it will rise the number
05:10 RealBadAngel but usual case is to update 1-2 blocks
05:10 RealBadAngel corner case is 4 blocks
05:11 Miner_48er joined #minetest-dev
05:12 hmmmm well, i've been digging nodes with that patch and i haven't found any graphical glitches
05:13 RealBadAngel removing a node can cause a mesh update of neighbouring blocks in 2 cases:
05:13 RealBadAngel change of lighting, that will work as before
05:13 RealBadAngel number of modified blocks will be processed as usual
05:14 RealBadAngel b) need of change the visible/invisible face of an adjacent node
05:14 RealBadAngel that apply only to blocks on the edge
05:14 RealBadAngel BUT not on all block edges for christ sake
05:15 RealBadAngel also the order of mesh updates was wrong
05:16 RealBadAngel you need to make adjacent node visible before you remove the dug one
05:16 hmmmm does addnode need to update all the edges as well?
05:17 RealBadAngel if this work flawlessly for folks (proven already) then yes
05:17 RealBadAngel i mean the same way as remove
05:17 hmmmm so you're just changing the remove and then waiting for a while to see the reaction, then try changing addnode if nobody complains, i take it?
05:18 RealBadAngel yes, i put it to testing
05:18 hmmmm why not just change both of them at once
05:18 RealBadAngel only change is major speedup
05:19 RealBadAngel digging nodes causing game choking is known for ages already
05:20 RealBadAngel well, i can see no reason now why not adding the second case
05:20 hmmmm yea
05:21 RealBadAngel meanwhile feel free to dig some nodes ;)
05:21 hmmmm hrm..
05:21 hmmmm http://i.imgur.com/N7I5gob.jpg
05:22 hmmmm unrelated but I think that's a regression
05:22 VanessaE the pattern in the glass?
05:22 RealBadAngel what da fuk is that??
05:22 hmmmm it looks like mipmapping gone wrong
05:22 VanessaE it's always been that way.
05:22 hmmmm are you sure
05:22 VanessaE nearest-neighbor scaling.
05:22 VanessaE yup
05:23 hmmmm i literally never noticed that happening.  it doesn't seem right at all
05:23 hmmmm are you SURE...?
05:23 VanessaE I've seen artifacts like that since forever.  in that precise pattern?  I don't know for sure - hard to tell from that angle
05:24 hmmmm it doesn't really depend much on the angle
05:24 RealBadAngel VanessaE, you know how to reproduce that issue?
05:24 RealBadAngel hmmm ?
05:24 hmmmm it's the distance
05:24 hmmmm realbadangel:  simple, just make a plane of glass, and increase your distance from the glass texture.
05:24 VanessaE hmmmm: I've seen that effect since before I added the aniso/mipmap flags to the engine
05:25 hmmmm at a certain point the texture starts getting weird
05:25 hmmmm vanessae, that effect is definitely not normal.
05:25 hmmmm at a certain distance the glass becomes invisible
05:25 VanessaE it not "normal" but it's always been that way
05:26 hmmmm i'm just saying i never saw this..
05:26 VanessaE you never play :)
05:26 VanessaE turning on mipmap et al. helps on some GPUs
05:27 VanessaE but for others it just turns everything blurry (and it doesn't fix that particular problem completely at any rate)
05:27 VanessaE because it doesn't scale against the transparent bits.  lack of alpha in the scaling algorithm I guess.
05:28 hmmmm mip_map = true doesn't help i noticed
05:31 VanessaE wow, so many dead links on the screenshots thread.
05:31 RealBadAngel confirmed, its mip mapping
05:31 sol_invictus joined #minetest-dev
05:31 RealBadAngel enabling it causing that glitch
05:32 hmmmm maybe it has to do with an updated version of my drivers
05:32 VanessaE http://666kb.com/i/by0ptdgtdg7ghuzxe.jpg
05:32 VanessaE that one was posted on Oct. 22, 2011
05:32 VanessaE you can see the broken effect there, too
05:32 hmmmm you can still SEE the glass though
05:32 hmmmm it doesn't become invisible at a certain distance
05:33 VanessaE hmmmm: it has for me since..forever
05:33 hmmmm this is F'd up
05:33 VanessaE agreed
05:33 hmmmm great, another thing to look into
05:34 VanessaE this one, doesn't show the errant effect... http://666kb.com/i/byl5xtlcgo1pbzqzm.jpg
05:34 VanessaE weird.
05:34 VanessaE (posted Nov 11, 2011)
05:34 hmmmm that's the way glass looked for me
05:34 hmmmm used to anyway
05:36 RealBadAngel i just asked the uncle, it is precisely how the mipmapping works
05:36 RealBadAngel the feature is unusable for low resoultion textures
05:37 RealBadAngel it is simply downscaling the high res ones for better performance
05:37 RealBadAngel downscaling 16x16 to 1x1 will result always with such glitches
05:39 VanessaE if you use a bad algorithm, yeah
05:39 VanessaE aren't there others that irrlicht could use besides nearest-neighbor?
05:40 RealBadAngel idk
05:40 RealBadAngel but definitely i can see mipmapping require more memory
05:41 RealBadAngel http://en.wikipedia.org/wiki/Mipmap
05:41 VanessaE sure, if it's enabled at all
05:41 VanessaE but, hmmmm aren't you implying it's *disabled* in your case?
05:42 hmmmm mipmapping was disabled for me
05:42 hmmmm i don't know man.
05:42 RealBadAngel i just checked, glitch is not here when disabled, showed up when enabled
05:43 hmmmm what graphics card do you have
05:43 RealBadAngel wait a sec
05:43 RealBadAngel i mean hold on
05:44 RealBadAngel theres something fucked up
05:45 RealBadAngel sometimes even with option in menu unchecked mipmapping is still on
05:46 RealBadAngel i turned it off, mm was still on. i enabled shaders it went off, disabled shaders it was still off
05:47 VanessaE turning on aniso on my GPU is enough to make the glass fade out
05:47 VanessaE mipmap alone isn't enough to cause it.
