Time  Nick          Message
00:01 hmmmm         how does regular glass overcome this problem?  (note:  at this point the only thing i'm doing is using EMT_TRANSPARENT_ALPHA_CHANNEL and drawing it with glasslike)
00:03 hmmmm         http://ompldr.org/vaTZ1aw
00:17 hmmmm         erm, it doesn't, nevermind
00:18 hmmmm         from what i understand, irrlicht does have transparency sorting but it's just broken
00:19 kaeza         how old is Irrlicht? because it seems pretty bugged for a project so widespread
00:19 hmmmm         *shrug*
00:20 kaeza         it just strikes me odd
00:21 kaeza         well, it's no surprise when MT is about 3yrs old and there are tons of bugs lol
00:23 hmmmm         note: EMT_TRANSPARENT_ADD_COLOR needs to be used for mobs shooting things, it looks awesome
00:30 hmmmm         ah, irrlicht does do transparency sorting, but only per-node, but the entire minetest world is a node
00:30 hmmmm         well crap this is dumb
00:31 hmmmm         how do we fix this.... what if we put all transparent nodes (water, this new node type, clouds, etc.) into their own separate irrlicht scene nodes
00:32 hmmmm         obviously this wouldn't work too well if someone were to mix and match transparent nodes with solid nodes
00:33 hmmmm         am i mistaken about this?  should irrlicht be able to do transparency sorting with varius materials on a single node
00:37 kaeza         https://github.com/minetest/minetest/pull/679
01:02 hmmmm         you know this might be a dumb suggestion but... doesn't metology use hardware lighting?  what if we just go and check out what they're doing for light occlusion?  i know they're not the greatest, but if they have this sorted out, i'd like to see how
01:05 kaeza         https://dl.dropboxusercontent.com/u/100008207/screenshot_doomtest.jpg
01:05 kaeza         :D
01:07 hmmmm         erm, not metology, terasology*
01:07 hmmmm         the hell is with the *ologies
01:07 hmmmm         hah
01:07 hmmmm         i assume you did that with the hud api
01:08 kaeza         yep :P
01:09 kaeza         with those changes much more flexible layout is possible
01:10 kaeza         you can even fake "wielditems" by using an overlay and fully transparent (i.e. empty) wield images
01:36 RealBadAngel  thats why we are having problems with too many vertices/indices sometimes. i guess mt way is an abuse of irrlicht. in their forums ive read an answer about this limit
01:36 RealBadAngel  "who would need so many indices/vertices for a single model??"
01:37 RealBadAngel  by now it is like we are displayin world same way as other irrlicht apps single mob or player model for example
02:16 hmmmm         let's just try hiding transparent faces that are behind other transparent faces
02:24 VanessaE      hi all.
02:24 VanessaE      hmmmm: what's news?  figured out the alpha channel issue?
02:25 hmmmm         i guess you can use it if you accept that it won't look right around water
02:25 VanessaE      (I saw the screeshot)
02:25 VanessaE      +n
02:25 VanessaE      not right, how?
02:25 hmmmm         thought you saw the screenshot
02:26 VanessaE      http://ompldr.org/vaTZ1aA\
02:26 VanessaE      http://ompldr.org/vaTZ1aA
02:26 VanessaE      this one, yes.
02:26 hmmmm         this http://ompldr.org/vaTZ1aw
02:26 VanessaE      ah, didn't see that one
02:26 VanessaE      that could be an issue :)
02:27 VanessaE      same thing that causes clouds to appear in front of water fountains and the like, I guess
02:27 hmmmm         yes, except this is much more irritating
02:27 VanessaE      lol
02:27 hmmmm         and since this entire thing was to be for ice, that make it pretty much unsuitable
02:27 VanessaE      rigt
02:28 VanessaE      I bet you'll figure it out though, given adequate time
02:28 hmmmm         [10:16 PM] <hmmmm> let's just try hiding transparent faces that are behind other transparent faces
02:28 VanessaE      right
02:30 RealBadAngel  http://i.imgur.com/t2HpgHT.jpg
02:30 RealBadAngel  this is where i am at the moment
02:30 RealBadAngel  im pretty done with it, just a single issue to solve
02:31 VanessaE      beautiful
02:32 VanessaE      *looks at screenshot again*...  what issue?
02:33 RealBadAngel  you cant see it there ;) sometimes, dunno why on the edges of the block check for neighbours doesnt see glasslike nodes
02:34 VanessaE      oh
02:34 RealBadAngel  then you need to punch it to update and form correct connections
02:35 RealBadAngel  im not sure what causing it, but its a really weird behaviour, when you place nodes from left to right it wont happen
02:35 RealBadAngel  if in reverse order, on block edge next wont see previous node
02:36 RealBadAngel  but if you punch it, it will recheck and look ok
02:38 RealBadAngel  http://i.imgur.com/NZUNRkE.png
02:39 RealBadAngel  this is huge glass surface, right after reload game
02:39 VanessaE      weird.
02:39 RealBadAngel  see artifacts?
02:39 VanessaE      yes
02:39 RealBadAngel  after some time they all will disappear
02:39 RealBadAngel  or when you come closer and punch it to force update
02:40 RealBadAngel  this is not a bug in my code
02:40 VanessaE      as far as you know ;)
02:40 RealBadAngel  n2 = data->m_vmanip.getNodeNoEx(n2p);
02:40 RealBadAngel  if (current == n2.getContent())
02:40 RealBadAngel  detNodeNoEx seems to fail sometimes
02:40 RealBadAngel  *get
02:41 hmmmm         by fail you mean return CONTENT_IGNORE?
02:41 hmmmm         or just return something wrong
02:41 RealBadAngel  this happens when blocks are fully loaded
02:41 RealBadAngel  i can reproduce this bug when placing nodes
02:41 hmmmm         that's a pretty huge bug if that's the case
02:42 RealBadAngel  im placing a line of nodes one by one
02:42 RealBadAngel  when im crossing block border this function doesnt see the neighbour
02:43 RealBadAngel  but in certain direction only
02:43 RealBadAngel  if i perform check from left to right it works
02:43 RealBadAngel  from right to left dont
02:46 RealBadAngel  http://i.imgur.com/KwYJEdv.png
02:46 RealBadAngel  i just placed 2 nodes inside to form connections, this one selected and one above
02:47 RealBadAngel  and they dont see neighbours to the left
02:47 RealBadAngel  however if i do punch the nodes to the left, they will see new ones
02:47 RealBadAngel  and connections will be fixed
02:48 VanessaE      well that means you could do the hacky thing and simulate a punch after each node is placed ;-)
02:48 RealBadAngel  left on the screenshot is x+, right is x-
02:48 VanessaE      but that's a horrible solution
02:48 RealBadAngel  its worse
02:48 RealBadAngel  punching the placed node wont help
02:49 ShadowNinja   RealBadAngel: normal glass has that issue too
02:49 RealBadAngel  you have to punch node next to it i mean in other block
02:49 RealBadAngel  the faulty connection is at the border of blocks
02:50 RealBadAngel  so, to the point, gedNodeNoEx is bugged imho
02:50 VanessaE      hey Mr. Literal, the ";-)" usually means "I'm not being 100% serious" :-)
02:50 RealBadAngel  this cannot be because block is not loaded or something
02:50 RealBadAngel  it surely is loaded
02:56 RealBadAngel  btw it will fix itself after some time
02:57 RealBadAngel  looks like getNodeNoEx is workin on kinda cache
02:57 RealBadAngel  and its got updated or somethin
02:57 hmmmm         by the way you do realize that a block can be blank if loaded with emergeBlock() anymore, right?
