Minetest logo

IRC log for #minetest-dev, 2013-01-21

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

All times shown according to UTC.

Time Nick Message
00:37 iqualfragile joined #minetest-dev
03:22 hmmmm joined #minetest-dev
07:41 Calinou joined #minetest-dev
07:45 celeron55 hmmmm: there is some really odd stuff in kwolekr/no_mgv7
07:45 celeron55 eg. https://github.com/kwolekr/minetest/commit/02c15321ba58f40a45bbfbe88039adc39bdb33fe
07:47 celeron55 i think i'll edit these commits to something reasonable
07:50 celeron55 this isn't very useful, as it doesn't even output any errors so the end result is likely only that the server jams due to trying to write too much stuff all the time: https://github.com/kwolekr/minetest/commit/556291100b689d6bfc0185fb37380757f8dd1087
07:50 celeron55 if it warned in errorstream, it would be a bit useful
07:51 * celeron55 edits
08:11 hmmmm weird... i could've sworn i actually added a line in guiFormSpecMenu.cpp in the former commit.  since a lot of time has passed since I first branched it, that could be the result of doing the rebase.. if you're wondering about the casts, those are there to stop warnings
08:12 hmmmm as for the latter commit, sorry, I didn't realize it would jam the server, but something needs to be done about that since I believe that was the root of the problem someone was having a while back where some blocks would be replaced
08:12 hmmmm if the disk was too full and the writes evidently failed
08:14 SiegeWizard joined #minetest-dev
08:14 celeron55 i'm also squashing serieses of fixup commits from here so the list of commits isn't so awfully long (and to reduce the amount of broken commits)
08:14 hmmmm yes, that would be great
08:14 hmmmm i wasn't really sure how to do that so to prevent breaking it all
08:14 hmmmm i just left it as-is
08:16 SiegeWizard http://minetest.net/wiki/doku.php?id=dev_index is this actual? it's at the description of the chat
08:18 celeron55 oh it's the old wiki
08:19 SiegeWizard that was my first guess
08:19 SiegeWizard should it be changed for the actual wiki?
08:19 Topic for #minetest-dev is now Minetest core development and maintenance. Chit-chat goes to #minetest. Consider this instead of /msg celeron55. http://irc.minetest.ru/minetest-dev/ http://dev.minetest.net/
08:20 SiegeWizard :)
08:22 celeron55 hmmmm: https://github.com/celeron55/minetest/commits/kwolekr_mg_r1
08:27 hmmmm looks great
08:27 hmmmm i need to get to sleep seriously
08:27 hmmmm no more of this 4am stuff
08:27 hmmmm sorry
08:31 celeron55 8)
08:32 Calinou https://github.com/celeron55/minetest/commit/25689e0c37ad6485ce2f73a60c800298f0b201e8 <= will it make the player spawn buried in ground?
08:32 celeron55 eh, what kind of a fix would that be?
09:11 Calinou wow, there is way too much jungle grass IMO
09:12 celeron55 i think it's dumb being everywhere
09:12 Calinou it should be made rarer and maybe have a shorter texture
09:12 celeron55 jungle grass should only be in jungles
09:12 celeron55 there could be other grass in some field-like places
09:13 Calinou then change the texture to a fieldgrass-like texture :P
09:13 celeron55 of course not?
09:13 celeron55 it's JUNGLE GRASS
09:15 celeron55 and jungle grass is meant to be just long enough to annoyingly block the player's vision if the player is not standing in a higher place
09:16 celeron55 i think it doesn't currently do that though
09:19 Calinou it's about 1.5 block high visually
09:19 Calinou player eye is 1.6 block high
09:19 celeron55 too short then 8)
09:35 SiegeWizard about the jungle, youll need biomes for that
09:37 celeron55 ............and?
09:42 celeron55 do you even know what a biome means, coding-wise, in practice?
10:16 BackupCoder joined #minetest-dev
10:16 SiegeWizard mmm
10:17 SiegeWizard a variable for each node?
10:17 SiegeWizard maybe?
10:17 SiegeWizard or you could make for each sector
10:17 SiegeWizard and make the border ones transitioning sectors
10:18 SiegeWizard i think that would be the simplest aproach to biomes codewise
10:19 SiegeWizard but the problem is that, you dont have biomes yet, do you? weren't they planned for the next release of mapgen?
10:22 celeron55 a non-moddable biome basically requires only a single line of code
10:23 celeron55 so "you need biomes" is quite a dumb thing to say
10:23 celeron55 we don't need practically anything for jungles, but we don't want to make them in the simple way
10:57 SiegeWizard i wasn't trying to say that
10:57 SiegeWizard i was trying to point out that, iirc, biomes were planned for v7 of mapgen
10:57 SiegeWizard and when that comes, you can create different types of grass depending of the biomes
10:58 SiegeWizard the jungle grass idea of blocking the player vision is a very good idea
11:02 celeron55 it's so in 0.3
11:03 celeron55 and about v7: yes, it is our (=hmmmm's) first attempt at definable biomes; it's still somewhat far from being actually usable though
11:03 SiegeWizard i think i may try to help him with that
11:04 SiegeWizard as I don't have time to read/understand the full minetest code for now
12:02 BackupCoder joined #minetest-dev
12:36 darkrose joined #minetest-dev
12:36 darkrose joined #minetest-dev
12:45 SiegeWizard does he have a github account with the project of the new mapgen?
12:45 SiegeWizard i haven't been able to find him
12:49 Progers joined #minetest-dev
12:51 Progers Hi :D
13:04 celeron55 SiegeWizard: this is my edit of his latest version: https://github.com/celeron55/minetest/commits/kwolekr_mg_r1
13:05 celeron55 (kwolekr is his github account)
13:09 SiegeWizard ok
