Time  Nick         Message
00:19 VanessaE     hmmmm: what's the story with mgv7 default biomes?
00:20 VanessaE     zat is modifying moretrees to be able to work on mgv7, without depending on plants_lib, but without some kind of reliable default biomes to grow on, well...  we're kinda stuck.
01:31 * zat        hears crickets
02:14 hmmmm        i'd recommend people to not use mapgen v7 for the time being
02:14 hmmmm        my life has been really crazy the past couple of months and i've just not really felt like working on minetest related things
02:14 hmmmm        sorry
02:15 hmmmm        see I'd go and push my better biome stuff right now, but I don't have a decent enough set of default biomes
02:16 * VanessaE   pokes zat
02:18 VanessaE     it's just a question at this point of having something to grow on at all really, even if the form the land takes will eventually change
02:18 VanessaE     you know, default:dirt_with_grass and so forth
02:19 VanessaE     but I understand quite well
02:21 zat          I am working on a set of biomes right now
02:21 zat          hmmmm: is there a way to get higher mountains in v7?
02:21 VanessaE     I suggested to zat that the biomes and the form the land takes should somewhat resemble mapgen v5
02:21 VanessaE     (since people generally liked that one better than v6)
02:22 zat          by the way I have no idea about v5 D:
02:22 VanessaE     zat: look at 0.3.x.
02:22 VanessaE     that's v5
02:23 VanessaE     but maybe it's better to use something like what paramat does with his paragen mod
02:23 VanessaE     but less complex maybe
02:25 hmmmm        zat:  i would hope that mgv7_np_mount_height would be pretty obvious
02:25 hmmmm        "hey, mount, mountain, height, that might be the noise params for mountain height"
02:26 zat          no idea... where do I see that lol
02:34 zat          nevermind I found it all :)
10:35 RealBadAngel #1308 is updated with some style fixes, .env removed and TTL for items added
10:36 ShadowBot    https://github.com/minetest/minetest/issues/1308 -- Item entity stacks merge on the ground, add TTL. by RealBadAngel
16:24 Megaf_       In thread 41761440:
16:24 Megaf_       /home/minetest/minetest/src/util/pointer.h:279: void SharedBuffer<T>::drop() [with T = unsigned char]: Assertion '(*refcount) > 0' failed.
16:26 Megaf_       (Leftover data: #3  void RemoteClient::GetNextBlocks(ServerEnvironment*, EmergeManager*, float, std::vector<PrioritySortedBlockTransfer>&))
16:26 Megaf_       maybe already fixed?
16:30 sapier       most likely not
16:31 sapier       megaf how often did you get that message about the missing player? as often as the old one?
16:33 Megaf_       sapier: not often at all, but when it happens is devastating
16:33 Megaf_       millions of lines every second
16:33 Megaf_       most likely to be fixed already
16:34 Megaf_       sapier: I get that on my Raspberry Pi server,
16:34 sapier       I don't think so, I made this an error because it's not supposed to happen, if you tell it's causing a lot of issues I may have missed a situation where this can happen
16:35 Megaf_       let me check,
16:35 sapier       could be a race condition about player beeing removed while something else is done ... chances a race condition does bad things are quite better on very very slow machines
16:36 Megaf_       sapier: I think you fixed that already, was just a matter of pulling a new snapshot
16:36 Megaf_       I will let you know if I see that again
16:36 proller      sapier, again things from your fantasy world ?
16:37 sapier       I didn't fix it on purpose
16:40 sapier       https://github.com/minetest/minetest/commit/960d731587f58036bd4957f4a77db41e145c2d04 I know you're blind proller but I didn't know you're that blind
16:40 proller      miss!
16:41 proller      its not my code
16:41 sapier       check blame
16:41 proller      check brain
16:41 sapier       you've been the one adding it
16:41 proller      but not me writing it
16:42 sapier       you expect me to fix all bugs for any code I touch why don't you do it yourself too?
