Time  Nick         Message
08:11 kaeza        sofar, in `if obj and obj:get_luaentity() then`, `obj` is guaranteed to be non-nil
08:11 kaeza        also, `obj:get_luaentity()` could be localized too :P
08:12 kaeza        (it's used both in the check and below to get the definition)
08:12 kaeza        s/definition/itemstring/
08:13 kaeza        I wonder where's paramat with his insightful comments on performance
08:55 everamzah    lotta issues and PRs, holy
09:55 RealBadAngel hi
09:55 RealBadAngel how do you like it? http://i.imgur.com/dCfPFYq.png
09:56 everamzah    a lot.  first impression: "why is wielditem not blurry" <-- but, details
09:56 everamzah    looks good
10:00 Calinou      yay DoF
10:00 Calinou      it focuses on the wrong thing, and is a bit too strong though.
10:00 Calinou      be sure to make the blur strength adjustable in minetest.conf
10:01 Fritigern    "It will be a cpuple of days before I have DoF working".... 6 hours later: "Hey guys, look at my DoF!" :-)
10:05 RealBadAngel ive used shader by martinsh, it has shitload of settings
10:06 nrzkt        RealBadAngel, seem player has drunk
10:06 RealBadAngel http://pastebin.com/9QcQruD5
10:06 RealBadAngel who wanna try to configure it? ;)
10:07 RealBadAngel Calinou, atm i let it autofocus on the middle of the screen
10:08 Calinou      RealBadAngel, sounds good, but be sure to add some tweening between values
10:08 Calinou      so that the change is never instant (it shouldn't be)
10:08 Calinou      ie. if focus distance goes from 5 units to 10, it should take like 0.2 seconds to interpolate
10:08 Calinou      ideally in a non-linear (cubic) way
10:09 RealBadAngel that would require some c++ code
10:11 RealBadAngel that can be developed later on
10:11 RealBadAngel right now autofocus is enough
10:44 RealBadAngel Calinou, actually it works quite nice without delays
10:48 RealBadAngel ive seen the delays like you described in terasology, thx to that game got sloooow feeling
10:48 Calinou      it's unrealistic to change focus instantly
10:48 Calinou      nobody can
10:49 Calinou      it reminds me of "cheap" DoF 2010 games had
10:50 RealBadAngel propably we can fine tune it later on
10:50 RealBadAngel meanwhile wanna try how it works?
10:51 Calinou      yes
10:51 RealBadAngel imho thx to this game gets unbeliveable 3d kick in the ass
10:51 Calinou      needs motion blur too :P
10:51 Calinou      when it's well done, it works well
10:51 Calinou      obviously not enabled by default, but it's something
10:51 RealBadAngel ok, gimme a moment, i will upload latest changes
10:52 nrzkt        this blur could be useful in water
10:53 nrzkt        and when you take damage
10:53 nrzkt        maybe add a packet to trigger a blur with a color on client could be useful :D
10:53 nrzkt        with a timer
10:53 nrzkt        this permit to add more client side effects :D
10:55 Calinou      the blur could be strengthened in liquids, indeed.
10:55 RealBadAngel Calinou, #3623
10:55 ShadowBot    https://github.com/minetest/minetest/issues/3623 -- Postprocessing (WIP) by RealBadAngel
10:55 RealBadAngel build in place, dont try to change screen size when in game ;)
10:56 RealBadAngel and start from main folder
10:56 Calinou      I used "git checkout https://github.com/minetest/minetest/pull/3623" :P
10:56 Calinou      thanks to hub
10:56 Calinou      why can't screen size be changed while playing?
10:57 RealBadAngel mrt is changing RTT sizes automatically, irrPP not yet
10:57 Calinou      http://hastebin.com/ojapiqomol.coffee :(
10:57 Calinou      build error
10:58 RealBadAngel lemme try too
11:01 RealBadAngel forgot to update irrPP, hold on
11:03 RealBadAngel Calinou, try now
11:04 Calinou      great, I can even git pull on your branch
11:04 nrzkt        RealBadAngel, except the indent, i don't see any stability problem atm nor performance issue in the coding style
11:05 RealBadAngel nrzkt, dont even try to check code style yet, please ;)
11:05 RealBadAngel to see why, check irrPP sources
11:05 RealBadAngel theyre all formated with spaces lol
11:06 Calinou      it seems to focus on the wrong thing…
11:06 Calinou      I'm looking at a cactus, it is very blurry.
