Time  Nick            Message
02:35 paramat         VanessaE, you might be interested to try this branch https://github.com/paramat/minetest/tree/novergen and try moretrees in mgv7. i may have found the cause of the shadow bugs in mgv7/v5/flat/fractal
04:57 paramat         .. documented here https://github.com/minetest/minetest/issues/3421
05:00 VanessaE        paramat:  I really REALLY wish you'd stop doing that :)
12:40 hmmmm           !!!
12:40 hmmmm           Zeno quit for good
12:42 hmmmm           Lol, minetest needs "more" code comments
12:42 hmmmm           Please if you're going to do that, fork the project and call it minetest-super-annotated or minetest-obnoxious-newbie-edition
12:43 hmmmm           hrmm, paramat, I see you added documentation for biomes... so does this mean the biome API is stable?
16:31 Calinou         hmmmm, why not? Minetest got quite sparsely commented over the years.
16:31 Calinou         it was more commented when only c55 worked on it, I guess
16:31 Calinou         please don't turn Minetest into another Cube 2 :)
16:31 Calinou         whose only 3% of lines are comments
16:36 celeron55       some parts might have the issue for sure
16:37 rubenwardy      good commenting, not more commenting
16:37 celeron55       those parts need to be addressed separately and not by talking about the whole codebase
17:16 rubenwardy      Wed Aug 29, 2012 4:54 pm
17:16 rubenwardy      Thu Dec 03, 2015 5:39 pm
17:17 rubenwardy      oops
19:02 est31           minetest code is commented well
19:02 est31           also, C++ has the concept of variable and method names
19:03 est31           if you set m_sun->is_shining to false then you dont need a comment explaining that it became night
19:03 est31           otherwise, if one has questions, one can look at google as c55 said, or if its minetest specific there is also grep -rIn
19:03 est31           and git blame
19:04 est31           so lots of ways to find out about something
19:04 est31           even asking here
19:06 asl97           est31: (#3417) i hope that what you mean by move the constructor from player.h to player.cpp and override behaviour of the superclass.
19:12 est31           code looks good now
19:12 est31           thanks asl97
19:13 est31           asl97, have you tested that it fixes the bug?
19:29 asl97           i did do a build and it seem to fix the bug but i should do clean build and make sure everything works
19:30 est31           nice
19:30 est31           well idk does sb want to look at it?
19:30 est31           or can I merge it with one approval
19:31 paramat         nore sfan5 or anyone, please can someone check/approve the implementation of https://github.com/minetest/minetest/pull/2827 'Allow overriding the hand'? seems an essential feature long overdue
19:33 thePalindrome   Oh wait, I can approve things! I'll checkem out
19:33 asl97           since when did thePalindrome became a core dev?
19:33 est31           he isnt
19:33 thePalindrome   I thought it was contributor+ that did that?
19:33 * thePalindrome checks documentation again
19:34 est31           but if there is a mistake, and he finds it, and it gets fixed, everybody is thankful :)
19:35 Krock           thePalindrome is not a palindrome.
19:35 thePalindrome   aye
19:36 RealBadAngel    hi
19:36 est31           hello
19:37 paramat         also any comments appreciated on https://github.com/minetest/minetest/pull/3418 'Mapgen: Add propagate_shadow bool to calcLighting'
19:37 RealBadAngel    esr31, fixed compability problems with shader thingy
19:38 RealBadAngel    CF version 7 is moved to old ones
19:38 thePalindrome   Eh, I guess I'm just crazy and mis-read something
19:42 RealBadAngel    paramat,  if light propagation is such a problem why just cut-off those calcuations for defined stacked realm height?
19:42 RealBadAngel    *why not
19:42 asl97           est31: i also did check to make sure that it was overriden on the server by adding a `actionstream<<player_max_speed<<std::endl;`in content_sao.cpp (checkMovementCheat), no idea how to check if it done anything on the client though
19:43 est31           it shouldn't change anything on the client
19:44 RealBadAngel    paramat, this way you wont bother bout a setting and propably calcuations could be made faster
19:58 paramat         because light calculations are important and should be active in all chunks, otherwise stuff looks like the math mapgen mandelbox, fullbright
20:00 hmmmm           paramat, are you sure that piece of code works as-intended?  the way it is now, you propogate sunlight even if the overtop has a shadow, but you don't stop from propogating shadows the rest of the way down after hitting a solid node
20:01 paramat         yes that's intentional
20:02 paramat         so that any structure just under the cutoff y has proper shadows
20:02 hmmmm           yeah but that causes an ugly line where shadows just stop
20:02 paramat         yes
20:03 paramat         but we have that anyway when generating upwards
20:03 est31           I do not think its the right way to do it
20:03 hmmmm           yeah you have a point
20:03 est31           have sunlight nodes instead
20:03 hmmmm           any way we look at it, lighting is broken
20:03 paramat         cutoff y will be chosen by the mapgen to occur in what is usually empty space, so the line is rarely seen
20:04 hmmmm           this makes it less broken for a certain case
20:04 paramat         oh yeah i thought of sunlight nodes :)
20:04 hmmmm           I think I know what I'm going to work on next...