16:43 sapier       https://github.com/minetest/minetest/commit/e47f6fdfe985b8907e0aa7372bb325ceb5ed5f4c btw you DID add it
16:45 proller      it was function from stackowerfow or google
16:46 proller      no
16:46 proller      https://github.com/minetest/minetest/commit/f764297be2d55c9b8350077e44e3d62a9e058bcf
16:47 proller      https://github.com/minetest/minetest/commit/f960c3be313c761fe02f9167dc1fe3fca5dc4ada
16:47 sapier       So you copy code from somewhere and don't even check it? That's not even close to an excuse ;-) ... well at least I don't have to check that code for copyright now ;-)
16:48 proller      fuck you
16:48 proller      it was squashed commit of https://github.com/matttpt/minetest/compare/ipv6-revised
16:48 sapier       well last one shows you're still guilty
16:49 sapier       previous ... all of those variables have been unsigned int prior your changes
16:49 proller      you still drugged
16:49 proller      and blind
16:50 sapier       OLD unsigned int a, b, c, d; ---> YOURS u8 a, b, c, d, addr;
16:51 proller      and stupid 8(
16:51 proller      again:
16:51 proller      it was squashed commit of https://github.com/matttpt/minetest/compare/ipv6-revised
16:51 proller      author = /matttpt/
16:51 proller      got it?
16:52 sapier       ther it's correct!
16:53 sapier       this is the one breaking it https://github.com/minetest/minetest/commit/f960c3be313c761fe02f9167dc1fe3fca5dc4ada
16:54 sapier       wait ok ok line 184
16:54 sapier       still you changed it 3 times ... and never did see it ;-)
16:55 sapier       but things like that happen that's not a problem at al
16:55 sapier       l
16:57 proller      yes, main problem - your overall shitcode
16:58 sapier       I wonder why I keep on trying to find even the slightest sign of self reflection from you ;-P
16:59 Megaf_       proller: make freeminer work without leveldb
16:59 proller      Megaf_, why?
17:00 Megaf_       or at least make it find that leveldb is already installed
17:00 sapier       megaf -->#freeminer ... this is about minetest ;-)
17:00 proller      its more possible
17:01 * Megaf_     looks forward for some MySQL support
17:01 proller      why?
17:01 proller      dont forget oracle and mssql
17:04 Megaf_       why not?
17:05 dsimon       hm
17:06 dsimon       in the far future, it might be nice to use something like mysql (or some other network-accessible db resource, like mongo or perma-backed redis)
17:06 dsimon       because that could allow minetest server sharding
17:06 proller      but why?
17:06 dsimon       however... that would be really really hard
17:06 proller      you cant run two servers with one database
17:06 dsimon       so the database thing would only get you like 1% of the way there :-)
17:06 proller      also leveldb is fastest
17:08 proller      and it not helps running server on very slow hardware, there is lot of slow places exept database
17:08 dsimon       yeah, indeed
17:09 dsimon       even if the disk were the bottleneck, you could fix that with a networked filesystem
17:09 dsimon       rather than having to mess with the actual app
17:09 dsimon       (or, you know, just getting a better disk...)
17:12 Megaf_       proller: why? To use different computers, one to run the server task and the other to run the database
17:12 proller      buy one corei7
17:12 Megaf_       this would take a huge load from the hardware where the server is running
17:12 proller      no
17:12 Megaf_       since everything run on ONE FUCKNG SINGLE CORE
17:12 Megaf_       for fucksake
17:12 proller      database load-save less than 1%
17:12 dsimon       Megaf_, databases are generally not CPU bound
17:13 proller      i have 8 cores on my mobile
17:13 proller      fuckin china 250$ phone
17:13 proller      you can also try run server on 486 cpu, but its stupid
17:14 proller      also this phone 40-50x faster than slowberry pi
17:15 Megaf_       proller: yep, my phone too
17:15 Megaf_       and my phone is x86
17:15 sfan5        dsimon: you can already use redis and you can already run 2 server with one redis server with that
17:19 proller      and network transfer will slower than local read
17:20 proller      in freeminer you can fal with 100-150 nodes per second speed without stuck
17:20 sfan5        <proller> and network transfer will slower than local read
17:20 sfan5        [citation needed]
17:20 proller      with loading blocks (generating slower)
17:20 Megaf_       why?
17:21 Megaf_       network can be as fast as local disk
17:21 sfan5        or even faster
17:21 Megaf_       all depends on impelemtation and protocol
17:21 proller      remote server = ask server -> server read -> server pack data -> server sends to you -> you recieve -> you unpack data
17:21 Megaf_       That may be a bad protocol
17:21 proller      leveldb =  you read from disk.
17:22 Megaf_       anyway, how do you think SATA works?