02:57 hmmmm         s/anymore//
02:58 hmmmm         that might be what's tripping you up
02:58 hmmmm         i changed it that way so nothing else besides the emergethread can generate map
02:59 RealBadAngel  it updated now after like 1 minute i placed the blocks
02:59 RealBadAngel  (nodes)
03:00 RealBadAngel  notice this wont happen if i place new nodes 1 to the left
03:00 RealBadAngel  they will always see neighbour in x- direction
03:00 RealBadAngel  and bug wont pop up
03:01 RealBadAngel  it happens only in + direction, imho on all axises
03:03 RealBadAngel  i will check it later today when i will be back from work
03:03 RealBadAngel  now have to go, cya later
03:19 VanessaE      bbl.  sleep.
04:02 kaeza         what is VoxelManipulator? some kind of "double-buffered system" for the world?
04:02 kaeza         just being curious
04:05 ShadowNinja   kaeza: It is a faster way of modifying the map
04:06 kaeza         ah I see that
04:15 Pentium44     ohh hi ShadowNinja
04:16 ShadowNinja   Hello Pentium44.
04:16 Pentium44     you have you been?
04:16 Pentium44     its madchicken13
04:16 Pentium44     or WiredOnLinux, or M13, or UbuntuNerd too
04:17 Pentium44     same person
04:19 ShadowNinja   Pentium44: huh, are you saying I am that person? I am not and this should be in #minetest.
04:20 Pentium44     sorry, and no 'I' am that person
04:20 Pentium44     that is what I meant
04:20 ShadowNinja   OK, well better stop chatting here.
04:21 hmmmm         hrm
04:22 hmmmm         how would you guys like to tell minetest to use the alpha channel of a texture?
04:22 hmmmm         would you set alpha to some value below 255?
04:22 hmmmm         or would you rather have a new boolean parameter, use_alpha (this would require a change in the ContentFeatures version)
04:24 ShadowNinja   hmmmm: Is there an advantage to the second version (if I am understanding you correctly)?
04:24 hmmmm         well it's not as hacky, and what happens is clearer.
04:27 ShadowNinja   Well I think you want Minetest to use a transparent texture if it is one, I can't think of any case where you wouldn't.
04:27 hmmmm         you think i would want that, but in fact i don't
04:28 ShadowNinja   Why not?
04:30 hmmmm         because the drawing mode used for this type of material is quite slow
04:31 ShadowNinja   Well that is what I was asking when I asked about the disadvantades ;-)
04:32 ShadowNinja   So add the uses_alpha or similar option
04:33 hmmmm         you must've understood me ...
04:34 VanessaE      hmmmm: re: alpha, it should be based on whether the node's textures have an alpha channel at all, imho
04:34 VanessaE      rather than an explicit flag
04:34 hmmmm         and i just explained why that's not happening
04:34 VanessaE      sorry, guess I missed it
04:36 VanessaE      *reads* again
04:36 VanessaE      ok, so a flag then.
04:36 hmmmm         and then i said adding a new flag would require a contentfeatures version change
04:36 VanessaE      saw that.
04:37 hmmmm         meh, i'll just add a flag in lua but it won't actually be part of contentfeatures.
04:38 hmmmm         and it'll just modify the alpha parameter on its own
04:40 VanessaE      now the real question:  can this be used arbitrarily e.g. with nodeboxes?  or strictly cubes?
04:40 hmmmm         can glass be used with nodeboxes?
04:41 VanessaE      well the texture certainly can ;)  but I mean for example, in the "fancy" door in homedecor.  The glass insert there should have been translucent, were it possible.
04:42 hmmmm         i remember seeing a glass tube nodebox for something
04:42 hmmmm         that's not glasslike?
04:43 VanessaE      that was probably pipeworks pneumatic tube.
04:43 VanessaE      and no, it ain't.
04:45 VanessaE      (that is, however, another example that would look beautiful with skilled use of alpha)
04:53 hmmmm         you might be able to use it without glasslike
04:53 hmmmm         you just need to make sure that the backface isn't culled
04:53 hmmmm         erm that's stupid
04:53 hmmmm         that comes automatically with nodebox
04:54 hmmmm         nvm, so yeah, you can use this with nodeboxes
05:04 VanessaE      seet
05:04 VanessaE      sweet, too.
05:38 hmmmm         i still have to make the shader for this
05:39 VanessaE      oh that'll be real fun
05:40 VanessaE      I don't think I'd worry about that, honestly.
05:40 VanessaE      more pressing is the need for smooth lighting across nodeboxes/2d objects
05:40 VanessaE      (frankly, the flat lighting among smooth-lighted objects makes those things look like total shit in some situations. :-/ )
05:42 VanessaE      (issues https://github.com/minetest/minetest/issues/104 and https://github.com/minetest/minetest/issues/125 )
05:46 khonkhortisan Quothe the raven, 404
05:48 VanessaE      khonkhortisan: 2d objects issue (look at the ceiling), http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/homedecor1.png
05:52 VanessaE      nothing handy that shows the nodebox issue but we've all seen that one.  same effect, but in 3d.
05:52 kaeza         there's one in the slabrealm mod
05:54 VanessaE      http://i49.tinypic.com/2v8oljl.jpg
05:55 kaeza         that one
05:55 VanessaE      (the other one is misleading because of the attempted recolored slab textures)
05:57 khonkhortisan it looks topographic
05:57 VanessaE      mmhmm
06:25 celeron55     i see quite little comments in here about https://github.com/minetest/minetest/pull/679 so i assume nobody opposes it
06:25 * celeron55   merges
06:26 VanessaE      merge it.  seen kaeza's joke image? :)
06:33 ShadowNinja   If you feel like merging some pull requests I have some that nobody oposes...
06:35 khonkhortisan repost: http://imgur.com/wqJaBg7
06:36 khonkhortisan I have some that people oppose :)
06:37 ShadowNinja   Me too (but for bad reasons imo).
06:38 khonkhortisan I put the cart before the horse (rail before the cart)
06:43 ShadowNinja   Well I hope they get merged, I had better go now.