11:06 Calinou      https://lut.im/kYqcqcaP0Z/DIBhtGhZmETFaWFE.png
11:07 Calinou      your blur is also a bit too strong IMO
11:08 RealBadAngel open dof,.frag in /client/shaders/postprocess
11:08 RealBadAngel the shader is highly configureable
11:09 RealBadAngel (as all others made by martinsh)
11:09 RealBadAngel you can play with settings
11:09 Calinou      "This work is licensed under a Creative Commons Attribution 3.0 Unported License."
11:09 Calinou      :(
11:09 Calinou      GPL-incompatible license
11:09 Calinou      also not suited for code
11:10 Calinou      I'm aware it is indirectly GPL-compatible since you can upgrade it to CC BY 4.0, but still.
11:10 RealBadAngel oops
11:11 RealBadAngel i really dont understand all this licensing shit
11:11 RealBadAngel "So you are free to share, modify and adapt it for your needs, and even use it for commercial use."
11:12 Calinou      we can use it legally, but it is tricky
11:13 Calinou      I suggest you contact the author and tell them to license their shaders under MIT or CC0…
11:13 RealBadAngel can you telll whats wrong for us in this sentece?
11:13 Calinou      attribution.
11:13 Calinou      it means we might need to add attribution in the game credits
11:14 RealBadAngel nah, since i do know how it works i will just write my own
11:14 RealBadAngel at least i have tested depth maps with it ;)
11:17 nrzkt        Calinou, stop drinking vodka
11:18 nrzkt        i think the blur should be < 5%
11:18 RealBadAngel lol
17:12 RealBadAngel hmmmm, the truth maybe something in between. its seems to be like kinda mix of both HW and software, im not sure anymore
17:13 RealBadAngel ive put that away to finish postprocessing
17:15 RealBadAngel one thing im sure, that mt's disabling/enabling shaders is just changing  their/our shaders
20:19 paramat      nore sfan5 i might merge this later game#810
20:19 ShadowBot    https://github.com/minetest/minetest_game/issues/810 -- Default/mapgen: Add gravel beach in Tundra by paramat
20:20 sfan5        hmm
20:20 sfan5        paramat: the gravel texture is ugly
20:21 paramat      perhaps we can change that later, how could it be improved?
20:24 sofar        I'd like to to look a bit more like real gravel: have pebble-like pixel clumps that are 2-4-6 pixels or so, and of various colors
20:24 sofar        black, white, reds, browns
20:24 sofar        hints of sand texture color, hints of stone color texture
20:25 sofar        maybe even make it by taking tiny cuts out of other textures
20:26 paramat      yeah this needs improving
20:26 paramat      btw there are some game PRs that just need approvals and don't need much code review, please could you lookk through them?
20:26 paramat      (look)
20:27 paramat      most of the time we can review code ourselves but i can't approve on my own :)
20:28 paramat      comments will be appreciated
20:29 sofar        paramat: happy to. especially if links are spammed in here ;)
20:29 paramat      hm ok
20:31 paramat      it's sort of all of them but i'll try to pick some..