17:22 sfan5        local disk = ask kernel -> kernel calls driver -> driver reads data -> driver gives the data back to the userspace -> you receive -> you unpack data
17:22 proller      you cant read and two way transfer faster than stupid not locked read
17:23 proller      sfan5, remote server do same
17:23 proller      but + transfer
17:24 sfan5        any good protocol will not add much metadata
17:24 sfan5        also redis keeps everything in memory
17:24 sfan5        and then getting the data is literally just accessing RAM
17:24 proller      you cant make protocol without network transfer
17:24 sfan5        did i say that?
17:25 proller      keeping 4g data in memory when you use 50m - bad idea
17:25 Hunterz      with redis will be for example put/get to chest faster? or still will see some delay ?
17:25 proller      and you can warm local storage
17:26 sfan5        local storage is still much much much slower than RAM
17:26 sfan5        RAM only exists because local storage is too slow
17:26 Megaf_       finally someone who knows something
17:26 proller      slow for what ?
17:27 Hunterz      are there some minetest benchmarks ?
17:27 sfan5        slow for reading/writing data
17:27 Megaf_       and HUGE latency
17:27 sfan5        (compared to RAM, yes)
17:27 Megaf_       I can't believe we are still using mechanical disks
17:27 proller      but it not a bottlneck
17:27 sapier       Hunterz:  none I know about
17:28 Hunterz      any public server running wit redis? I would like to compare as player
17:28 Megaf_       minetestserver is jusr slow by design
17:28 * sfan5      uses redis for floatlands survival server
17:28 sfan5        but my upload speed is not that gret
17:28 sfan5        great*
17:28 proller      Hunterz, try any setun.net server ;)
17:28 * Megaf_     tested redis, leveldb and sqlite3, on a local network, everything is slow
17:28 Hunterz      ok, thank
17:28 proller      and compare with any redis
17:29 Hunterz      without redis i see lag during chest access
17:29 sfan5        it would be better to test it on a device that is actually has good upload speeds
17:29 proller      Hunterz, chest access - other problem
17:29 Hunterz      i have i7 with ssd and much ram
17:29 sfan5        proller: we are comparing database backends only, not all improvement freeminer might or might not have over MT
17:29 proller      sfan5, very important.
17:30 Hunterz      what problem have chest and propably others containers?
17:30 sfan5        can you reword your sentence?
17:30 RealBadAngel anybody checked #1308 ?
17:30 ShadowBot    https://github.com/minetest/minetest/issues/1308 -- Item entity stacks merge on the ground, add TTL. by RealBadAngel
17:30 sfan5        Hunterz: formspec needs to be sent to client
17:33 sfan5        <Megaf_> I can't believe we are still using mechanical disks
17:33 sfan5        that is just because mechanical disk space is so cheap
17:33 sfan5        and flash memory has it's disadvantages too
17:33 sfan5        we've talked about that already
17:37 Megaf_       I'm not a fan of SSDs either
17:37 Megaf_       ferromagnetic memory is a cood thing
17:38 Megaf_       reliable as hdds and fast as dram
17:38 Megaf_       but it's density is nothing close to the ssds
17:39 sfan5        throw some money at the part of freescale that researches this thing, maybe they'll do it faster
17:45 dsimon       sfan5, two minetest on one redis?
17:45 dsimon       i take it you don't mean they're running the same world
17:51 Megaf_       http://en.wikipedia.org/wiki/IBM_3592
17:51 Megaf_       250 MB/s on a tape
17:51 Megaf_       not bad
17:51 Megaf_       and 4 TB
17:52 dsimon       wow
17:52 dsimon       what's the seek speed on that, though?
17:53 Megaf_       a couple of seconds
17:53 Megaf_       this one is pretty impressive too, 8.5 TB http://en.wikipedia.org/wiki/StorageTek_tape_formats#T10000
17:55 sapier       I guess a desaster ;-) high continous reading bandwidth is completely to unrelated to read positioning ... in worst case this is as slow as reading from beginning to the location you wanna read ;-)
17:55 sapier       -to
17:55 dsimon       so, fantastic for dumps, not so great for random access
17:55 Megaf_       they are not made for random access
17:56 sapier       tapes are archive only so random access isn't at all a optimization target
17:56 dsimon       Megaf_, yah, i guess i'm just curious why you bring it up in a context when we were talking about a random-access database
17:56 dsimon       i guess just cause it's cool? :-)
17:56 Megaf_       Yep
17:56 Megaf_       isnt it?