05:48 VanessaE http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/Screenshot%20-%2011082014%20-%2012%3a48%3a33%20AM.png
05:48 RealBadAngel weirdo
05:49 VanessaE turn on trilinear along with those two and the glass fades out gradually rather than cutting off as you see in the image
05:49 VanessaE but hmmmm's point stands.
05:49 VanessaE mipmap shouldn't do that at all really.
05:49 RealBadAngel will try to find a reason
05:49 RealBadAngel hmmm, btw i can see a cave on VE
05:49 RealBadAngel ... screenshot
05:49 VanessaE you mean there behind the glass?
05:50 RealBadAngel ofc
05:50 VanessaE what about it?
05:50 * VanessaE flies over to look at it
05:50 RealBadAngel one day i will be pissed enough and i will force disabling that shit generation in the sky
05:50 VanessaE well if you disable THIS part, you'll disable the ability for caves to carve holes in the ground surface
05:51 RealBadAngel no
05:51 RealBadAngel i mean the fucking sky
05:51 RealBadAngel carving has the radius for that
05:53 RealBadAngel mg6 happily creates caves in content ignore, pisses on conditions never met (ie air, stone etc) and has no fucking limits
05:53 hmmmm it needs to create caves in content ignore.
05:53 hmmmm shit wouldn't work right if it didn't.
05:53 hmmmm in fact i think you tried to disable that a while back... i blocked it
05:54 RealBadAngel the last attempt worked quite nice
05:54 RealBadAngel it was the second one already ;)
05:54 hmmmm attempt?
05:54 hmmmm no, stop.
05:54 hmmmm you're walling in caves
05:54 RealBadAngel the second one was working as expected
05:55 RealBadAngel just try it
05:55 hmmmm look, RBA, I completely 100% understand the ramifications behind making that modification to the cave generator
05:56 hmmmm it doesn't "just work"
05:56 RealBadAngel mgv7 works
05:56 hmmmm the concept of not writing to content ignore does not work
05:56 RealBadAngel why v6 cannot work the very same way?
05:56 RealBadAngel and that was my last attempt
05:57 hmmmm mapgen v7 looks like it works because i have code that stops it from caving too far above the surface
05:57 RealBadAngel isnt that the point?
05:57 hmmmm look, the problem the cavegen is suffering from is the same as every other problem ever
05:57 hmmmm you need to predict the future of what the other neighboring node is going to contain
05:58 hmmmm this happens with lighting too and lots of other stuff'
05:58 RealBadAngel ofc, i know
05:58 hmmmm the only real solution here is to make caves noise-based
05:58 RealBadAngel but there have to be limits for predictions
05:58 RealBadAngel otherwise we will have shadows bugs forever
05:59 hmmmm no, we're not going to have shadow bugs because we're going to stop using this shitty lighting model someday
05:59 RealBadAngel and code that will carve caves (read WASTE time)
05:59 hmmmm and we're not going to have this problem because we're going to generate caves the smart way
05:59 RealBadAngel lighting is one problem
06:00 hmmmm the fundamental problem is the same, as well as the solution
06:00 RealBadAngel but executing complicated code and carving thousands of caves for nothing
06:00 RealBadAngel is another one
06:00 hmmmm switch to a method that doesn't depend on the contents of other non-existing blocks so you don't need to predict the future
06:01 RealBadAngel real lighting will just take shadows out
06:01 RealBadAngel carved in the sky caves will remain there
06:01 RealBadAngel slowing down world generation as hell.
06:04 RealBadAngel so making it smarter is the solution
06:06 kaeza joined #minetest-dev
06:15 paramat joined #minetest-dev
06:15 * paramat reads logs
06:16 paramat hmmmm, > 'you want to add schematics to minetest game by default?' yes basic trees for mgv5/7/8...
06:21 RealBadAngel paramat, schematics for trees? what for?
06:21 paramat RBA > caves in air cant be fixed in mgv6 due to the need to generate caves into ungenerated neighbour mapchunks
06:22 paramat for biomes in mgv5/v7
06:22 RealBadAngel you do have lstystem trees
06:22 RealBadAngel they dont need any schematics
06:22 RealBadAngel well, they do have own ones ;)
06:23 paramat yep but they are slower
06:23 VanessaE borrow the beech trees from moretrees but "scale them down" a bit?
06:23 VanessaE if lsystems would be faster than schematics, that is
06:23 paramat and look worse to me
06:24 paramat anyway hmmmm im stuck with mgv5, details coming ...
06:24 RealBadAngel paramat, lsystem trees can be anything you code
06:25 paramat yeah
06:25 RealBadAngel lsystem is in fact programming language for plants
06:26 paramat golden ratio something to do with branching ...
06:27 hmmmm don't L-system plants need lua callbacks?
06:27 hmmmm the whole point of the decoration system is to be fast
06:27 RealBadAngel no they dont need it
06:27 hmmmm jeez if we wanted to build trees up in Lua, we have on_generated, don't we?