20:04 hmmmm           this lighting model NEEDS to die
20:05 hmmmm           all of these crappy problems can be fixed by adding a single dynamic light for the sun
20:05 hmmmm           only problem is, for people with GPUs that don't support dynamic lights, they'll be stuck with the old lighting model
20:06 paramat         nice. i'm used to the buggy lighting it doesn't bother me much
20:06 est31           <hmmmm> all of these crappy problems can be fixed by adding a single dynamic light for the sun
20:06 est31           wrong
20:06 est31           take large caves
20:06 hmmmm           it'll fix sunlight propogation problems
20:07 hmmmm           it is only one half of the solution though
20:07 paramat         hmmmmm see https://github.com/minetest/minetest/issues/3421
20:08 est31           nice work
20:08 paramat         mgv5/v7 biome/decoration API and biome system are all very nearly stable, but i hate having to decide something is stable :P something may come up that needs changing
20:09 paramat         like recently i made caves much more dense, after feedback
20:11 paramat         hopefully 0 up 2 down overgen will also fix shadow bugs caused by schematic placement and use of non-mapgen object voxelmanip, since those also use 'updateLighting'
20:12 paramat         still needs lots of work and testing though
20:13 paramat         so yes it was my fault all along =}
20:18 paramat         RBA >"cut-off those calcuations for defined stacked realm height?" maybe i misunderstood, the mapgen itself will define the cutoff y. later if we have a realms API that can set the cutoff as you suggest
20:19 paramat         but a core or lua mapgen itself may have multiple layers itself so this will also be needed
20:19 paramat         hmmmmm i think multi-level worlds are a priority over lighting
20:19 RealBadAngel    paramat, i was think about realms as layers, 256 or 512 height, so limiting such calculations could be trivial
20:21 paramat         well a realms API could have varying realm thicknesses, but yes it would auto-set cutoff to just below a realm border
20:21 paramat         .. maybe 128 nodes below whatever
20:23 RealBadAngel    this you will have shadows from layer above just on the clouds of the one below ;)
20:23 paramat         i won't wait for realm support though i might create a multi-level core mapgen
20:24 paramat         well clouds will be much lower, i was thinking 1000-1500 nodes of air between realms
20:25 RealBadAngel    what for? who would ever need that space?
20:25 paramat         for kickass mountains, as in my rivergen branch
20:26 RealBadAngel    thats a point to define layer height
20:26 paramat         mgv7 mountains can already hit 500
20:29 paramat         this commit allows existing mapgens to add their own floatland realms, if wanted
20:29 paramat         i already have code for mgv7 floatlands but have been feeling 'meh' about it :(
20:29 est31           realm height should be something configured per map not per mod setup
20:29 paramat         yes
20:30 est31           lazy travis buildbot
20:31 est31           https://github.com/minetest/minetest/pull/3422
20:32 RealBadAngel    https://github.com/minetest/minetest/pull/3391
20:32 RealBadAngel    so what about this one?
20:35 paramat         what about est's issue in second to last comment?
20:37 est31           RealBadAngel, why can't you just add a simple if case?
20:37 est31           duplicating all the code is total bloat
20:37 RealBadAngel    all other versions are made this way, just moved there
20:37 est31           I mean you only make small changes, no?
20:38 RealBadAngel    yes
20:39 RealBadAngel    but motice ive moved to v8 not only my changes
20:39 RealBadAngel    all the other ones waitin in line too
20:39 est31           A simple if case, around these lines: https://github.com/minetest/minetest/pull/3391/files#diff-70868aa6d6b96c0c1623c761500d23c4R399
20:39 est31           done
20:40 est31           duplicating code makes writing backwards compatible code hard
20:40 RealBadAngel    not rly, old protocols are stored in serliaze/deserialize old
20:40 RealBadAngel    you dont care about them anymore
20:41 RealBadAngel    current one is not bloated with compability code
20:41 est31           <RealBadAngel> you dont care about them anymore <--- perhaps you don't, but I still do care.
20:41 RealBadAngel    i meant, you dont care about them in current serialize routines
20:41 RealBadAngel    for that you have Old
20:42 RealBadAngel    thats not my invention, i added new version and followed the way already used
20:42 est31           RealBadAngel, Old is only for stuff thats really old and in a way where simple if cases wont help
20:43 RealBadAngel    6 is older than 7 and is in old
20:43 RealBadAngel    so 7 too
20:43 est31           only if that person has done it this completely ugly way doesnt mean you have to do it too :)
20:43 * est31         blames the code
20:44 RealBadAngel    i can do that your way, but then current contsruct will be broken
20:44 RealBadAngel    it is already so that 7 doesnt mean compability 100%
20:44 est31           yea lets "break" it.