17:56 dsimon       yah
17:57 sapier       in generally yes but not at minetest-dev ;-)
17:57 RealBadAngel sony has annouced 250TB tape based data storage...
17:57 Megaf_       RealBadAngel: Oh, I read about there somewhere
17:58 Megaf_       Quite easy to extend the capacity of tapes, just roll on more tape :P
17:58 sapier       for what I remember sony did it a different way
17:58 dsimon       i kinda like disc burner jukeboxes for that too
17:59 RealBadAngel not just that, it has density 148 GB per square inch
17:59 dsimon       though the write speed there is abysmal
18:00 RealBadAngel i think thats rather for archives, not regular usage
18:01 * dsimon     nods
18:03 RealBadAngel also they announced 1TB optical discs
18:03 RealBadAngel http://interestingengineering.com/sony-and-panasonic-announce-1tb-optical-discs/
18:04 RealBadAngel i dont even have blue ray yet, and it looks like it will become outdated before i get one ;)
18:09 Megaf_       RealBadAngel: but for now we have to use HDDs...
18:14 RealBadAngel ok, ive added config example to #1308, it is now complete
18:14 ShadowBot    https://github.com/minetest/minetest/issues/1308 -- Item entity stacks merge on the ground, add TTL. by RealBadAngel
18:17 sapier       RealBadAngel: did you remove the newline at the end of file?
18:18 RealBadAngel lemme check
18:19 sapier       imho 900 seconds are  a little bit small I'd suggest something around half an hour
18:20 sapier       and using on_activate to decrement age counter too
18:21 RealBadAngel 900s is 15 minutes
18:21 RealBadAngel and yup, newline is gone, shall i re-add it?
18:22 RealBadAngel on_activate is called only when item got loaded
18:22 RealBadAngel it has nothing to do with item's age (only just that it loads the variables)
18:22 sapier       hmm ok I somehow had 7.5 min in mind but you're right ... hmm 15 minutes is short too but not as short as what I thought
18:23 sapier       yes but you store item age so why don't you update it on_activate too?
18:23 RealBadAngel it depends on game time
18:23 RealBadAngel so no need to update it
18:24 sapier       hmm that's not good
18:24 sapier       if time changes or overflows that's gonna fail
18:24 RealBadAngel if it wont fail imho
18:24 RealBadAngel time = current time - time of creation
18:24 sapier       simple and for sure error free would be setting ttl to 900 and just do a count down
18:25 sapier       self.age = self.age -dtime
18:25 sapier       if self.age < 0 -->remove
18:25 sfan5        dsimon: no I don't
18:26 RealBadAngel but then unloaded items will live for way longer than they should
18:26 sapier       no
18:26 RealBadAngel ofc
18:27 sapier       on_activate provides time passed since deactivation
18:28 RealBadAngel hmm, how much is dtime?
18:28 RealBadAngel in seconds
18:28 sapier       unit is seconds
18:28 sapier       it's a float value
18:28 RealBadAngel 1 step != 1 second
18:29 sapier       yes that'S why you do a self.age -dtime
18:29 sapier       dtime is time since last step
18:29 RealBadAngel ah, ok
18:30 RealBadAngel but i dont want it to be decremented anyway
18:30 sapier       why not?
18:30 RealBadAngel because it serves also another purpose
18:30 sapier       which one?
18:30 RealBadAngel its "immune" time for item's auto pickup
18:31 RealBadAngel so one can throw away an item
18:31 sapier       and what'S the problem on decrementing?
18:31 RealBadAngel i assume you cant auto pickup item with time <3s
18:31 sapier       "core.get_gametime() - self.age" will cause an error on unit overflow
18:32 sapier       well that check would be (start_ttl - ttl) > 3
18:33 RealBadAngel ok, i will change that
18:41 BlockMen     RealBadAngel, why do you add this stacking stuff with TTL? that makes no sense at all
18:42 RealBadAngel ofc that makes sense, both solutions are to make server's life easier
18:43 RealBadAngel also having one bigger cobble rotating on the ground is way better than 99 ones
18:43 RealBadAngel even for just 15 minutes
18:43 BlockMen     that is your opinion
18:44 RealBadAngel not only mine, so far youre the only ONE that is against that
18:45 sapier       I share RealBadAngel's opinion, but I see reason to get a more wide acceptance of that feature prior adding/not adding it too
18:45 BlockMen     have you any proove that it lowers the server workload noticeable?