06:53 celeron55     umm well
06:54 celeron55     i guess if i merge some small scriptapi stuff, it won't conflict too much with sapier's rework stuff
06:54 celeron55     it might though; i already said him there won't be changes if he gets it done fast
06:55 celeron55     well the hud improvements already conflict with such so i'll just do this
11:03 PilzAdam      RealBadAngel, for your problem with the new glasslike: I had the same problem when fixing new_style_water
11:04 PilzAdam      the solution is to treat CONTENT_IGNORE as a node of the same kind
11:04 PilzAdam      so all glasslike nodes "connect" to CONTENT_IGNORE
11:24 kahrl         that may be an improvement but I wouldn't call it a solution
11:25 kahrl         a solution (although expensive) would be to regenerate the meshes of all connected mapblocks when a block is received
11:25 Calinou       kahrl is back ö_ö
11:26 kahrl         I've been for a while :P
11:26 Calinou       +1 for regenerating mapblock mesh when there's glasslike nodes at the edge
11:26 PilzAdam      kahrl, yea, but at least for water that works quite good
11:27 Calinou       water doesn't suffer from that problem too, why?
11:27 Calinou       does it regenerate mesh like I said?
11:28 kahrl         there also used to be edges with the smooth lighting depending on the order of block loading, because of CONTENT_IGNORE. But I haven't seen those in recent versions
11:30 Calinou       while we're talking at rendering issues, there are two bugs related to nodeboxes and other "transparent" drawtypes:
11:30 Calinou       - any "transparent" drawtype is drawn 1 light level darker than it should be. as a result of this, I'm forced to use sunlight_propagates for stairs+ nodes (in moreblocks now) so that textures are not darkened
11:30 Calinou       fix: draw them 1 light level brighter, except if they're in light level 0?
11:31 Calinou       - plantlike/signlike/nodebox/glasslike/allfaces/allfaces_optional/fencelike nodes should have sunlight_propagates set to true by default
11:31 Calinou       (airlike too?)
11:44 PilzAdam      kahrl, c55 has fixed the "shadow edges" in 0.4.2 or so
11:55 celeron55     yeah, i added the diagonal neighbors to the mesh generator voxelmanipulator as it turned out it didn't slow it down at all
14:04 bookwar1      celeron55: I unblocked the domain name. As I understand, now you need to ask your registrar to initiate transfer, then they will ask my registrar and then I should approve
14:52 celeron55     bookwar1: replied by e-mail
15:11 sfan5         is dtime in milliseconds?
15:12 celeron55     in minetest almost all times are in seconds; IIRC that isn't any different
15:12 sfan5         ok
15:55 RealBadAngel  hi all
15:55 RealBadAngel  PilzAdam, thx for tip, but this is rather very nasty hack
15:56 PilzAdam      yea, it seems like it works only for water
15:56 RealBadAngel  im pretty sure by now under which certain circumistances this bug occurs
15:56 RealBadAngel  so hopefully i will be able to find the reason
16:24 hmmmm         ah thanks for merging that celeron
16:28 hmmmm         you also added the 'don't prepend ~!~ Server' thing... personally i wanted there to still be some way to differentiate between server and client messages, like slightly different-colored text
16:29 celeron55     i'd think of that more like differentiating what the player writes from other text
16:30 celeron55     (that's how irc clients do it; your nick or lines are in different color)
16:30 Exio          hmmmm: by default the messages will still have -!- server
16:30 PilzAdam      can we change "PM from player:" to ">player<"?
16:31 RealBadAngel  use bold to highlight broadcaster
16:31 hmmmm         can't without making a bold font
16:31 Jordach       cant we just colour it yellow
16:31 celeron55     i'd say just don't touch it anymore
16:31 celeron55     it's good enough
16:31 celeron55     :P
16:32 hmmmm         how about this https://github.com/minetest/minetest/pull/670
16:32 RealBadAngel  so use irc way then
16:32 RealBadAngel  server: blah blah
16:32 RealBadAngel  player: Hello
16:32 celeron55     the mods are responsible for differentiating themselves somehow if they disable the regular server notice text
16:33 celeron55     and i belive they will do so because otherwise anyone will see it doesn't make any sense
16:34 celeron55     hmmmm: how is that not a problem on ARM?
16:34 Jordach       lemme just compile this on arm
16:34 hmmmm         ARM is usually run in little endian mode
16:34 celeron55     (i think minetest has been succesfully built on android, albeit doesn't actually work)
16:34 Jordach       celeron55, correct
16:34 Jordach       Irrlicht just wont :P
16:34 RealBadAngel  again on the problem with getNodeNoEx
16:35 RealBadAngel  celeron55, have you read what i wrote today in the morning about it?
16:36 PilzAdam      hmmmm, why do we need snow and dirt_with_snow?
16:36 hmmmm         because that's what's needed for a snow biome..?
16:36 hmmmm         isn't it?
16:36 PilzAdam      i think dirt_with_snow isnt needed
16:36 PilzAdam      just add a snow layer on top of it
16:36 Jordach       yeah, we have a hot place, but no cold place
16:37 RealBadAngel  should be added with flag?
16:37 hmmmm         but how will people build on the ground with that
16:37 Jordach       PilzAdam, you do realise the grass side needs re-texturing right
16:37 Jordach       hmmmm, buildable_to
16:37 hmmmm         oh okay
16:37 Jordach       snow acts like water or a liquid
16:37 hmmmm         sounds good then
16:37 RealBadAngel  flag snow=1 then dispaly it with snow texture added
16:37 PilzAdam      Jordach, all other nodes need side retexturing too if snow is on them
16:38 RealBadAngel  same way as ores are added to stone
16:39 RealBadAngel  ofc snow will require top, and side textures, top will be replacement, side mixed with original side one
16:40 celeron55     PilzAdam: because dirt_with_snow looks good
16:41 PilzAdam      yea, dirt_with_gras with snow over it doesnt look good
16:41 Jordach       i agree with celeron55
16:41 celeron55     i agree that there could be a better way, but i also think many other ways will be just hackier
16:41 Jordach       anyways, MC has dirt_with_snow
16:41 Jordach       so why are we complaining
16:42 celeron55     Jordach: because we can sometimes do better
16:42 Calinou       MC doesn't
16:42 Jordach       celeron55, we did beat them to modding :P
16:42 hmmmm         hmmmm
16:42 RealBadAngel  celeron55, long story short: consider 2 blocks next to each other. check at pos from node placed in direction x+ to node in block in direction x-, will always see correct content
16:42 Calinou       grass blocks look like they have snow on their sides when there is a snow cover above
16:42 celeron55     we'd need some kind of a group -> grass re-texture definition table
16:42 Calinou       and there are (stackable!) snow covers
16:42 celeron55     and then make the client apply a different grass based on it
16:42 celeron55     but it's kind of arbitrary
16:42 hmmmm         if we can semi-fix the transparency sorting problem and add a colored drawtype, biomes would be in good shape
16:43 RealBadAngel  when you do the very same in differen direction x- to x+, it WILL ALWAYS give you content ignore
16:43 celeron55     RealBadAngel: just fix it, i don't want to think of that
16:43 hmmmm         we are getting snow and all these neat blocks
16:43 hmmmm         it's all coming together
16:43 Jordach       hmmmm, were looking more polished
16:44 RealBadAngel  celeron55, i will try
16:44 RealBadAngel  i really dont like solution PA suggested to treat ignore as same node we would want to have there
16:45 hmmmm         one question though
16:45 hmmmm         is there a way to pass my color-to-blend to the shaders?