20:33 sofar        I'll chomp through it from the top
20:35 paramat      sfan5 game#731 needs another -1 to close
20:35 ShadowBot    https://github.com/minetest/minetest_game/issues/731 -- Smooth boat turning to reach player yaw as he accelerates forward by beyondlimits
20:35 paramat      assuming 3 out of 5 devs against to close
20:36 sofar        yeah, that one should get closed
20:36 sfan5        paramat: agree with your reasons, you can close game#731
20:36 ShadowBot    https://github.com/minetest/minetest_game/issues/731 -- Smooth boat turning to reach player yaw as he accelerates forward by beyondlimits
20:36 paramat      ok same with game#615
20:36 ShadowBot    https://github.com/minetest/minetest_game/issues/615 -- Replace bronze with gnucydur by FreeLikeGNU
20:36 paramat      nore and i against
20:36 sfan5        >gnucydur
20:36 sfan5        lel
20:36 sfan5        -1 from me too
20:37 paramat      game#778
20:37 ShadowBot    https://github.com/minetest/minetest_game/issues/778 -- ability to dye coloured wool white by tenplus1
20:37 sofar        I can't believe people come up with stuff like that. Then again, we have mese X/
20:38 paramat      game#743 but i suspect you're against
20:38 ShadowBot    https://github.com/minetest/minetest_game/issues/743 -- Create API for fence.register, and use it. by sofar
20:39 red-001      game#602 ?
20:39 ShadowBot    https://github.com/minetest/minetest_game/issues/602 -- Add access privilege that allows players to bypass protection. by red-001
20:39 sofar        I'm not a fan of having to redo the recipe
20:39 red-001      I would help server owners
20:40 red-001      It*
20:40 paramat      sofar you mean the fallback recipe suggested by pilzadam?
20:41 paramat      i agree with you there
20:42 paramat      sfan5 game#785 i have no idea if this is a good idea
20:42 ShadowBot    https://github.com/minetest/minetest_game/issues/785 -- Move character.png to textures. by sofar
20:42 sfan5        me neither
20:44 paramat      sfan5 game#792 just needs a +1, sounds are nicely balanced now
20:44 ShadowBot    https://github.com/minetest/minetest_game/issues/792 -- Snow footsteps: lower volume by sofar
20:45 sofar        a fallback recipe would be possible thought, in c++ (I confirmed in the code that it's doable - not easy, but doable)
20:45 paramat      sloppy idea though
20:46 sofar        it's not so bad. You'd call the recipe lookup code twice - once with a flag "don't use groups" and once with "use groups" if the former didn't return a recipe
20:46 sofar        no changes in lua needed
20:47 sofar        and the hit would be minimal
20:47 sofar        anyway, I think the idea of adding conflicting recipes is problematic by itself
20:48 paramat      then 2 big changes, comments appreciated game#789 game#793 i provisionally support both
20:48 ShadowBot    https://github.com/minetest/minetest_game/issues/789 -- New mesh door models, and extensive door API by sofar
20:48 ShadowBot    https://github.com/minetest/minetest_game/issues/793 -- Improve Chest appearance - opening chests. by sofar
20:49 paramat      those are still somewhat WIP
20:49 paramat      and/or waiting on engine PRs
20:49 sofar        it looks like that timer thing will go through though
20:49 paramat      yeah
20:49 sofar        so we can wait for that
20:50 sofar        was there anything remaining for the door one though?
20:50 sofar        I even posted changes to mesecons to make it work
20:50 paramat      backface culling?
20:50 sofar        hmmmm technically yes
20:51 sofar        not sure why the last PR is held up at this point. it should be good
20:51 paramat      lack of devs heh
20:51 sofar        yeah, I'm not helping either. Pushing more PR's and I'm not an accepted reviewer yet ;)
20:52 sofar        the one PR I'd like to see merged though is the timer one
20:52 sofar        my bet is that is causing all sorts of perceived server lag
20:52 paramat      sfan5 apart from those perhaps sometime you can quickly go through all game PRs and state your opinion or +1 -1
20:53 sofar        I mean, the connected node boxes one should just get merged as well, it's going to help a ton going forward making servers load faster
20:53 sofar        but I can hold out
20:55 paramat      yeah there are PRs i have not mentioned and should have, but again all of them need atention and game devs are very elusive at the moment
20:57 sofar        #768 should just get merged
20:57 ShadowBot    https://github.com/minetest/minetest/issues/768 -- Added show_debug to minetest.conf.example by 0gb-us
20:58 sofar        game 768
21:00 red-001      sofar I would prefer the blast intensity at one node from center to be 1 and not 0.4 recursive
21:00 red-001      809
21:01 red-001      that also seems to give nicer number in general
21:02 red-001      maybe the number should be rounded?