18:46 proller      TTL is wery bad idea
18:47 sapier       sorry RealBadAngel I guess I'm responsible for the second "no" proller always says no to what I suggest ;-)
18:57 RealBadAngel nore, and whats your opinion on #1308 ?
18:57 ShadowBot    https://github.com/minetest/minetest/issues/1308 -- Item entity stacks merge on the ground, add TTL. by RealBadAngel
18:57 RealBadAngel proller, you can always disable TTL
18:58 proller      its ok, but anyway loosing entities - not good
18:58 sapier       proller you can disable it
18:59 RealBadAngel its not good either players dropping their stuff everywhere and not takin care of them
19:00 proller      you can disable lot og users WTF about loose items
19:00 proller      RealBadAngel, why?
19:00 sapier       proller you're gonna loose them right now too
19:01 sapier       just drop it while there are already 49 entities (or whatever you configured)
19:01 RealBadAngel try dropping your diamond pick on the street, we will see how long it will last before some1 make it disappear ;)
19:01 proller      sapier, first problem with entities - your mobf
19:02 sapier       If I'd get a single cent for everytime I'm accussed for no reason I'd be rich ;-P
19:03 proller      no reason? lol
19:03 RealBadAngel what mobs have to do with items ttl for christ sake?
19:03 proller      mobf alvays overflow item limit everywhere
19:03 sapier       who told you you have to use mobf? if you don't like it don't use it
19:04 RealBadAngel proller, so just code better mobs, easy as that
19:04 sapier       or write valid error reports and suggestions ... just plain telling "it's bad" wont improve anything
19:04 proller      sapier, while you not use your soft all suggestions useless
19:04 RealBadAngel and we are not talkin bout mobs right now
19:05 sapier       sorry proller I can't play minetest 12h a day I have to do some work too ;-P
19:05 RealBadAngel but having less entities thx to mergeing and ttl will surely help mobs
19:05 proller      nothing helps mobs
19:05 proller      merging when digging - very good
19:06 proller      but dropping faster than ~week - bad
19:06 proller      i'm about defaults
19:06 RealBadAngel week? lol
19:06 proller      also sometimes lot of items located underground - they can be cleaned
19:06 BlockMen     i would be fine when they are merged with player action (lets say when droped by player)
19:06 BlockMen     but not e.g. for sapling
19:06 BlockMen     or other non player items
19:07 proller      RealBadAngel, on Sky server map near year old, and you can find year old entites in some places
19:07 RealBadAngel what about 4k of leaves falling down from oak?
19:07 RealBadAngel have you ever thought about that?
19:07 BlockMen     RealBadAngel, then redo the leave decay code, not the item code
19:07 RealBadAngel no
19:08 BlockMen     yes
19:08 RealBadAngel this is the common way how it should be done
19:08 BlockMen     better than this merge stugg
19:08 BlockMen     *ff
19:08 BlockMen     no, dont merge "natural" items
19:08 BlockMen     it is just odd
19:09 RealBadAngel if some1 wanna have nonsense behaviour, thousands of items spinnin in one pos, let him mod that
19:09 RealBadAngel current items behaviour is just pain in the ass
19:09 BlockMen     just because you say its "nonsense" it has not to be
19:09 BlockMen     and explain why it is currently a pain
19:10 RealBadAngel just stand in place and drop several stacks of cobble one by one piece
19:10 BlockMen     ^ lol, what?
19:10 RealBadAngel and see how it works and looks like
19:10 BlockMen     who would do that?
19:10 BlockMen     that is just stupid
19:10 RealBadAngel this is how it works
19:10 BlockMen     but no one does (except you maybe)
19:11 RealBadAngel me? game does
19:11 BlockMen     single items of stacks dont drop without player action
19:13 RealBadAngel btw, have you actually tried to play with that code?