16:45 RealBadAngel  propably it will work in this very case, but who knows what it could break
16:46 hmmmm         I figure since i'm mucking around with shaders anyway, it'd probably be a good time to add the color drawtype in too
16:46 celeron55     hmmmm: that's a complicated subject
16:47 hmmmm         well as it stands, if i can't pass my color along to the shader, then this feature just won't happen
16:47 celeron55     currently two light values (IIRC) are passed as the vertex color components R and G, to be mixed by the shaders according to a global which contains the current day cycle brightness
16:48 hmmmm         according to a global
16:48 hmmmm         hmm
16:48 celeron55     so there is B left in that
16:48 Exio          can't you pass that as a "variable"?
16:48 Exio          (to the shader, i mean)
16:48 celeron55     (A is obviously used for actual alpha)
16:48 celeron55     well shaders are kind of limited in where they can get data
16:49 hmmmm         it's probably not going to work out if i make a constant called "blend color" that i change every time i'm shading a pixel from a different node
16:49 Calinou       color... drawtype? shouldn't it be a paramtype?
16:49 celeron55     also i am quite unknowledgeable about shaders, and don't know much about actual modern graphics programming
16:49 hmmmm         paramtype, yeah, whatever
16:49 hmmmm         i don't know anything about graphics programming
16:50 celeron55     i think there must be some way to associate more data per vertex, but that is going to get complicated in some way that i don't have any idea of
16:50 hmmmm         a game is pretty foreign to me... but that's fine because i can pick up on things quickly enough
16:50 hmmmm         people are probably going to hate me for asking this, but....
16:50 celeron55     i suggest just implementing the oldskool implementation first (probably with texture modifications); a shader implementation can be figured out later
16:50 hmmmm         .....how does minecraft do it?
16:51 hmmmm         lol
16:51 hmmmm         this wouldn't be a problem if we had lighting
16:52 celeron55     (even with shaders being used, any oldskool stuff can be in use, so that's the right way to do it anyway)
16:52 hmmmm         oh my god.  if we can just get lighting, all of our problems would disappear
16:52 sfan5         celeron55: you should be able to pass any integer/floating point number to the shader
16:52 hmmmm         mauvebic's whining included
16:52 celeron55     sfan5: yes one global, but one per vertex is a different story
16:53 celeron55     there of course *are* ways to do it
16:53 celeron55     because obviously the complex grass and whatever rendering systems use that much data
16:53 celeron55     modern ones i mean
16:53 * sfan5       guesses there complicated
16:53 hmmmm         erm
16:53 celeron55     but i have no idea how they do it
16:53 sfan5         *they are
16:53 celeron55     somebody needs to read up on this
16:53 hmmmm         kahrl did the current shaders, right?
16:53 sfan5         yes
16:53 hmmmm         i wonder if he'd be willing to help here
16:53 celeron55     i implemented the current shaders
16:53 hmmmm         oh
16:53 Calinou       hmmmm: minecraft uses no shaders at all
16:53 celeron55     kahrl implemented much of the framework
16:53 Calinou       minecraft do what, too?
16:54 * Calinou     knows quite a lot about MC, as a regular player and *cough
16:54 Calinou       MCP user*cough*
16:54 Calinou       (fail enter key)
16:54 hmmmm         well we're figuring out how to colorize individual nodes
16:54 hmmmm         at draw time
16:55 hmmmm         that way we can 1). reduce definitions of all the same node with the only difference being color, and 2). have grass/water/etc. colored differently in different biomes
16:55 hmmmm         and this isn't working with shaders
16:56 hmmmm         (just as a recap to anybody lost in the conversation)
16:56 celeron55     apparently one can bind arrays of miscellaneous structures of data to opengl to be used per-vertex: http://stackoverflow.com/questions/15102593/how-to-pass-per-vertex-attributes-to-glsl-shader
16:56 hmmmm         neato
16:56 celeron55     dunno how new shader support that requires (might be newer than we want to support) and might be confusing to implement in irrlicht
16:57 celeron55     that's to be found out by somebody
16:57 celeron55     (miscellaneous structures = structures containing a miscellaneous amount of named floats :P)
16:58 celeron55     if that isn't possible, then i am guessing at least passing a second array of colors (one per vertex) will be possible with a version-to-be-supported)
16:58 celeron55     -)
16:59 PilzAdam      I have added a function to snow that turns grass under it to dirt_with_snow when placed; now I need to change the ABM that grows grass
17:00 PilzAdam      should it just grow dirt_with_snow if snow is on top of it?
17:00 hmmmm         yeah
17:00 hmmmm         that's good
17:01 RealBadAngel  funny fact: last sunday came into my room 2 little girls, daughter of my sister, and her friend. both aged 5-6yrs
17:01 hmmmm         why is that funny
17:01 RealBadAngel  saw me workin on mt, i was just codin and testing
17:02 RealBadAngel  and those girls said: oh! youre playin minecraft
17:02 RealBadAngel  we do play too!
17:02 hmmmm         hah wow.  they know what minecraft is?
17:03 sfan5         "best comment award goes to hmmmm"
17:03 RealBadAngel  yeah, and they started to explain me what they build in there
17:03 RealBadAngel  5 yrs old kids
17:03 celeron55     humm... this is going to get ugly
17:05 celeron55     quickly looking, it seems irrlicht doesn't have built-in support for such modern shader stuff (as it hasn't been designed for flexible shaders in the first place)
17:06 celeron55     but irrlicht has some more complicated vertex data formats (= contains some more data fields) that can be used
17:07 PilzAdam      should dirt_with_snow turn into something else if the snow above it is dug?
17:07 celeron55     http://irrlicht.sourceforge.net/forum/viewtopic.php?f=4&t=27802
17:07 PilzAdam      and I also gonna add a snowblock, crafted with 9 snow
17:09 celeron55     (eg. this http://irrlicht.sourceforge.net/docu/structirr_1_1video_1_1_s3_d_vertex_tangents.html)
17:09 hmmmm         that's not horrible though
17:09 celeron55     by the way, i hate that new doxygen page format of irrlicht
17:09 celeron55     the old one was like 10x better
17:10 RealBadAngel  about lights: http://www.youtube.com/watch?v=Y7AR7oxLR_4
17:10 celeron55     those frames are like back to 90s
17:10 hmmmm         we'll just use a S3DVertex2TCoords
17:11 celeron55     well, ugly is not always horrible 8)
17:11 hmmmm         it's the best we can do given the restrictions imposed upon us by irrlicht
17:11 Calinou       hmmmm: minecraft never colorizes individual nodes, wool just uses 16 textures like we do
17:11 Calinou       in 1.9pre2 part of the cauldron texture was colorized
17:12 Calinou       and potions are still colorized dynamically, but that's an item
17:12 hmmmm         well
17:12 hmmmm         this has lots of applications
17:12 hmmmm         not just potions and whatever
17:13 hmmmm         potions have to be colorized dynamically though because they're animated
17:13 Calinou       some minecraft packs animate potion textures, too
17:14 Calinou       talking about animation: it'd be nice if instead of definiting in lua code, we used text files, client-side
17:14 celeron55     hmmmm: you do understand that there absolutely needs to be a non-shader implementation of the coloring of whatever too?