20:44 RealBadAngel    theres 7.0, 7.1. etc
20:45 RealBadAngel    throw, catch..
20:45 RealBadAngel    i moved it to old because of protocol v27
20:45 RealBadAngel    tile def has changed too
20:46 RealBadAngel    thought its a good ocasion to move things out of throw-catch and make clean v8
20:46 est31           yes no problem with tht
20:46 est31           that*
20:47 est31           just dont implement backwards compat the way that you copy the whole damn list.
20:49 est31           man, great, travis is broken, this means I can't check that PR.
20:49 RealBadAngel    lately is more broken than operational
20:51 RealBadAngel    est31, so you want cases for another, future protocols inside current serialize/deserialize routines, right?
20:51 est31           RealBadAngel, as long as the change isnt too intrusive yes.
20:52 est31           then stuff can be moved to the old version of the routine
20:52 paramat         hmmmm see this lua mapgen that uses slopes, new screenshots https://forum.minetest.net/viewtopic.php?p=200541#p200541
20:53 RealBadAngel    est31, ok, i will redo compability commit then
20:54 RealBadAngel    https://github.com/minetest/minetest/pull/3391/files#diff-70868aa6d6b96c0c1623c761500d23c4R340
20:55 RealBadAngel    that looks i will have to redesign lotsa stuff in here
20:55 est31           Just keep the PR focused on what it does
20:55 est31           you can make a separate PR with the redesign
20:56 RealBadAngel    im focused on shaders and gfx
20:56 RealBadAngel    and yet more it takes to care about ancient clients ;)
20:57 RealBadAngel    which was not my goal at all
20:59 rubenwardy      as a server owner, I'd like 0.4.12+ to be able to connect, I don't care about older versions.
21:00 est31           RealBadAngel, sorry that cant really discuss about these graphics changes.
21:00 est31           that I*
21:00 est31           bc I dont know about gfx alot
21:00 rubenwardy      however other server owners may want to support older if they use less new things
21:00 RealBadAngel    youre blind or something?
21:01 RealBadAngel    :)
21:02 RealBadAngel    gfx is about how it looks i think ;)
21:02 rubenwardy      Actually, I'd like 0.4.10+
21:02 rubenwardy      http://minetest.rubenwardy.com/stats/
21:02 rubenwardy      we should persuade the old clients to upgrade, though
21:03 RealBadAngel    https://github.com/minetest/minetest/pull/3391/files#diff-70868aa6d6b96c0c1623c761500d23c4R1306
21:03 RealBadAngel    this is one of the reasons to really move to 8
21:03 RealBadAngel    6 is sent there twice
21:03 RealBadAngel    once for tiles, not taken care of even
21:04 RealBadAngel    second for special tiles
21:04 Fixer           we need at least some kind of non-intrusive message that will ask to update to a newer version
21:04 rubenwardy      https://gist.github.com/rubenwardy/9859d09c2ff1748e64e5
21:04 RealBadAngel    2 reasons, 2 bytes to move that code to old ;)
21:05 rubenwardy      !title
21:05 rubenwardy      ~title
21:05 rubenwardy      ok...
21:05 ShadowBot       rubenwardy: Error: The command "title" is available in the Format and Web plugins.  Please specify the plugin whose command you wish to call by using its name as a command before "title".
21:05 ShadowBot       rubenwardy: Error: The command "title" is available in the Format and Web plugins.  Please specify the plugin whose command you wish to call by using its name as a command before "title".
21:05 rubenwardy      ShadowBot sucks
21:05 rubenwardy      "Send chat messages to old clients serverpackethandler.cpp.diff"
21:05 rubenwardy      ~web title
21:05 ShadowBot       rubenwardy: web title [--no-filter] <url>
21:06 rubenwardy      that shouldn't really be included in core, as it's not good for forks
21:07 RealBadAngel    est31, anyway, i will change that code tommorow, going to sleep now
21:08 RealBadAngel    atm its maybe not nice, but its working and folks can test the shaders
21:29 Robert_Zenz     Any chance that I can get another pair of eyes on #3358, please?
21:44 rubenwardy      Is m_lag in Server a reliable indicator of how much the server is struggling?
21:44 est31           pushing in 10 mins https://github.com/est31/minetest/commit/d8975eabf95fac1881f75ab149d0cebeff3d10f3
21:53 nrzkt2          rubenwardy i did exactly same thing since 0.4.13, except i detect android version and send to players the right app link :)
21:54 rubenwardy      you were the motivation of that change
21:54 nrzkt2          rubenwardy, for m_lag it depends