19:13 RealBadAngel because i have the feeling you have not
19:15 RealBadAngel https://github.com/minetest-technic/item_tweaks
19:16 RealBadAngel and add that to config:
19:16 RealBadAngel enable_item_drops = true
19:16 RealBadAngel enable_item_pickup = true
19:17 RealBadAngel also about TTL, that can be used instead of bones
19:17 RealBadAngel drop the player's inv on the ground, add waypoint to death pos
19:18 RealBadAngel that will bring back to game some risk
19:19 RealBadAngel by now game feels like kindergarten, no risk at all, no reason to care about anythin
19:19 RealBadAngel and that makes game boring
19:22 BlockMen     game is frozen, so bones will stay
19:22 BlockMen     and no, but you know what
19:22 BlockMen     idc ...merge it or not
19:23 BlockMen     i will override it in my games anyway, so do what you want
19:23 RealBadAngel whats frozen can be melted ;)
19:23 RealBadAngel just jokin, but thats an idea to discuss
19:23 BlockMen     no, re-read the log of yesterday
19:24 BlockMen     lets game dead for now
19:24 BlockMen     *-s
19:24 RealBadAngel fighting boredom in gameplay is much more important than whacky idea of freezing the code
19:24 BlockMen     more important now is to figure out a way how to chose the games that will be bundled with 0.4.10
19:25 RealBadAngel that will bring much more fights to the channel than mt game changes imho ;)
19:25 BlockMen     read this
19:25 BlockMen     https://forum.minetest.net/viewtopic.php?p=138276#p138276
19:25 RealBadAngel i wont even start proposing technic_game for that
19:26 RealBadAngel mainly because technic_game is around 250mb right now
19:27 RealBadAngel but wont be released until 0.4.10 comes
19:28 BlockMen     well, we found a reason for not including games now: SIZE ;)
19:29 Calinou      not all games are 250MB
19:29 RealBadAngel ofc not
19:29 Exio4        even 20MB is heavy Calinou
19:29 BlockMen     i wouldnt include games > 5 MB
19:30 Calinou      Carbone is not much heavier, without .git
19:30 Calinou      I did everything to compress it :P
19:30 RealBadAngel but mine will come with high res tp, normals and parallax support
19:30 RealBadAngel and all the eyecandies possible
19:30 RealBadAngel including music
19:30 Exio4        do we have a moddb for games?
19:30 BlockMen     RealBadAngel, then its IMO not wise to bundle it with the engine
19:30 BlockMen     its just toooooo big
19:30 RealBadAngel i know
19:31 RealBadAngel technic game is not for basic download
19:31 RealBadAngel its a total option
19:31 RealBadAngel weaker machines wont even handle it
19:33 RealBadAngel BlockMen, so, what about trying item_tweaks?
19:33 RealBadAngel can you give it a shot?
19:34 RealBadAngel it has some more features not mentioned in this conversation...
19:35 BlockMen     RealBadAngel : http://irc.minetest.ru/minetest-dev/2014-05-21#i_3714187
19:38 RealBadAngel so just watch, if you dont want to try: https://www.youtube.com/watch?v=mnwk5XZ6DQE
19:40 BlockMen     ik the videos
20:05 RealBadAngel ok
20:06 RealBadAngel i guess you gonna just stay as one of the very few that dont like the feature ;)
20:07 RealBadAngel one thing wonders me, why proller in private said he like it but in public said he does not ;)
20:07 RealBadAngel maybe just for sake being opposite to sapier as usual ;)
20:17 iqualfragile RealBadAngel: is ther efinally real light in minetest?
20:18 RealBadAngel not yet, its a long road for just one coder...
20:19 RealBadAngel i am at the doorstep of rendering buffers and image postprocessing
20:20 RealBadAngel for that i had to rebuild whole shaders handling
20:20 RealBadAngel which is done and waiting for feedback
20:22 BlockMen     RealBadAngel, is it possible to disable waving for lava now?
20:22 RealBadAngel with the code that waits in line it will be
20:22 RealBadAngel shaders are assigned there by tile
20:22 RealBadAngel or by specific node
20:23 BlockMen     very good
20:23 RealBadAngel in the future i plan to introduce own lava shader
20:24 BlockMen     how would i disable it for lava then?
20:24 BlockMen     waving = 0?
20:24 RealBadAngel by now its assigned by liquid
20:24 RealBadAngel so all liquids do wave
20:25 RealBadAngel https://www.youtube.com/watch?v=zxUXNNFpPpc
20:25 RealBadAngel ^^ example lava shader
20:25 BlockMen     with your current pull, can i set lava to not wave at the same time having water waves?