17:14 Calinou       eg. default_stone.png, default_stone.txt, it contains the total time of animation
17:14 hmmmm         obviously
17:14 Calinou       (in seconds)
17:14 hmmmm         and that's much simpler as well
17:14 celeron55     good
17:14 Calinou       why can't we use a shader-less implementation regardless of whether shaders are enabled or not? is it faster?
17:14 Calinou       would be simpler
17:14 hmmmm         it's just that i gave to give special attention to the shader case because it's the main complicating factor
17:14 celeron55     a shader one will be useful for those using super-high-resolution textures though, as otherwise it's going to take enormous amounts of texture space 8)
17:15 hmmmm         s/gave/have/
17:16 hmmmm         erm this is probably a stupid question
17:16 hmmmm         but how do i do translucency with shaders?
17:17 hmmmm         gl_Color.a always seems to be 1.0
17:19 hmmmm         erm nevermind, that's just the shader failing to generate...
17:23 hmmmm         :D
17:24 hmmmm         suck it, transparency sorting problem!  http://ompldr.org/vaTc3OA
17:24 PilzAdam      hmmmm, change the red texture aplha to 0
17:25 PilzAdam      can you see to the ground of the sea then?
17:25 hmmmm         probably
17:25 hmmmm         that's hardly a problem though
17:26 PilzAdam      is the alpha set in the texture file or in the alpha field in nodedef?
17:26 hmmmm         texture file
17:26 hmmmm         well, obviously, you won't be able to replace regular glass with this
17:31 hmmmm         nor should you
17:31 hmmmm         (for obvious performance reasons)
17:37 celeron55     there is 100% surely going to be a mod for those who wish to kill their performance with making all grass and all leaves and all everything translucent 8)
17:37 celeron55     grass? ehm, i mean glass
17:38 celeron55     also, actually it doesn't even affect performance
17:39 hmmmm         maybe if you're using shaders it won't
17:39 hmmmm         but EMT_TRANSPARENT_ALPHA_CHANNEL is definitely slower than EMT_TRANSPARENT_ALPHA_CHANNEL_REF
17:45 celeron55     yes, on old hardware, definitely
17:45 Calinou       on today's hardware texture size doesn't even matter
17:45 Calinou       just use the thing that looks best
17:46 celeron55     it matters because of memory consumption
17:46 Calinou       today's graphics cards have 1GB of VRAM in the low end range, 2GB in the mid end
17:46 celeron55     also mesh sizes matter mostly because of memory consumption
17:46 Calinou       good luck filling that, even with 4× AA on 2560×1440
17:46 hmmmm         we are not optimizing for high-end systems, that's against the minetest philosophy
17:47 hmmmm         frankly i don't care about "hardware these days"
17:47 celeron55     how about one thousand pieces of VanessaE's 512x512px textures? 8)
17:47 celeron55     that's 1GB
17:47 Calinou       512× is pointless, 256× is enough, by far :P
17:47 hmmmm         i don't really like that i have such a powerful processor on this computer for testing minetest... but compiling is great (:
17:47 celeron55     hmmmm: in linux there is a way to select how many cores to use for a process 8)
17:48 hmmmm         everything ought to be near-instantaneous on any processor from 2005
17:48 celeron55     setting that to one for MT makes a good difference
17:48 hmmmm         still won't simulate an athlon 64 3500+ very well
17:48 celeron55     not really, today's cores are a couple of times faster
17:49 celeron55     >near-instantaneous on any processor from 2005
17:49 celeron55     any web application these days lags like hell eating 100% CPU on those
17:49 Calinou       architecture is very important, much more important than amount of cores :P
17:49 celeron55     they just rely on the processing power :P
17:49 Calinou       today's quad core intels beat eg. a 990X (some 6 core from intel)
17:50 celeron55     well, a pentium 4 is like a low-end atom these days
17:50 celeron55     that's how it goes
17:50 celeron55     (a nice 10x difference in power consumption)
17:51 celeron55     (well, more like 5x)
17:51 celeron55     actually... more like 20x
17:53 hmmmm         did anybody notice that people do stupid things like "hey imma run a minetest server on my 600mhz laptop"
17:53 hmmmm         they figure minetest shouldn't be intensive at all
17:53 hmmmm         can anybody say with honesty that minetest deserves to be as intensive as it is?
17:53 hmmmm         maybe they are justified
18:08 celeron55     a minetest server will run on a 600MHz laptop
18:08 celeron55     as long as it has RAM in the range of 100MB
18:08 celeron55     i broke a HDD of one of mine by running an MT server on it 8)
18:09 celeron55     poor old thing
18:09 sapier        but if you want to use mods even 1.8ghz won't be enough for server
18:09 sapier        at leastwhile magen is running
18:10 celeron55     well yeah, all of my servers have always been quite vanilla
18:11 sapier        yes vanilla is quite fast ... swamp mod is horrific and snow mod not that much better
18:11 PilzAdam      we have snow now in vanilla :-)
18:12 sapier        snow biomes?
18:12 PilzAdam      no
18:12 PilzAdam      snow nodes
18:12 sapier        hmm maybe I had both :-)
18:13 sapier        does vanilla create sea ice?
18:13 PilzAdam      the snow and ice nodes arent generated yet
18:14 sapier        did anyone look at the scriptapi subfolder changeset yet?
18:14 PilzAdam      yes, its broken
18:14 sapier        a little bit more details plz ;-)
18:15 PilzAdam      https://travis-ci.org/minetest/minetest/jobs/6547231
18:17 sapier        what the hell?
18:18 sapier        wonder why gcc 4.7 doesn't have any problems with this
18:22 celeron55     the standard does say that class should always be "friend class X", functions are "friend <function declaration>"
18:23 sapier        gcc 4.7 doesn't even think this is worth a warning
18:25 celeron55     actually no, it doesn't say so
18:25 celeron55     but
18:25 celeron55     ah, here's the key
18:26 celeron55     there doesn't need to be "class" if the class is in scope
18:26 celeron55     is this related to your case?
18:26 celeron55     hmm, it seems you do have declarations for them
18:27 sapier        hmm class is forward declared not quite sure if this counts as "in scope" ... I've added the class already hopefully this is only 4.6 4.7 incompatibility
18:27 celeron55     it does count
18:28 celeron55     (as per C++98)
18:28 sapier        maybe this is a gcc 4.6 bug ... would explain why 4.7 doesn't even issue a warning
18:28 celeron55     anyhow, we support 4.6 so just fix it 8)
18:28 sapier        already done ;-)
18:31 sapier        does travis build start automaticaly?