06:27 RealBadAngel you can use it directly from engine
06:27 RealBadAngel it is just exposed
06:28 RealBadAngel static const with settings, call with pos and voila
06:30 paramat hmmmm, this commit returns mgv5 to having a separate function for generating underground dirt/lava blobs, and makes blobs optional with 'blobs' spflag. this commit doesnt work currently, but works if i exchange the 'if (spflags & MGV5_BLOBS) { generateBlobs(); }' in makechunk with just a non-optional 'generateBlobs'
06:30 hmmmm meh
06:30 paramat https://github.com/paramat/minetest/commit/3cce1c2cd726b52ab98fd45c0c1c12ca1c74b34d
06:30 hmmmm shouldn't blobs be an ore type anyway
06:31 hmmmm anyway
06:31 RealBadAngel https://github.com/minetest/minetest/blob/master/src/treegen.cpp#L120
06:31 hmmmm my plan was to remove dirt/gravel blobs from mapgen v6 as well and reimplement it with ORE_CLAYLIKE
06:31 RealBadAngel this is for lua
06:32 RealBadAngel https://github.com/minetest/minetest/blob/master/src/treegen.cpp#L153
06:32 RealBadAngel this is engine
06:33 paramat well for the moment mgv5 can be faithful to the original? even though its a bit crazy using eased 3d noise for an ore shape
06:35 paramat your easing results in larger noise values, i had to retune the caves threshold to compensate, the tunnels had become too narrow
06:35 paramat also retuned the maximum height of lava blobs as they were appearing above water level and setting fire to biomes =)
06:45 RealBadAngel #1812
06:45 ShadowBot https://github.com/minetest/minetest/issues/1812 -- Speed up removing a node (less block mesh updates). by RealBadAngel
06:45 RealBadAngel updated
06:47 RealBadAngel i applied the same changes to placing a node
06:47 RealBadAngel but, theres also a small question
06:48 RealBadAngel update lighting code adds modified block to the vector of modified blocks, what for?
06:59 hmmmm sure
07:00 hmmmm paramat, you realize you can mix and match ease/no ease for 3d noise, right?
07:01 hmmmm in any case I have a LOT of stuff to do...
07:02 hmmmm handling peoples' pull requests... generation element architecture stuff, new noiseparams, biomes, ore_blob, deco_lsystem, biome revamp, etc...
07:02 hmmmm and all that on top of work work
07:03 exio4 joined #minetest-dev
07:05 paramat bye hmmmmm O/ and yes im choosing 4 eased noises out of the original's 5, 'wetness' doesn't deserve it
07:07 paramat left #minetest-dev
07:17 Krock joined #minetest-dev
07:31 sol_invictus joined #minetest-dev
07:35 Hunterz joined #minetest-dev
07:46 kilbith joined #minetest-dev
08:35 ImQ009 joined #minetest-dev
08:40 OldCoder The RESEND RELIABLE bug continues. Sapier's patch helps but is incomplete.
08:45 Calinou joined #minetest-dev
09:14 OldCoder I am testing a new RESEND patch that combines my patch, Sapier's, and Zeno's
09:18 casimir joined #minetest-dev
09:36 ShiningRice joined #minetest-dev
09:45 Krock joined #minetest-dev
10:04 ImQ009 joined #minetest-dev
10:22 PenguinDad joined #minetest-dev
10:23 rubenwardy joined #minetest-dev
10:44 RiZom-91 joined #minetest-dev
10:54 exio4 joined #minetest-dev
10:56 PilzAdam joined #minetest-dev
12:03 FR^2 joined #minetest-dev
12:21 RiZom-91 joined #minetest-dev
12:24 RiZom-91 Hi all. I'm trying to add a "flammable" property for entities, is content_sao.cpp the best place to add a function that check if there is some lava or fire in the collisionbox ?
12:24 RiZom-91 something like bool LuaEntitySAO::touchFireOrLava()
12:26 RiZom-91 this function would be called in the LuaEntitySAO::step() function to apply damage is entity is flammable
12:26 Amaz joined #minetest-dev
12:27 RiZom-91 (btw flammable is false by default so it sould not break anything)
12:27 Krock #1791 thumbs up and down please
12:27 ShadowBot https://github.com/minetest/minetest/issues/1791 -- Add factor of maximal player transfer distance by SmallJoker
12:30 Amaz +1
12:41 PenguinDad Krock: could you add the value the old limted player transfer distance used to minetest.conf.example please?
12:41 Krock PenguinDad, that's infinite, right?
12:41 PenguinDad Krock: I said limited
12:41 Krock please explain what I should add exacty
12:41 proller joined #minetest-dev
12:46 PenguinDad Krock: add something like "# Use <value> to make it behave like the old unlimited_player_transfer_distance = false"
12:47 Krock PenguinDad, that depended on the object send range - there's no static value to "port" this
12:48 PenguinDad that's bad :/
12:51 kahrl Krock: what's the meaning of the magic constant 160 in the code?
12:51 Krock kahrl, minimal 1 chunk/mapblock.
12:52 kahrl Krock: write it as BS * MAP_BLOCKSIZE
12:52 Krock or as < 0
12:52 kahrl if that works, that would be even better
12:53 Krock so should I make those two commits to one or should I keep those single steps?
12:53 kahrl whoever merges it can squash
12:53 kahrl but if you want you can squash too
12:54 Krock k
12:57 Krock updated
13:01 kahrl haven't tested it but from the code it seems good to me now
13:01 GhostDoge joined #minetest-dev
13:01 proller joined #minetest-dev
13:01 Krock I tested the two commits a week ago
13:02 Krock the only fail was when the transfer distance is less than the player size
13:02 Krock but that's no longer possible
13:02 Krock s/is/was/
13:07 kahrl Krock: minetest.conf.example should state that the setting is measured in blocks
13:07 kahrl or perhaps the setting should be called player_transfer_distance_blocks
13:08 Krock hmm
13:09 Krock updated
13:16 kahrl would it be a good idea to do  if (player_radius == 0 && g_settings.exists("unlimited_player_transfer_distance") && !g_settings.getBool("unlimited_player_transfer_distance))  player_radius = radius;
13:16 kahrl as a compatibility measure?