21:02 sofar        in any case we should aim for intensity to be [0, 1] at all times
21:02 red-001      I agree
21:02 sofar        we can play with the math
21:03 red-001      It has to be relatively  fast
21:03 sofar        math is fast
21:03 paramat      oh crumbs game 768 my nemesis
21:04 sofar        paramat: just merge it.
21:04 sofar        performace overhead is nil
21:04 sofar        those messages are already there
21:04 paramat      well just item type might be ok
21:06 paramat      i should ask a trusted server owner ie VanessaE how useful the extra info is
21:06 sofar        red-001: intensity = 1 / math.pow(math.min(1, dist), 2)
21:06 sofar        red-001: just do that, then
21:07 sofar        damn
21:07 sofar        red-001: intensity = 1 / math.pow(math.max(1, dist), 2)
21:07 sofar        always mix min/max :(
21:07 kaadmy       heh me too
21:08 sofar        red-001: could scale differently on dist as well
21:08 sofar        red-001: intensity = 1 / math.pow(math.max(1, dist / 2), 2)
21:08 sofar        would make blast intensity go double as far
21:09 paramat      yes intensity 1 at 1 node
21:09 sofar        yeah, I think that's OK
21:09 sofar        that would be
21:09 paramat      as the centre is tnt itself
21:09 sofar        intensity = 1 / math.pow(math.max(1, dist), 2)
21:09 sofar        well everything within 1.0 of the center of the tnt block would be 1
21:09 paramat      yes
21:09 sofar        including the 6 neighboring nodes
21:10 sofar        that seems reasonable
21:10 paramat      yes
21:10 red-001      rounding can be done by mods
21:10 sofar        I put in a comment
21:12 sofar        somewhere between 3-4 dist it'll drop to below 0.1, which seems ok
21:12 sofar        I think you want to calculate player damage with the same curve
21:13 sofar        make it kill at 2 nodes guaranteed, maybe even 3
21:13 sofar        damage = 100 * intensity, or something like that, is what you'd end up with
21:13 red-001      yes
21:13 sofar        anyway, I suspect we'll do many more TNT prs
21:14 red-001      tnt doesn't check node further then 3
21:14 sofar        it should damage players and objects beyond that radius
21:14 red-001      it does
21:14 red-001      radius 6 for damage
21:15 red-001      radius 3 for block damage
21:15 red-001      but the second one is also a setting
21:15 red-001      so....
21:15 red-001      I need to add support for that
21:16 sofar        each tnt block should have an intrinsic power
21:17 sofar        and that should be used for calculating any radius
21:17 red-001      ?
21:17 red-001      what do you mean?
21:17 sofar        radiuses should not be hardcoded
21:18 sofar        we should put the radius at the point where 1/dist^2 = 0.05
21:19 sofar        or something like that
21:19 red-001      thats what I'm doing
21:19 sofar        so that results in dist = sqrt(20)
21:20 sofar        so the radius for block breaking would be 1/dist^2 = 5%
21:20 sofar        and the damage radius would be 2x the block breaking radius
21:21 sofar        or maybe, 1% -> r=10
21:21 red-001      ohh
21:21 red-001      I see
21:22 sofar        and because you pick 0.01, you can multiply intensity with 100 to get damage == 1
21:22 sofar        which is a nice round simple number of damage points at max range
21:22 red-001      that might be a bit harder for server owners...
21:22 red-001      but idk
21:22 sofar        math is fast
21:23 red-001      it is indeed
21:23 sofar        so at r=1, damage == 100 (boom, dead)
21:23 sofar        r = 2, damage == 25 (dead too)
21:24 sofar        r = 3, damage == 11 (ouch)
21:24 red-001      well something to add to tnt api pr
21:24 sofar        4 -> 6
21:24 sofar        5 -> 4
21:24 red-001      but I will add an option to override but radius
21:25 sofar        (3, 2, 2, 1 ....)