18:32 celeron55     yes
18:32 sapier        btw celeron you've been right problem with entity duplication is not exactly beeing saved in different block
18:33 sapier        seems problem was a entity was not staticized while still beeing in use by a client but left active ... noone ever deactivated it
18:36 celeron55     i'm not sure about the correctness of your patch
18:37 sapier        I've been thinking about it too but the more I think about it the more I do think it's correct
18:38 celeron55     IIRC it's intended to continue storing the object as a counterpart of an active object (=having static id) until all clients have been informed it, and when it has happened, then it should finally store it as static-only (id=0)
18:38 sapier        problem is that there is no static data within this block once it's deactivated
18:38 celeron55     the active part is kept in the pending state until all clients know it will be removed
18:38 sapier        while the entity still is active
18:39 sapier        but a remove message is sent to client
18:39 sapier        if this block is activated again there is not static object to be activated --> client doesn't get informed this object needs to be added
18:39 celeron55     i would guess you have tested it enough to see that it does work though
18:40 sapier        I've only tested singleplayer
18:40 sapier        so if there are race conditions with multiple clients I wouldn't have run into it
18:40 celeron55     well then, if somebody cares to do some testing with some mob mod in multiplayer, that would be highly valued
18:41 celeron55     also, the dungeon game of mine always spewed some static object errors
18:41 celeron55     that would be a good test too
18:42 sapier        testing first part of bug is really simple
18:42 sapier        add a cart somewhere run away until this cart is unloaded
18:42 sapier        then run back
18:42 sapier        in current version you'll never get the cart back on client until you log out
18:43 celeron55     i think you actually do sometimes; i was testing that once and it behaved really oddly
18:43 sapier        I tried yesterday I didn't have a single situation where I got it back ...
18:44 sapier        of course that doesn't tell it never happens
18:48 kahrl         yeah, I remember seeing mixed results when I tried the same
18:48 kahrl         (with item objects instead of carts)
18:49 sapier        as far as I understand that shouldn't be any difference
18:49 ssieb         any idea what would cause this: 20:27:38: ERROR[ServerThread]: ServerEnv: Trying to store id=3 statically but block (-2,1,3) already contains 49 (over 49) objects. Forcing delete.
18:49 ssieb         I investigated this once and the location was in a solid area of stone...
18:50 sapier        it's exactly what is written in the message
18:50 sapier        too many objects at one place
18:50 kahrl         object duplication, probably
18:51 ssieb         but nobody had been there and there was nothing there
18:51 ssieb         I teleported into that spot and dug out the stone and there was nothing...
18:52 kahrl         did you dig out the entire 16x16x16 mapblock?
18:52 ssieb         oh!
18:52 ssieb         am I confusing the terms?
18:52 kahrl         block (-2,1,3) is at coordinates (-32,16,48)
18:53 ssieb         ok, now that's makes a lot more sense
18:53 ssieb         there's a limit per mapblock of how many loose items can be around?
18:54 kahrl         yeah, it was added in the 0.3 days when there was a rat plague on the gameboom server
18:54 ssieb         I'll tell the kids that they really need to clean up there mess then. ;-)
18:54 sapier        maybe we should disable this limit, in 0.3 ther haven't been item entities mob entities ...
18:55 ssieb         s/there/their/ :-/
18:55 sapier        now if your inventory is full all nodes are dropped as entities so limit is reached very soon
18:55 ssieb         yes
18:55 ssieb         and TNT!
19:13 celeron55     kahrl: no
19:13 celeron55     it was added when my server got riddled with dungeon master duplication
19:14 celeron55     or misspawning
19:14 celeron55     it's been there waiting for a time when we are confident with object handling actually doing it's job reasonably
19:15 sapier        hmm so if my patches already cought all bugs this could be disabled completely?
19:15 celeron55     there should probably always be some limit, but maybe we can raise it a factor of some after testing sapier's fix
19:15 sapier        i don't wanna tell I already fixed everything
19:15 celeron55     no, it should never be disabled completely
19:15 sapier        of course some limit is reasonable
19:16 celeron55     i don't want anyone to have a broken world because of a broken mod or something
19:16 celeron55     the limit should be in the hundreds
19:16 celeron55     once there are no bugs
19:16 sapier        mods always can break a world
19:16 celeron55     yes but it's too easy to break a world by flood-spawning entities
19:17 sapier        it's even more easy by using an abm to replace nodes ... but I got what you want to tell
19:17 celeron55     it should cap at something that most computers will still handle
19:17 celeron55     sapier: that's hard to get accidentally wrong
19:18 sapier        one typo in replacement node ;-)
19:19 sapier        but back to entity problem once those bugs are fixed player experience with mobs way better than now
19:19 sapier        I've been playing only 3 days now in order to release new mobf version and got duplicated mobs to an extent it wasn't fun anymore
19:52 celeron55     okay so i play minetest a bit after like a month
19:52 celeron55     and i think that this end-of-fall camera jiggling is ridiculous
19:52 celeron55     this is the most ridiculous thing i have seen
19:52 celeron55     what the fuck man
19:53 celeron55     1) it isn't even close to being anywhere properly synced with any momentum, 2) it would be stupid anyway
19:53 StrayBytes    It does get really annoying.
19:53 iqualfragile  PilzAdam wasnt too hapy about that either
19:53 celeron55     who even agreed in putting this in
19:54 celeron55     i'm going to just revert this unless somebody gives any reasonable explanation for this in an hour
19:54 Randy_        Hey PilzAdam Can you add farming as a default? and maybe some animals/mobs? in the next Dev.?
19:55 PilzAdam      ehm
19:55 iqualfragile  Randy_: animals and mobs are not yet stable/performant enought
19:56 Randy_        For like 0.4.7 MineTest maybe then just Farming
19:56 PilzAdam      its Minetest, not MineTest
19:57 Randy_        Same difference. tomato,tamoto
19:57 Jordach       celeron55, ask Taoki
19:57 Jordach       he put it in
19:57 sapier        I'd like to expose collision handling support to lua in order to make mobs more performant what do you think about this idea?
19:57 Taoki         celeron55: It's just an effect for realism, and it's very welcome
19:57 Jordach       bull,
19:57 iqualfragile  Randy_: in minecraft mojang makes the game, in minetest the server admin does, so just add it to your local installation
19:57 celeron55     lol what
19:57 Taoki         Why does every good feature have to be considered wrong in this project
19:57 Jordach       i dont mind it: but it does annoy me slightly
19:58 Taoki         Seriously? A simple camera jiggling when you fall is that stupid?
19:58 sapier        e.g. once collision is detected call to lua telling which entity collided
19:58 celeron55     it's like the player deliberately gently curtseyed after each landing
19:58 Taoki         Maybe we should indeed fork this project if that's what the current admins think...
19:58 Taoki         Who don't want any parts of it to evolve
19:58 celeron55     it's plain stupid :D i don't even have anything else to say about it
19:58 Jordach       realistically, its correct
19:58 Jordach       but as this is not IRL, there is no need for it
19:59 celeron55     it's utterly wrong, it should not work even close to this if it wanted to act like in real life
19:59 Taoki         celeron55: A shocking amount of games (both open and closed source) have such "plain stupid" effects. I guess Minetest needs to be kept at stupidly simplistic level however
19:59 Taoki         Really? Like, every project I saw so far has something like this
19:59 Taoki         FPS or anything else
19:59 Jordach       Taoki, red eclipse does not
19:59 ShadowNinja   Taoki: I like it but falling off even one node activates it, I think it should take a little more
19:59 Taoki         You're the first I see who finds this plain stupid
20:00 celeron55     where's the code to make this disabled by default and enabled by a configuration setting that nobody can find
20:00 Jordach       Taoki, if the force was like falling from 8+ nodes, then yes
20:00 Jordach       it would be nice
20:00 Taoki         But I'm more and more starting to support the idea of a fork. Since the current administration doesn't seem to want anything to evolve properly
20:00 iqualfragile  Taoki: just go and help minetest black
20:00 Taoki         Everything is seen as stupid
20:00 celeron55     >Everything is seen as stupid
20:00 iqualfragile  its basicly minecraft build via minetest
20:00 celeron55     >everything
20:00 Jordach       because it doesnt fit currently.