13:16 kahrl s/\./->/
13:17 kahrl the downside would be a possibility of confusion: "I set player_transfer_distance to 0, why doesn't it work?" when there's still an unlimited_player_transfer_distance = false around
13:18 kahrl (I mean adding that code at server.cpp:714)
13:20 kahrl Krock: ^
13:22 Krock kahrl, yes I understand that. I could try to add compatbility
13:31 kahrl I think compatibility is not that important in this particular case
13:33 Krock done
13:33 kahrl okay, that's good too :)
13:35 Krock a +1 now please? *big eyes*
13:35 Krock *begging eyes
13:36 kahrl another one? :P
13:38 * Krock looks around for more +1
13:39 GhostDoge Krock: I can only give you a *1 :P
13:39 Krock at least something :)
13:40 kahrl alright alright
13:40 kahrl Ook. Ook.
13:40 GhostDoge Ook!
13:43 * Krock looks at sfan5
13:43 sfan5 hm?
13:43 Krock Could you donate me a +1 for the pull #1791
13:43 ShadowBot https://github.com/minetest/minetest/issues/1791 -- Add factor of maximal player transfer distance by SmallJoker
13:44 Krock ?
13:47 sfan5 Krock: sorry, I can only give a :+1:
13:49 Krock many thumbs up do not exactly do what I tried to reach
13:49 Krock s/tried/try
13:49 Krock thx
14:00 Krock \o/
14:37 CraigyDavi joined #minetest-dev
14:40 RealBadAngel joined #minetest-dev
14:42 RealBadAngel kahrl, sfan5 hi
14:43 RealBadAngel any comment on #1812?
14:43 ShadowBot https://github.com/minetest/minetest/issues/1812 -- Speed up removing(placing) a node (less block mesh updates). by RealBadAngel
14:46 Krock +1
14:50 Amaz joined #minetest-dev
15:00 RealBadAngel nobody can comment a commit with +5 and -6 lines ?
15:05 sfan5 RealBadAngel: why remove the TimeTaker?
15:06 nore joined #minetest-dev
15:14 RealBadAngel sfan5, why it should be initialized when not used?
15:14 RealBadAngel need it then turn it on]
15:15 RealBadAngel all around timetakers are commented out, just here it wasnt
15:16 RealBadAngel nore, please test #1812
15:16 ShadowBot https://github.com/minetest/minetest/issues/1812 -- Speed up removing(placing) a node (less block mesh updates). by RealBadAngel
15:18 kaeza joined #minetest-dev
15:28 iqualfragile joined #minetest-dev
15:35 hmmmm joined #minetest-dev
15:39 Sokomine joined #minetest-dev
15:52 VanessaE RealBadAngel: I see no obvious problems with the revised #1812 patch.
15:52 ShadowBot https://github.com/minetest/minetest/issues/1812 -- Speed up removing(placing) a node (less block mesh updates). by RealBadAngel
15:52 VanessaE (that is, in practice)
15:54 Taoki joined #minetest-dev
15:54 alexxs joined #minetest-dev
16:06 alexxss joined #minetest-dev
16:07 diemartin joined #minetest-dev
16:11 Megaf_ joined #minetest-dev
16:59 Calinou joined #minetest-dev
17:04 hmmmm RealBadAngel, LGTM
17:06 kahrl does that mean Lesb... oh
17:06 hmmmm Looks Good To Me
17:07 kahrl I always mix up those two acronyms
17:08 hmmmm do you mean LGBT?
17:08 hmmmm haha
17:08 kahrl yes
17:09 kahrl although "M" could mean "Metro" :P
17:25 MinetestForFun joined #minetest-dev
17:27 * Krock would like to see pulls 1812 (node speedup) and 1813 (mgv5) merged. C'mon devs, those things have been tested successfully
17:28 ShadowNinja RealBadAngel: I think you have enough approval to merge that speedup PR.
17:40 hmmmm wait, shadowninja, where did you get the "use two tabs for function indentation" from?
17:41 hmmmm BSD userland/kernel, linux kernel, and windows kernel all use single tabs for indentation
17:41 hmmmm well all microsoft code
17:41 ShadowNinja hmmmm: Do you mean two tabs for arguments past ~80 chars when calling a function?
17:42 hmmmm no, when defining a function
17:42 hmmmm int foobar(ReallyLongHugeBigTypeHere *really_big_long_type_here,
17:42 hmmmm \tint next_parameter_here;
17:42 hmmmm erm, remove that ;
17:42 Calinou I don't like double-tabs much either
17:42 hmmmm yeah.
17:43 hmmmm I make mine single-tabbed continuations for function definitions
17:43 ShadowNinja hmmmm: Oh, I got it from the code style guide.  It helps to differentiate from initializer lists which use only one tab.
17:43 hmmmm what code style guide??
17:43 Calinou http://dev.minetest.net/Code_style_guidelines
17:43 ShadowNinja ^
17:43 Calinou http://dev.minetest.net/Lua_code_style_guidelines
17:43 hmmmm yes, but YOU wrote that shadow
17:43 hmmmm that's why I'm asking
17:45 hmmmm wait actually...
17:45 ShadowNinja hmmmm: 217.226.206.67 wrote that.
17:45 hmmmm you just modified it, it was actually 217.226.206.67 who edited that
17:45 ShadowNinja Not my IP.
17:45 hmmmm okay
17:46 hmmmm so some jokester modified the code style to his liking
17:46 hmmmm it's a single tab, not multiple
17:47 Calinou do the current code use single tab or double tab?
17:47 hmmmm it's a huge mix
17:47 hmmmm minetest's code style at the moment is a free-for-all
17:47 hmmmm i use a single tab, i've seen people modifying existing code to have two tabs, etc.
17:48 hmmmm what do you guys think about the windows kernel function definition style?
17:51 ShadowNinja hmmmm: I don't know what that style is...
17:51 hmmmm int
17:51 hmmmm KiDoSomeFoobar(
17:52 hmmmm \tPVOID SomeThingHere,
17:52 hmmmm \tPKPROCESS TargetProcess,
17:52 hmmmm <so on and so forth>
17:52 hmmmm ) {
17:53 Calinou what's the difference between the usual style?