20:27 RealBadAngel https://github.com/RealBadAngel/minetest/blob/shaders_unite/src/nodedef.cpp#L699
20:27 RealBadAngel we can add there another shader for just lava
20:28 RealBadAngel here water_shader is assigned just for one tile (because of water reflections, that will be shown just on surface)
20:29 BlockMen     and/or apply the shader only if waving = 1 (like for plants&leaves)?
20:29 RealBadAngel this is done in shaders code
20:29 sapier       maybe we should think about a way to specify shaders in nodedef?
20:29 RealBadAngel shaders are instanced
20:29 RealBadAngel for each drawtype and material type
20:30 RealBadAngel but from single source one
20:30 RealBadAngel https://github.com/RealBadAngel/minetest/blob/shaders_unite/client/shaders/nodes_shader/opengl_vertex.glsl#L57
20:31 RealBadAngel so all the transformations are depending on passed types to shader
20:32 BlockMen     https://github.com/RealBadAngel/minetest/blob/shaders_unite/src/nodedef.cpp#L687,L688 i mean like that, so having a liquid and a liquid_waving material
20:32 BlockMen     then depending on nodedef you can choose liquid or liquid_waving as material
20:36 RealBadAngel BlockMen, yeah, we can do that this way, no problemo
20:38 BlockMen     good :)
20:39 RealBadAngel this way we will separate all the liquids
20:39 RealBadAngel btw, there was an idea out there
20:40 RealBadAngel regarding water at all
20:40 RealBadAngel to introduce ocean water (waving) and still water - lakes, rivers, artificial channels - not waving
20:41 RealBadAngel but that should be made at mapgen level
20:41 RealBadAngel and then just handled
20:42 RealBadAngel sapier, looks like im done with using dtime for ageing
20:42 RealBadAngel im suprised because code become even simplier with that
20:44 sapier       I thought so that's why I suggested it
20:44 sapier       did you already update the pull request?
20:46 RealBadAngel not yet, adding now proper age for unloaded entities
20:55 * VanessaE   peeks in
20:55 VanessaE     is it safe now?
20:57 BlockMen     VanessaE, its never safe on #-dev :P
20:57 RealBadAngel lol
20:57 RealBadAngel hi VanessaE
21:00 VanessaE     hi
21:25 RealBadAngel sapier, something is wrong with dtime_s in on_activate
21:27 RealBadAngel http://pastebin.com/TwctaMkp
21:27 RealBadAngel it prints always 4
21:27 sapier       I'm gonna check
21:29 RealBadAngel it works when i let them unload without leaving to menu
21:29 RealBadAngel but when i exit to menu and go back
21:29 RealBadAngel it is always 4
21:29 sapier       sounds like something not beeing persisted correctly
21:31 sapier       that's strange
21:31 sapier       dtime_s isn't a float but a unsigned int value
21:33 RealBadAngel looks like, its always rounded
21:34 RealBadAngel  4
21:34 RealBadAngel  4
21:34 RealBadAngel  4
21:34 RealBadAngel 40
21:34 RealBadAngel 34
21:34 RealBadAngel 34
21:34 sapier       can you upload your current code for me to test?
21:34 RealBadAngel sure
21:34 sapier       dtime_s is only seconds
21:34 RealBadAngel ok, but why always 4 on start?
21:35 sapier       that's a good question
21:35 RealBadAngel propably time since server starts
21:36 sapier       maybe
21:36 sapier       that'd mean ttl will run in "gametime" instead of realtime
21:37 RealBadAngel https://github.com/minetest-technic/item_tweaks
21:37 RealBadAngel uploaded
21:37 sapier       that's not what I had in mind but it'd be not that bad
21:37 RealBadAngel not really
21:37 sapier       actually that's quite strange
21:38 RealBadAngel if player dies, he could just exit the game and go back
21:38 RealBadAngel to let his items live longer
21:38 RealBadAngel and thats kinda cheat
21:38 sapier       well that'd happen for gametime usage too
21:38 RealBadAngel actually no
21:38 sapier       that's a basic flaw in block timestamps
21:38 sapier       why not?