20:00 celeron55     >EVERYTHING
20:00 celeron55     :D
20:00 celeron55     LOOK AT THE WORD
20:01 Jordach       were going into chit chat territory
20:01 Taoki         celeron55: Also, I wanted to add a setting for it. PilzAdam didn't agree. Blame him that it's enabled only in the code and there's no setting
20:01 Taoki         If you need to blame someone
20:01 Jordach       blame yourself.
20:01 Taoki         celeron55: Unfortunately, that is the vision of the current administration. And things such as "who needs / cares about the user base"
20:01 celeron55     Taoki: EVERYTHING is seen as stupid!
20:01 iqualfragile  Taoki: s/administration/bunch of random blockheads/g
20:01 celeron55     i
20:01 celeron55     lol
20:02 celeron55     where's all sense from any of this
20:02 Taoki         iqualfragile: I simply don't understand why anyone would be so pissed about a mininal camera fall effect, which is sane in any game to keep falling when jumping from looking horrible and fake
20:02 celeron55     everything!
20:02 celeron55     everything is stupid!
20:02 Taoki         celeron55: Anyway, if you must remove it, please add a setting and turn it off by default. Other people like me (stupid ones) would like to be able to enable it, even if manually
20:03 celeron55     i mean, have you ever even glanced at the commit log?
20:03 celeron55     or pull requests
20:03 iqualfragile  Taoki: im sorry, i misstook you for Randy
20:03 sapier        :-) I think EVERYONE got the point ;-P
20:03 Jordach       celeron55, if youre going to be acting crazy, go to minetest
20:03 Taoki         Ah, ok. No I don't use that name
20:03 Jordach       we dont need your kind here!
20:04 Taoki         Jordach: He is the main developer and I appreciate what he did. But I don't agree with how he admins all the time myself... maybe its just me. Then again this isn't my project / fork and that's the most I can say / complain about.
20:04 Randy_        Oh and i can't log on the forum it says im a bot?
20:04 sapier        currently most ppl do only add new features bugs are tracked down by very few ppl ... taoki when did you look for an anoying bug last time?
20:04 Taoki         sapier: I'm all for bug fixing, but it doesn't mean new features aren't good too
20:04 PilzAdam      Randy_, thats offtopic
20:05 sapier        no but new features do mean NEW bugs ... this fixing the old ones will get more and more complicated
20:05 celeron55     so here's a poll: will i either disable this by default or simply revert it
20:05 Taoki         celeron55: Can you add a setting and disable it instead of removing it, so we can still use it optionally? Maybe let it adjust the intensity as well, like view bobbing works
20:06 Taoki         I wanted to do it like that, but PilzAdam said it won't go in if I add a setting for something so baisc. Not to blame anyone, just how things happen. So you might want to ask him if it's ok to just add a setting
20:06 Taoki         **happened
20:07 celeron55     do you understand that i am not blaming anyone, i am blaming this stupid thing
20:07 PilzAdam      if its enabled by default, then a setting is useless
20:07 Taoki         Other than that, I'm starting to wonder if my vision (of maybe a more modern and advanced Minetest) isn't really in match with the artistic style of Minetest. Would be good to know just to be sure
20:07 PilzAdam      but if its disabled by default, then go ahead and add the setting
20:07 PilzAdam      I would be fine with removing also
20:07 celeron55     Taoki: that may be very true
20:07 celeron55     Taoki: but that's more of you to decide, because i have never seen your full vision
20:07 Taoki         celeron55: It's a small new feature that got added. I don't see why it's so stupid and evil and horrible. It's ok if people don't like it otherwise... which is why I suggested a setting
20:08 sapier        PilzAdam I don't understand your logics if anything default doesn't need a setting why add any setting?
20:08 PilzAdam      sapier, because noone would enable it
20:08 sapier        using this logics security is to be added for everyone or noone too ...
20:08 Taoki         celeron55: Don't really have a full vision. Just that I'm hoping Minetest can get to be a very modern engine (even if using blocks for terrain, etc). Modern visual effects, camera effects, hardware lighting, etc. Much of what I'm doing is to lead to that.
20:09 hmmmm         [04:01 PM] <Taoki> celeron55: Also, I wanted to add a setting for it. PilzAdam didn't agree. Blame him that it's enabled only in the code and there's no setting
20:09 hmmmm         a setting
20:09 hmmmm         a setting isn't magically going to reverse the damage that you've done to the infrastructure to support fake colored lights
20:09 celeron55     what
20:10 celeron55     now there is absolutely no sense in here
20:10 Taoki         hmmmm: Firstly, I haven't done anything... just supported the idea of the feature being done. Second, I understand your point on this... like I said I think colored lights would be worth the extra info in mapgen but that's just me
20:10 PilzAdam      why are you changing the topic?
20:10 hmmmm         oh
20:11 celeron55     i don't understand anything anymore, i'll just make this and go to sleep i guess
20:11 hmmmm         sorry, i thought that's what he was on about
20:11 hmmmm         i didn't realize the topic was something less controversial like the fall bobbing
20:11 hmmmm         no, it's not just "extra info in the mapgen"
20:11 Taoki         celeron55: If you wish, make a setting instead of removing. It's your choice though... I'd prefer that though.
20:11 hmmmm         it's "doubling the size of everything in minetest"
20:11 Taoki         hmmmm: Ah no, it's a about a camera effect I added for falling
20:12 hmmmm         i personally don't have any problems with that.
20:12 hmmmm         a setting is always nice though
20:13 hmmmm         [03:53 PM] <celeron55> who even agreed in putting this in   i agreed to putting it in because it didn't seem very harmful for one, and for two, everybody was talking about it as if they didn't have any problems either
20:14 sapier        maybe noone tested it?
20:15 * Taoki       is a bit upsed because I don't understand why a minor feature like that could be so hated for no reason. That and I'm still unclear about other things in vision of the dev team... especially after the who needs / cares about a community mentality I heard here. Or the everything is stupid line c55 mentioned
20:15 Taoki         **upset
20:15 hmmmm         there's that word "everything" again
20:15 hmmmm         i don't think you people know what it means
20:15 sapier        I haven't tested it by now so I can't really tell
20:16 Taoki         Minetest is a great project. But sadly, I no longer feel the positive direction behind it. Maybe it's just me though.