17:54 hmmmm int KiClass::doSomeFoobar(void *something_here, KPROCESS *target_process)
17:54 hmmmm {
17:55 ShadowNinja hmmmm: I pefer the closing parenthesis on the last parameter -- and I like to put more than one parameter on a line.
17:55 ShadowNinja So, no, because that's just about the only difference I see.
17:56 ShadowNinja Oh, and I like the type on the same line.
17:56 ShadowNinja template<foo> goed on a seperate line though.
17:56 ShadowNinja s/d/s/
17:57 hmmmm you mean like this:
17:57 hmmmm int KiClass::doSomeFoobar(void *something_here, KPROCESS *target_process) { ????
17:57 ShadowNinja hmmmm: Yes, I then normally break with two tabs if it gets too long.
17:57 hmmmm I used to do that but i don't like it because it looks worse and it's not 1TBF, BSD kernel style, linux kernel style, etc.
17:57 hmmmm K&R also
17:58 ShadowNinja hmmmm: Um, it looks like it is Linux kernel coding style.
17:59 ShadowNinja At least acording to https://www.kernel.org/doc/Documentation/CodingStyle
17:59 hmmmm ????
17:59 hmmmm However, there is one special case, namely functions: they have the
17:59 hmmmm opening brace at the beginning of the next line, thus:
17:59 hmmmm int function(int x)
17:59 hmmmm {
17:59 hmmmm body of function
17:59 hmmmm }
18:00 hmmmm that's the third code block under Chapter 3
18:00 ShadowNinja hmmmm: Yes, exactly.  But the closing parantheses ')' is on the same line.
18:00 hmmmm oh
18:00 hmmmm oh derp
18:00 hmmmm i thought you were talking about { for some reason
18:03 _Esteban joined #minetest-dev
18:22 _Esteban left #minetest-dev
18:27 kahrl sfan5: I've just installed weblate 2.0 on my test server
18:27 kahrl seems to work well, no problems yet
18:28 kahrl the email registration issue is also fixed
18:28 kahrl I've updated http://dev.minetest.net/Installing_Weblate already
18:30 Calinou cool
18:30 * Calinou promotes kahrl
18:31 kahrl I don't wanna be a manager :(
18:32 Calinou people don't want me to manage it :(
19:13 ShadowNinja sfan5: ^ Can you set that up?
19:14 sfan5 probably
19:29 EvergreenTree joined #minetest-dev
19:29 casimir_ joined #minetest-dev
19:35 cg72 joined #minetest-dev
19:35 cg72 left #minetest-dev
20:16 RealBadAngel https://github.com/minetest/minetest/pull/1804
20:17 RealBadAngel can we have this merged?
20:19 RealBadAngel also #1799
20:19 ShadowBot https://github.com/minetest/minetest/issues/1799 -- Implement WieldMeshSceneNode which improves wield mesh rendering by kahrl
20:19 RealBadAngel and #1809
20:19 ShadowBot https://github.com/minetest/minetest/issues/1809 -- Stop inserting empty lines at end of .conf file each time it is rewritten by Zeno-
20:19 VanessaE I noticed that yesterday
20:19 VanessaE thought it was something I did :P
20:19 RealBadAngel the last one is rly annoying
20:22 RealBadAngel also #1813
20:22 ShadowBot https://github.com/minetest/minetest/issues/1813 -- Add mgv5 by paramat
20:22 hmmmm YES, already, 1799 can get merged
20:22 hmmmm 1804 is approved
20:23 hmmmm 1809 is fine too
20:24 hmmmm mapgen v5 looks good and should be ready to go
20:24 Calinou looks far from the 0.3 one, especially caves
20:24 Calinou 1799 breaks something
20:25 hmmmm what does 1799 break
20:25 Calinou There are some of cases where one may use NPOT textures (eg. mob/model texture as wield_image). You are breaking functionality that is currently being used by some mods…
20:25 Calinou so do a news post for this
20:25 Calinou (or on release)
20:25 RealBadAngel Calinou, models are not affected
20:25 Calinou also mgv5 is overrated, but add anyway
20:25 Calinou RealBadAngel, you got it completely wrong
20:25 Calinou I'm talking about wield_image
20:25 Calinou can you read?
20:26 RealBadAngel ofc, but the problem is only for extrude code
20:26 RealBadAngel not the wield mesh at all
20:26 RealBadAngel if item or a node has own mesh, it is not called
20:27 RealBadAngel capisci? :)
20:28 RealBadAngel for other things we do use pow2 textures, no exceptions
20:28 VanessaE there *are* a scant few objects in a few games that use npo2 textures but those can just be replaced.
20:28 VanessaE (recall that default junglegrass was at one time)
20:29 kahrl I can easily make any npo2 texture use the 512x512 extrusion mesh
20:29 RealBadAngel cost of it?
20:29 kahrl that will be a bit slower but the problem should be nearly invisible
20:30 kahrl RealBadAngel: well that clearly depends on your GPU
20:30 VanessaE kahrl: why even extrude in that case?
20:30 VanessaE better to render npo2 as flat.
20:30 kahrl flat looks even more horrible
20:31 VanessaE not really, but think of it as a warning that looks less bad than the artifacts that were shown in those screenshots
20:32 RealBadAngel in general im against of using npot2 textures
20:32 RealBadAngel theres too many things depending on that
20:32 kahrl or I could make it generate an extrusion mesh on use, let me test how that performs
20:33 VanessaE kahrl: until you did 1799, I was using that no-extrude hack and actually, it really didn't look bad most of the time.  1799 makes that hack pointless now, save for this idea that is.. just saying.
20:34 RealBadAngel im waiting with 1799 then, others im going to merge
20:36 kahrl do we have a function that checks if a number is a power of two?