21:25 red-001      both
21:25 sofar        I would make it all use physics only, with the exception of the tnt base power value
21:26 sofar        I can imagine a firecracker with 0.1 power
21:26 sofar        it would do 10 damage if you'd stand on top of it :^)
21:26 red-001      minetest isn't that relativistic
21:26 sofar        no reason to avoid these models
21:28 red-001      you are right about that
21:29 red-001      but what if I want to make highly unrealistic explosion?
21:29 sofar        large?
21:29 paramat      math is slow in lua
21:29 sofar        no problem, just set power to 1000000
21:30 paramat      hm i see your point though
21:31 red-001      one that only damages blocks for example
21:32 paramat      with simple inverse square, as tnt power and radius increases intensity remains constant at a cetain distance
21:33 sofar        intensity = tntpower / sqr(max(1, dist))
21:34 sofar        to make things more realistic, you could give blocks a "resistance" value
21:35 sofar        e.g. stone.resistance = 0.5
21:35 sofar        sand.resistance = 0.1
21:35 red-001      so my newest mod?
21:35 sofar        and you could break blocks way beyond the 2-3 node range
21:36 red-001      maybe a radius and pressure parameter for an explosion?
21:37 sofar        intensity = math.min(tntpower / sqr(max(1, dist)), 1)
21:37 sofar        hmmm
21:37 red-001      not fully realistic but it allows a lot a of different types of tnt
21:37 sofar        base tnt should just have power = 1
21:37 red-001      yes
21:37 sofar        no matter what
21:38 red-001      well
21:38 sofar        I don't see why you can't have intensity over 1
21:38 red-001      it has a range parameter
21:38 sofar        1 just means to break everything
21:38 sofar        maybe 2 means destroy and do not drop items
21:38 red-001      +1
21:38 sofar        that would work well I think
21:38 sofar        you'd get dropped itemstacks at the edges
21:39 red-001      yes it would
21:39 sofar        entity ejection is broken
21:39 sofar        like I said, tnt is not even half done at this point
21:39 red-001      local intensity = (radius / 3 / dist ^ 2)?
21:40 sofar        why are you putting radius back in?
21:40 sofar        max_search_radius = math.sqrt(20) * tnt.power
21:41 sofar        that's how large the "hole" will be
21:41 red-001      it should be something that can be   overridden
21:41 red-001      if you want a tnt api that is
21:42 sofar        that doesn't make sense
21:42 red-001      why not?
21:42 sofar        bombs don't explode one block and leave the one next to it untouched :)
21:42 sofar        your formula
21:42 sofar        local intensity = (radius / 3 / dist ^ 2)
21:43 sofar        that basically says
21:43 sofar        tnt.power = radius / 3
21:43 sofar        I'm saying, tnt.power = 1
21:43 sofar        and from that results
21:43 sofar        calculated at boom time
21:43 sofar        that radius = tnt.power * sqrt(20) = 4.47
21:43 sofar        (rounded = 4)
21:44 sofar        and damage radius = tnt.power * sqrt(100) = 10
21:44 sofar        then use those radiuses to find entities and blocks
21:44 sofar        a fire cracker would be tnt.power = 0.01
21:44 red-001      I see your point
21:45 sofar        a handgrenade would be tnt.power = 0.5
21:45 sofar        a stick of C4 would be tnt.power = 5
21:45 sofar        only one value ever needed to calculate everything
21:45 red-001      would a mod be able to change radius or damage radius?
21:46 red-001      or not?
21:46 sofar        by changing power, yes
21:47 paramat      yes tnt power is linked to radius so we only ever need 1 value, might as well use radius as this is already used
21:48 sofar        plus, physics and math are fun when they're used correctly in games
21:48 red-001      no as in change radius independent from damage radius to for example create admin tnt that only damages blocks
21:48 red-001      after all why not?
21:49 sofar        because math
21:49 twoelk       could an explosion that does not kill a player maybe push him?
21:50 sofar        but adding resistance values to blocks would make sense
21:50 sofar        twoelk: code is supposed to already do that (but bugged, badly)
21:50 sofar        I intend to fix that
21:52 red-001      decrees modablity  for the sake of realism?