20:18 hmmmm         it really is just you
20:18 Taoki         if you say so
20:18 hmmmm         you and mauvebic in particular try to paint the portrait of the dev team being negative and unreceptive to outside ideas
20:18 PilzAdam      Taoki, c55 just quoted you with the "everything is stupid"
20:18 Taoki         PilzAdam: Ok, misunderstood then
20:19 hmmmm         because we're not allowing a feature to be implemented in the naive way
20:19 Taoki         hmmmm: Maybe it's a bit of how I feel... I don't have a purpose of painting such a thing
20:19 hmmmm         and because someone else has a personal preference to not have fall bobbing enabled
20:19 hmmmm         fucking a
20:19 hmmmm         it certainly is ant season now
20:20 Taoki         The personal preference is ok >_> Was a bit surprised the feature got so hated all of a suggen, when it was just a small part of me trying to optimize and modernize several things on my list
20:20 jin_xi        such things should be configurable. no need to speculate if users will turn it on if someone already made the feature
20:20 Taoki         I have other reasons of my own to be upset too... but those aren't relevant here
20:20 hmmmm         hold on i need to take care of this
20:21 celeron55     done
20:22 kahrl         I don't mind the fall bobbing, but I agree with Jordach, the effect for 1-2 node falls and simple jumps is way too much
20:24 Taoki         Anyway, I don't really plan to upset part of the team as much in the future with features they consider stupid. As ffar as the engine is concerned, I'm most interested in dynamic lighting now. Then effects like reflections, gloss, bloom, parallax / bump mapping, and what else can be done and allowed. My ideas and plans will be more on the LUA side in the future
20:24 Taoki         Though if I can help with the engine too at any time, I like to
20:24 hmmmm         taoki, have you looked at what Terasology does for hardware lighting?
20:25 Taoki         But before coding new features, I'll probably ask first from now on. I spoke about this with PilzAdam, but I can't rely on every main dv telling me if a feature is right or wrong since not everyone is online all the time
20:25 Taoki         hmmmm: I'm not familiar with its code, but I'm totally impressed by its graphics
20:25 PilzAdam      celeron55, care to put a rangelime arround the setting?
20:25 hmmmm         well they seem to have the problems we've discussed figured out already
20:26 celeron55     we have a simple rule: if two in the core team agree on something, it can be merged; but as i did now, it can of course be reverted or disabled by default later
20:26 celeron55     PilzAdam: why?
20:26 Taoki         ok
20:26 PilzAdam      fall_bobbing_amount = 100 <- I just tried that
20:27 Taoki         PilzAdam: That will probably cause the view to go way in the ground :P
20:27 celeron55     well that's user error (or if someone is crazy enough, that may be exactly the wanted thing)
20:27 celeron55     my policy is to not restrict anything unless actually necessary
20:27 hmmmm         also what's up with the mese pick not being able to mine cactuses anymore
20:27 PilzAdam      its illogical to go deeper than the ground
20:28 hmmmm         that's a bit redic.
20:28 kahrl         it could be seen as a cheaty way to find caves
20:28 PilzAdam      hmmmm, you need an axe for this
20:28 Taoki         PilzAdam: Sure. Which is why one shouldn't set the setting too high... maybe beyond 2 or so
20:28 iqualfragile  great… now everybody is unhappy about it not beeing on by default
20:29 celeron55     oh god
20:29 celeron55     i thought making the amount configurable was a good idea
20:29 celeron55     ...
20:29 Taoki         It is a good idea. View bobbing is configurable too
20:29 Taoki         It's how I intended it initially also
20:29 kahrl         nothing is ever easy :P
20:29 celeron55     move on, it's done and perfect now
20:30 * celeron55   moves hands like a police guiding traffic
20:30 Taoki         :)
20:31 celeron55     if somebody wants to limit it, the limit must be done so that it's barely limited to not go through ground
20:31 celeron55     anything else is overly restrictive
20:36 celeron55     i'm open to testing improvements on it, if somebody cares to attempt such
20:36 celeron55     i'm not hopeful at all about that though
20:48 celeron55     anyway, there are at least three problems with it: it does not properly match the fall velocity of the player, the acceleration curve of it doesn't make much sense (it uses sin(); that's not right in any way) and it shouldn't really happen at all unless the player is continuously walking
20:49 celeron55     when testing it when walking without jumping at all, one can kind of see what it tries to do
20:49 celeron55     but it's too wrong to be enjoyable even then
20:56 celeron55     so my "2) it would be stupid anyway" part might be wrong; but i am yet to see
21:04 Exio          :D perfect, now i can haz fall_bobbing_amount = 2.0
21:28 kahrl         ugh, not having any luck whatsoever trying to reproduce the F10 crash
21:28 kahrl         has it been fixed? I haven't seen any mention of that in chat
21:29 kahrl         though I managed to fix an unrelated bug that few people will ever notice, but anyway: https://github.com/kahrl/minetest/commit/104fe65b4155f9341e3e66bdf11b5d1385b3e61a
21:30 kahrl         hope it's okay I fixed a typo in the same commit
21:33 sapier1       is there any scheme when a travis build is run?
21:55 hmmmm         erm
21:55 sapier1       yes?
21:55 * VanessaE    looks at hmmmm
21:55 hmmmm         kahrl, vanessa gave us a backtrace for the F10 bug and the problem seems to be in a call to isspace(), but the string it was passing the character to was wide
21:56 hmmmm         i figured that was probably the problem, but because we can't reproduce it, i didn't touch it
22:08 kahrl         right, that should maybe be iswspace
22:08 kahrl         don't see how it could crash though
22:08 hmmmm         me neither
22:08 hmmmm         but it was a segfault, and it happened in isspace
22:09 kahrl         pretty it's memory corruption that for some reason causes a crash when submitting chat
22:09 kahrl         pretty sure*
22:10 hmmmm         by submitting chat you mean opening up the console, right?  that's why it's called the F10 bug, because you press F10, it instantly crashes (for certain people)
22:10 hmmmm         but i do agree that it must be memory corruption, otherwise there would be no explanation to why the majority of people don't have the problem
22:10 kahrl         oh, no I didn't mean that
22:11 kahrl         because for me when it crashed (years ago) it was always when I pressed enter in the console
22:11 ShadowNinja   There is also a chat-submit bug
22:11 hmmmm         yeah, those are two different things then
22:11 kahrl         possibly caused by the same memory corruption
22:12 ShadowNinja   Also a rare bug, and sometimes the same message won't cause the crash
22:26 hmmmm         ugh
22:26 hmmmm         i need to finish up with what i'm doing
22:31 hmmmm         whatever, i'm done.
22:32 hmmmm         i'm going to stop fiddling around and just commit it.  it does have its problems though, you can pick your poison though by disabling or enabling shaders (i'll solve the problems at some later point in time if it's really that bad)
22:39 hmmmm         https://github.com/kwolekr/minetest/commit/7bd10de45d9a012be53528d2b7c15544989940b2
22:48 ShadowNinja   hmmmm: use_texture_alpha_channel is terribly long
22:51 ShadowNinja   just use_alpha or use_texture_alpha would work fine I think
23:29 hmmmm         yeah but this is self-descriptive