20:36 kahrl something like bool is_power_of_two(u32 n) { return n != 0 && (n & (n-1)) == 0; }
20:37 VanessaE kahrl: can't just take the x rez, compute sqrt() and check if the result is an int?
20:37 VanessaE right, like that :)
20:37 kahrl power of two != square :P
20:37 hmmmm hrmm, kahrl not to my knowledge
20:37 hmmmm otherwise it'd surely be in util/numeric.h
20:37 kahrl I'll just add that to util/numeric.h then
20:38 kahrl great minds think alike ;)
20:38 VanessaE well ok compute *both* dimensions and compare then :P
20:38 kaeza FWIW, I recall there was an issue in Android where npot textures were drawn as part garbage (and minor issue, but burningsvideo seriously breaks with npot textures, but who uses BV anyway?)
20:39 hmmmm minetest is pretty much broken with BV
20:39 VanessaE kaeza: sapier fixed that by matting such textures onto a lager canvas
20:39 hmmmm i don't know why we have it as an option even
20:39 VanessaE I think
20:44 RealBadAngel hmmmm, i was against adding that
20:45 RealBadAngel having an option for regular user in main menu, that will surely fuck up things is shooting own leg
20:45 Calinou BV?
20:46 RealBadAngel burningsvideo
20:46 RealBadAngel software also is a no go
20:46 sfan5 software graphics suck
20:46 sfan5 except llvmpipe
20:46 sfan5 but that is (as expected) slow
20:47 RealBadAngel that setting shall be removed from menu (or at least tweaked to contain only useable ones)
20:48 RealBadAngel btw, if we do support only OpenGL, why the duck such setting is needed?
20:48 kaeza RealBadAngel, because D3D
20:48 sfan5 Direct3D9 runs faster for some people
20:48 RealBadAngel we dont support d3d
20:48 Calinou sfan5, on machines other than netbooks?
20:48 Calinou no shaders, remember
20:49 kaeza RealBadAngel, "we don't support d3d [shaders]"
20:49 kaeza no shaders ain't the end of the world
20:49 sfan5 Calinou: I duno
20:49 RealBadAngel yes, ofc, but shaders are essential
20:49 sfan5 dunno
20:51 Calinou shaders may get mandatory some day
20:51 * Calinou prediction
20:51 kaeza in any case, removing BV/D3D from the main menu is useless; a standard Irrlicht build will contain those drivers anyway, so you won't save anything
20:51 Calinou you'd save people from choosing them
20:53 RealBadAngel kaeza, point is user given such option in game menu can make the game not working
20:53 RealBadAngel even main menu is fucked up then
20:54 kaeza alright
20:55 RealBadAngel i could live with that option in dev, but its unaceptable for release
20:55 RealBadAngel i dont want to see "omfg, the game is broken with this setting"
20:56 RealBadAngel and that will surely happen
20:58 casimir With v5 I get an error
20:58 casimir ERROR[ServerThread]: WARNING: saveBlock: Block failed to save (-8,-2,-8): SQL logic error or missing database
20:58 sfan5 thats just the DB backend
20:59 sfan5 sqlite likes to break
20:59 casimir Yes, but I get this for all the blocks. And they regenerate later.
21:00 sfan5 does this only happen with v5?
21:02 kaeza while at it, can some core dev please comment on #1770? (i.e. what's the reason for the far clip distance being that low, and if it can be removed/enlarged)
21:02 ShadowBot https://github.com/minetest/minetest/issues/1770 -- Remove wield item camera clipping.
21:02 RealBadAngel kaeza, kahrl's pull fixes that
21:02 kaeza huh?
21:02 RealBadAngel #1799
21:02 ShadowBot https://github.com/minetest/minetest/issues/1799 -- Implement WieldMeshSceneNode which improves wield mesh rendering by kahrl
21:03 kaeza ah
21:03 RealBadAngel kaeza, so just hold on a bit, until we get to a consensus what to do with npot2 textures issue
21:04 casimir only v5. And it says it can not load map_meta.txt
21:04 casimir but map_meta.txt exists
21:05 kaeza RealBadAngel, thanks. did not notice it included a fix for that
21:05 SmugLeaf joined #minetest-dev
21:05 Krock casimir, I can not reproduce this bug
21:05 RealBadAngel can it be related to backend used?
21:06 Krock using sqlite3 - no error
21:06 casimir I use sqlite3
21:07 Krock deletes map.sqlite - no fail
21:07 PenguinDad casimir: could it be possible that your sqlite3 is broken?
21:09 casimir ok, it happens for the other mapgens too. But only since v5 was merged.
21:09 casimir I will try to reinstall sqlite3
21:09 casimir <
21:11 kahrl ugh... it now somehow leaks the mesh scene node
21:14 casimir joined #minetest-dev
21:14 RealBadAngel looks like v5 is not working as expected
21:15 RealBadAngel http://i.imgur.com/i35GD89.png
21:16 RealBadAngel i suppose that was not a goal of v5 ;)
21:16 Krock I alrady saw that in paramat's screenshot
21:16 Krock https://cdn.mediacru.sh/Fe9gWHUvdDTI.png
21:18 RealBadAngel land shapes are nice, but i think paramat should work a bit to make v5 useable for regular playing
21:20 hmmmm RealBadAngel, how is v5 not working as expected?
21:21 Calinou caves are too tight
21:21 RealBadAngel see the screenshot
21:21 Calinou doesn't feel like 0.3 at all
21:21 hmmmm yes RealBadAngel... I am asking you to explain what is wrong with the screenshot
21:22 RealBadAngel no trees, no plants, no biomes, sharp cut out areas
21:22 hmmmm did you realize that trees, plants, and biomes are the responsibility of the game and not the engine?