21:53 red-001      punching trees is a thing
21:54 twoelk       regarding admin tnt, I wonder wether you could write a list of blocks on paper, add that to an admin tnt block and have it only destroy those blocks listed
21:54 sofar        decrees modablity? what does that mean?
21:54 red-001      make it harder to mod
21:55 sofar        on the contrary, it would make new tnt types easier to add
21:55 twoelk       punching trees is the entry portal to this class of video games :-D
21:55 sofar        since there's now a simple way to make a bigger or smaller blast
21:56 red-001      I said to make it possibly to override the vaule
21:56 sofar        just give a different power value
21:57 sofar        I don't get your point. Do you want a bomb that incinerates everything in 5x the radius but never kills players?
21:57 red-001      If i want to yes
21:57 sofar        that's not a tnt bomb then
21:58 sofar        should be a different weapon imho
21:58 sofar        maybe add a damage-scale value
21:58 sofar        tnt would have 100
21:58 sofar        and others can override it
21:59 red-001      that would be ok
21:59 sofar        you could make a lethal thing by setting power to 0.1 (unlikely to break a block) but  damage scale of 1000
21:59 red-001      yes thats my idea
22:00 sofar        and I want to fly with my arms on earth
22:00 sofar        that only happens in my dreams, though :P
22:00 * red-001    /grant sofar fly
22:00 red-001      ;)
22:02 paramat      KISS
22:02 sofar        paramat: you deserve one
22:02 paramat      lol
22:03 sofar        looks like 778 is a go-ahead
22:07 paramat      duane and i are working on developing mgvalleys and since hmmmm is not active i would like to have authority to merge PRs on this for some progress, i hope other devs are okay with this
22:07 sofar        sfan5: can you close #3626 quickly please?
22:07 ShadowBot    https://github.com/minetest/minetest/issues/3626 -- Update license year by pra85
22:08 sofar        sfan5: ty
22:09 paramat      ah nicely caught
22:09 sofar        too obvious
22:09 sofar        something I could have come up with ;)
22:09 paramat      yeah i got a weird vibe from that too
22:11 est31        paramat, can you ask hmmmm whether its okay with him
22:12 est31        he's the only other mapgen guy, and if you fuck up, I dont really care
22:12 est31        but he will
22:12 paramat      sure
22:13 est31        mhhh I could have put it more politely...
22:13 paramat      no problem :)
22:14 paramat      i'll continue to ask for 2nd approvals, but might sometimes act on my own, but will be clear and document what i do
22:14 est31        as long as its _only_ about mapgen, its fine with me
22:15 est31        emerge manager stuff is no mapgen already for example
22:15 est31        (for me)
22:15 paramat      ok
22:19 paramat      i actually dislike approving on my own so will be cautious
22:20 paramat      anyway to get this moving i'll probably merge #3590 later
22:20 ShadowBot    https://github.com/minetest/minetest/issues/3590 -- Add V5 caves to Valleys mapgen. by duane-r
22:22 paramat      just 1 step in ongoing development, will be refined for sure
22:23 paramat      .. in later PRs
22:23 est31        okay, what to do with 3534
22:24 est31        Well, I'll do what's allowed, and merge SN's implementation.
22:25 est31        if gregorycu has a problem with it, e.g. its still slower than his implementation, he can say something.
22:25 est31        ("allowed" meaning that I have SN's and my +1)
22:25 est31        although I'm not sure about his
22:26 est31        ShadowNinja, do you want to merge this: https://github.com/ShadowNinja/minetest/commit/9a3bf97ce0137743640b27fad8208d11fe56b4ea
22:57 est31        paramat, can you have a look at https://github.com/est31/minetest/commit/860d70bd0e228ee542e3e559961bfc7e56888d77
23:20 paramat      sure i'll look
23:21 est31        To give you context, its about: https://github.com/minetest/minetest/issues/3613
23:24 paramat      and if SN's implementation of 3534 tests to be good then do go ahead with that
23:29 paramat      okay your fix for 3613 seems reasonable to me +1
23:33 paramat      merging 3590 will be delayed, it needs improvement