21:39 RealBadAngel game time flows
21:39 sapier       no
21:39 RealBadAngel reseting it to 4s each load is different
21:39 sapier       sorry don't get it yet
21:39 RealBadAngel it flows when game is loaded
21:39 sapier       gametime starts on each game
21:40 sapier       so if a block containing is unloaded he'd not recognize the change too
21:40 sapier       I suggest fixing the timestamps to be real timestamps not gametime timestamps
21:41 RealBadAngel lemme check something
21:41 sapier       dtime_s = m_game_time - block->getTimestamp();
21:41 sapier       is the code calculating the dtime_s
21:42 RealBadAngel youre wrong
21:42 RealBadAngel game time is time since world creation
21:42 RealBadAngel not since it is loaded
21:42 sapier       then there has to be another bug
21:43 RealBadAngel not really
21:43 sapier       true, it's correct
21:43 sapier       if you exit and enter that may be at least 4s
21:43 sapier       well on a very fast machine you could get shorter values
21:44 sapier       you could try by droping something, teleport away, exit, wait, telpor back
21:44 sapier       dtime_s should be the time you've been away (almost)
21:45 RealBadAngel it is when youre still ingame
21:45 sapier       it should be the time the game was running
21:45 sapier       if you stop your server say for a day, that time won't be honored
21:45 RealBadAngel ofc
21:46 sapier       I'd be fine with that behavioud, ok loosing subsecond precision wasn't anticipated but imho there could be worse
21:46 RealBadAngel but when servers shuts down it shall write current dtime_s
21:47 sapier       why? the dtime_s depends on gametime
21:47 RealBadAngel exactly
21:47 RealBadAngel but some of it got lost when server exits
21:47 sapier       and gametime is time of world existance? or did I missunderstand
21:48 RealBadAngel look, i flew away from items
21:48 RealBadAngel block got unloaded
21:48 sapier       ok
21:48 RealBadAngel i get back and had dtime_s around 40s
21:48 RealBadAngel now im exiting the game
21:48 RealBadAngel and loading it back
21:48 sapier       while beeing there or while beeing away?
21:48 RealBadAngel i do have dtime_s = 4s
21:49 sapier       if you exited right next to your items the 4s are most likely fine
21:49 sapier       dtime_s is time since last deactivation
21:50 RealBadAngel so i have to use game time for ttl no matter what
21:50 sapier       no
21:50 sapier       if you leave the game right next to the items
21:50 sapier       then gametime is stopped on exiting game
21:51 sapier       if you enter again it continues ... then once the items are activated you get told how much (game) time passed since they've been unloaded
21:52 sapier       or do I missunderstand what you wanna tell me?
21:53 RealBadAngel damn, now it looks like it started to work
21:53 RealBadAngel time seems to be correctly preserved
21:54 RealBadAngel hell, no
21:54 RealBadAngel something weird happens here
21:54 RealBadAngel lemme pastebin some timings
21:54 sapier       ok :-)
21:54 sapier       I wonder if there'S still something broken about entity de/activation
21:58 RealBadAngel http://pastebin.com/GuTumnvB
21:58 RealBadAngel ive dropped a few items
21:58 RealBadAngel exited the game
21:58 RealBadAngel and had proper timings
21:58 RealBadAngel now did that again
21:58 RealBadAngel and all is reseted
21:59 RealBadAngel and each time i get different values
22:00 sapier       let me have a look at your code I don't understand
22:01 RealBadAngel https://github.com/minetest-technic/item_tweaks/blob/master/item_entity.lua#L75
22:02 RealBadAngel it looks like sometimes it deserialzes data.age, sometimes dont
22:03 RealBadAngel definitely something is wrong with that part
22:04 sapier       that'S strange, what's the itemstring supposed to be? legacy?
22:08 sapier       hmm I have a suggestion to check, maybe entity staticdata isn't updated because of position not having changed ... that'd be a ugly bug ... but could be true
22:11 sapier       self:set_item(self.itemstring)
22:11 sapier       you reset the age on each activation
22:12 sapier       last line of set_item function
22:17 RealBadAngel huh?
22:18 sapier       you set self.age at end of set_item to 0 and you always call set_item in on_activate
22:18 sapier       that's resulting in self.age always beeing 0 after on_activate
22:19 sapier       that's why you see correct times on first join too as they're restored correctly ... but overwritten right after printing them
22:19 RealBadAngel stupid bug :)
22:19 sapier       all bugs are stupid once you've found them ;-)