21:22 Calinou you need to define them like in v7
21:23 RealBadAngel ah ok
21:23 Calinou so make some v6-like biome definitions if you wish
21:23 Calinou I'd like that a lot
21:23 Calinou would probably be default in minetest_game
21:23 Calinou to make it integrate seamlessly with v6
21:23 Calinou (perhaps an additional snow biome is welcome)
21:23 hmmmm I can't stand it when people criticize mapgens using the biome infrastructure because no biomes have been added by the games
21:24 hmmmm mapgen ain't broke:  it's YOUR fault it's all stone.
21:24 casimir It still complains about missing map_meta.txt but sqlite issues are gone somehow...
21:25 VanessaE hmmmm: in mapgen v5??
21:27 VanessaE with v7 it makes sense if it's all stone, etc etc etc.  but with v5, it's it kinda expected to at *least* have something that resembles a 0.3.x game right out of the proverbial box?
21:27 hmmmm then add them to minetest_game.  there is absolutely no excuse.
21:28 kahrl ok, memleak is fixed, #1799 updated
21:28 ShadowBot https://github.com/minetest/minetest/issues/1799 -- Implement WieldMeshSceneNode which improves wield mesh rendering by kahrl
21:28 VanessaE hmmmm: sure there is.  haven't you said more than once that the APIs needed to put stuff on such a bare map are unstable?
21:28 kahrl it always creates a new texture whenever an item with npot texture is selected
21:28 hmmmm sure
21:28 kahrl a new extrusion mesh *
21:29 hmmmm i have a new idea:
21:29 hmmmm why don't you stop bitching and start coding
21:29 hmmmm same goes for everyone
21:29 IvanMaldonado joined #minetest-dev
21:29 VanessaE I'm bitching?  O_o
21:29 IvanMaldonado sup
21:29 VanessaE getting a little defensive here, aren't you?
21:29 hmmmm yeah, i'm a little moody right now
21:30 * VanessaE wanders off
21:30 hmmmm i'm just tired of too much talking and not enough coding'
21:30 IvanMaldonado After adding minetest ppa, how can I install minetest from the terminal?
21:30 hmmmm IvanMaldonado, sounds like a #minetest question
21:30 IvanMaldonado hmmmm, Get over it, Vanessa is a woman, you can't win
21:31 VanessaE IvanMaldonado: that was uncalled for, and you do NOT know me.
21:31 kilbith IvanMaldonado : go to #minetest please
21:31 IvanMaldonado hmmmm, you see? I'm already in troubles now, lol
21:31 IvanMaldonado ok, where can I get help?
21:31 IvanMaldonado sure
21:31 IvanMaldonado thanks
21:31 hmmmm bye
21:31 IvanMaldonado left #minetest-dev
21:32 hmmmm good grief
21:33 BlockMen joined #minetest-dev
21:34 VanessaE hmmmm: if anyone's qualified to add those settings to minetest_game, HE is
21:34 * VanessaE points at BlockMen  :P
21:34 BlockMen wut?
21:34 VanessaE BlockMen: default biomes, decorations, etc for mapgen v5 and v7
21:36 BlockMen ah, we have mapgen v5 back
21:36 BlockMen nice
21:36 BlockMen well, maybe better paramt adds those and makes a pull
21:37 BlockMen or hmmm
21:38 BlockMen or who else is mapgen pro here
21:40 kahrl can http://dev.minetest.net/TODO#Mapgen_V5 be removed?
21:41 BlockMen is there a release day set already?
21:41 kahrl no
21:41 BlockMen k
21:44 BlockMen sfan5, did you agree on https://github.com/minetest/minetest_game/commit/dcce6e97959647fd625c1ce219721c0b800f4938 ?
21:45 sfan5 no
21:45 sfan5 but I'm fine with that
21:45 EvergreenTree joined #minetest-dev
21:47 BlockMen the reason why i didnt "fix" it was that it was a effect like the fire "burns" the grass
21:51 Taoki joined #minetest-dev
21:51 kilbith BlockMen: https://github.com/minetest/minetest_game/pull/331   Leaves updated, i'm waiting for your feedback please.
21:51 kilbith nore & sfan5 are OK with it
21:52 BlockMen kilbith, already merged :P
21:52 kilbith uh ?
21:52 kilbith oh ok
21:53 BlockMen what the?
21:53 BlockMen great...history rewrite :\
21:56 kilbith BlockMen: I apologize for my git crap :/
21:57 RealBadAngel bout shaders and removed from there finalcolorblend. shall i add them again now?
21:58 RealBadAngel if we are going to make 0.4.11 imho it should be added back
21:58 RealBadAngel if we are going to make the lights and then 0.4.11 then no point doing this
21:59 BlockMen kilbith, then squash you other pulls ;)
21:59 RealBadAngel hmmmm, kahrl?
22:00 BlockMen RealBadAngel, you mean #1705?
22:00 ShadowBot https://github.com/minetest/minetest/issues/1705 -- Regression in smooth day-night transition
22:00 kilbith BlockMen: my other pulls are clean
22:01 RealBadAngel BlockMen, yes
22:01 BlockMen kilbith, e.g. 322 has still 4 commits for one file changed
22:02 BlockMen RealBadAngel, then yes ;)
22:02 kilbith BlockMen : 1 commit is for reverting the cobbles, the other ones are for updating the ladder, that's all
22:02 RealBadAngel any other dev vote for that?
22:02 kahrl RealBadAngel: I don't know what the current release schedule is
22:03 RealBadAngel kahrl, imho we shall do what rumours are saying ;)
22:03 BlockMen kilbith, and those can be squashed into 1 commit
22:03 kilbith BlockMen: should I close / re-open the pull for the ladder ?
22:03 RealBadAngel make one release now (or soonish) and the xmas edition as next ;)
22:04 RealBadAngel and i would rly like to light the lights on xmas :)
22:04 BlockMen kilbith, no. just rebase them
22:04 kilbith ehh...
22:06 kilbith i'm not a git expert :/
22:06 kahrl RealBadAngel: I'd be ok with that
22:08 BlockMen kilbith, ever heared of google(or other web search providers)? ;)
22:08 BlockMen http://davidwalsh.name/squash-commits-git
22:08 kilbith i was searching on google
22:08 kilbith unfortunally i still dont understand
22:09 kahrl RealBadAngel: what do you think, should I merge #1799 now or should I wait for more testing
22:09 ShadowBot https://github.com/minetest/minetest/issues/1799 -- Implement WieldMeshSceneNode which improves wield mesh rendering by kahrl
22:10 RealBadAngel merge it
22:10 RealBadAngel if something pops out, more testers will help fix it
22:11 proller joined #minetest-dev
22:16 RealBadAngel #1808
22:16 ShadowBot https://github.com/minetest/minetest/issues/1808 -- Adjust the values of dirs1 and dirs2 so that rotate_and_place orients textures correctly by dvere
22:16 RealBadAngel any objections?
22:22 RealBadAngel #1734
22:22 ShadowBot https://github.com/minetest/minetest/issues/1734 -- Add tooltips to main menu subgames button bar by Wuzzy2
22:22 RealBadAngel the same question
22:31 loggingbot_ joined #minetest-dev
22:31 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/
22:31 RealBadAngel kahrl, indeed
22:32 RealBadAngel shall i modify it and merge then?
22:32 kilbith BlockMen: I'll close / re-open the ladder pull, after many tries, I can't understand at all the squashing
22:32 BlockMen kilbith, no
22:32 BlockMen then leave it as is
22:32 kilbith ok, np
22:33 kahrl RealBadAngel: do it
22:33 RealBadAngel kahrl, on it
22:36 BlockMen ''tell ShadowNinja plz don't push things to minetest_game without agree of _game devs
22:37 sfan5 oh
22:37 sfan5 the bot is not there
22:37 BlockMen :\
22:37 sfan5 ShadowNinja: bring H-LuaBot here
22:37 kaeza it is, just not voiced
22:38 sfan5 BlockMen: try using 'ShadowBot: tell'
22:38 kaeza ''(help tell)
22:39 kaeza :/
22:39 BlockMen ShadowBot: tell ShadowNinja plz don't push things to minetest_game without agree of _game devs. or add a
22:39 ShadowBot BlockMen: O.K.
22:39 BlockMen ah :)
22:40 BlockMen sfan5, https://github.com/minetest/minetest_game/pull/326 ?
22:41 sfan5 BlockMen: looks good
22:41 BlockMen merge? ^
22:41 sfan5 yes
22:45 Krock I thought minetest does not save incomplete mapblocks to the database; but still, a Lua error can corrupt some chunks
22:45 Krock ^ hmmmm
22:46 Krock I've got the situation where the mapgen threw an error and after rejoining, there was a big hole of air.
22:47 hmmmm why are you pointing that at me
22:48 Krock hmmmm, because of mapgen stuff and so.
22:48 hmmmm that doesn't have anything to do with mapgen
22:48 Krock well, it has to do with the state of a mapblock
22:48 Krock who could I punch with this problem?
22:49 VanessaE BlockMen:  :'(
22:49 BlockMen VanessaE, what?
22:49 Krock *creates issue first*
22:52 cg72 joined #minetest-dev
22:56 cg72 is the skybox setup to handle animated textures???
22:56 RealBadAngel no
22:57 RealBadAngel some1 should code day/night cycles for skyboxes and transitions
22:57 RealBadAngel and maybe make them layered
22:57 RealBadAngel for clouds for example
22:58 cg72 color changing clouds for sunsets :)
22:58 RealBadAngel terasology has that code, easy to import
22:58 RealBadAngel if some1 *understands* java ofc ;)
22:59 cg72 lol i know c++ and java is close enough lol
23:01 cg72 rba is it posiable to have the clouds set up an a .ong file and be maybe plantlike or some 3d ish thing???
23:02 RealBadAngel clouds are not customizeable
23:03 RealBadAngel plantlike drawtype is just showing 2 flat images, its not 3d
23:03 RealBadAngel for 3d use meshnodes
23:04 RealBadAngel http://i.imgur.com/fQeMN8o.png
23:04 cg72 well i understand but they look more 3d than just flat lol, and i was wondering if it would be easy to do the c++ in the engine for the clouds to be changed
23:04 RealBadAngel or that http://i.imgur.com/maxiwQZ.png
23:05 RealBadAngel clouds are object that are far away, they dont need to be 3d at all
23:05 cg72 if i only knew how to use blender lol
23:06 cg72 so  it can be made to have a png not just a box lol i hate the clouds now  and am trying to do something about them ;)
23:07 RealBadAngel cg72, https://www.youtube.com/watch?v=_UO_8U39qIM
23:18 chchjesus joined #minetest-dev
23:22 OldCoder VanessaE, Sapier's patch was not sufficient. I am testing a version that combines his changes. Zeno's and mine
23:22 VanessaE ok
23:33 CraigyDavi joined #minetest-dev
23:34 cg72 RealBadAngel: how??? ;)
23:34 RealBadAngel how what?
23:34 cg72 the clouds fron the vid
23:35 RealBadAngel just a skybox
23:35 RealBadAngel only difference i added there rotations in core
23:35 cg72 but it moves!!!
23:35 RealBadAngel one liner
23:35 cg72 how hard is that and dese the latest have it?
23:43 RealBadAngel i will look for the code and post the link here later on
23:43 RealBadAngel this is really one liner (ofc without Lua api for it)
23:43 cg72 lol i wish the api had all the goodies in it
23:44 RealBadAngel one day
23:44 RealBadAngel but at some point we need to make it proper way
23:44 RealBadAngel instead of simple skybox, make it layered, animated and blending
23:45 RealBadAngel just a skydome
23:45 cg72 yes, i wish i could but with graphics im really useless

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