13:10 SiegeWizard ill try to takje a look at mapgen.{cpp/h} from that fork and see if i can hel him with something
13:11 celeron55 you probably want to look at this though https://github.com/kwolekr/minetest/tree/yay_new_branch
13:11 celeron55 it has the code of the v7 generator included
13:11 celeron55 (kwolekr_mg_r1 is just the framework and optimizations and stuff)
13:17 SiegeWizard i found that by myself but thx
13:18 SiegeWizard well not that branch actually
13:18 SiegeWizard im still getting usaed to github
13:27 rsiska joined #minetest-dev
13:34 ptimatth joined #minetest-dev
13:35 ptimatth left #minetest-dev
14:31 SiegeWizard could you list the files that take part of the mapgen? so that i can copy them into my iPad and read them while travelling?
14:32 SiegeWizard related to the branch and repository you linked
14:35 celeron55 SiegeWizard: i am not too familiar with the changes, but mapgen.cpp/h and the headers included in there are a good start
14:35 celeron55 also hmmmm wrote this http://pastebin.com/ZhDtdWuQ
14:35 celeron55 that describes the changes from 0.4.4 and some of the design
14:36 celeron55 i guess i'll merge the server list now
14:37 celeron55 as nobody else seems to be doing it
14:45 RealBadAngel celeron55, please do that
14:46 RealBadAngel this stuff is very useful
14:46 sfan5 yeah
14:46 celeron55 it's there
14:46 RealBadAngel +100 for ya
14:47 sfan5 GIT: Jeija commited to celeron55/minetest: Add a list of servers to the "Multiplayer" tab 6f93c01af9 2013-01-21T06:42:56-08:00 http://git.io/QRBMHg
14:47 sfan5 time for another minetest build!
14:47 RealBadAngel yup
14:48 RealBadAngel i like jeija's commits
14:48 RealBadAngel particles are great
14:49 RealBadAngel and this one is even better
14:50 RealBadAngel btw, what timezone is hmm?
14:50 celeron55 now i'd like to have someone's input on https://github.com/celeron55/minetest/commits/kwolekr_mg_r1 (description here http://pastebin.com/ZhDtdWuQ)
14:51 celeron55 i've browsed through it and changed what i thought needed changing, but it's so large it's a good idea to use a couple of pairs of eyes on it
14:51 RealBadAngel i tested this with moretrees and plantlife we are workin on with VanessaE
14:51 RealBadAngel its really faster
14:51 celeron55 that is the version that does not actually implement any of v7, but sets up the framework
14:52 RealBadAngel yes, but the speed of v6 is noticeable
14:53 RealBadAngel i tested trees on current tree and his hmmm's build
14:53 RealBadAngel it was way way faster
14:54 RealBadAngel havent encountered any crashes with it also
14:54 RealBadAngel and i by weekend generated like 100+ new worlds
14:54 RealBadAngel half of them on hmmm's build
14:58 celeron55 now i recall one important thing
14:59 celeron55 hmmmm: (hope you read this) in map_meta.txt, the map generator should be referred by a textual name, not a number
14:59 celeron55 it is a very good idea to allow more native generators than just some sequential upstream ones
15:01 celeron55 the current one could be called "vanilla6", and if your biome one ends up upstream, it is "vanilla7" (currently that name is reserved for whichever first succeeds implementing nice biome support)
15:04 RealBadAngel well, we do have ready pretty nice implementation of biomes vs V6
15:04 RealBadAngel but we have to calculate everythin in Lua
15:04 RealBadAngel which is slow
15:05 RealBadAngel i think hmmm shall finish the work with c++ supported biomes. we can give him contens to fill them up
15:06 RealBadAngel for example, we cant really turn off default trees generator
15:07 RealBadAngel we can fake it by makin air aliases, but it still is workin, takin time
15:07 celeron55 hmm, now where was that mod enable/disable gui stuff
15:08 RealBadAngel that idk
15:09 RealBadAngel btw, celeron55, would you like to see world with mapgen generated L-system trees?
15:09 RealBadAngel 11 kinds of trees
15:09 celeron55 i'd like to see one that does not have any default trees and only L-system ones
15:10 RealBadAngel thats ours
15:10 RealBadAngel we do only L-systems
15:10 celeron55 but where the hell is that branch
15:10 celeron55 i want to see the status of that
15:10 RealBadAngel its all Lua except for generator itself
15:11 celeron55 < celeron55> hmm, now where was that mod enable/disable gui stuff
15:11 celeron55 talking about that
15:11 RealBadAngel https://github.com/VanessaE/plantlife
15:12 RealBadAngel https://github.com/VanessaE/moretrees
15:12 celeron55 ah, it was doserj
15:12 celeron55 https://github.com/celeron55/minetest/pull/416
15:12 RealBadAngel celeron55, please install those 2 mods
15:12 RealBadAngel and run new world
15:14 RealBadAngel you may encounter lags with generation, but it is workin almost fine
15:14 RealBadAngel Lua is not good for everythin sadly
15:20 RealBadAngel about #416, also pretty nice feature, but more useable for single/local players
15:20 RealBadAngel server managers wont use that, or wont need to
15:23 RealBadAngel celeron55, http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/screenshot_1503697289.png
15:24 celeron55 ...
15:24 celeron55 what kind of a "but" is local playing?
15:25 PilzAdam joined #minetest-dev
15:26 RealBadAngel local player likes to experiment
15:26 RealBadAngel and choose which world to play, right?
15:26 RealBadAngel server is run with one given set of options once
15:26 celeron55 i mean, local playing is obviously what this is for
15:26 RealBadAngel indeed
15:27 RealBadAngel seen the picture?
15:28 celeron55 yes?
15:29 celeron55 the size difference in trees is silly
15:29 RealBadAngel not when default are replaced with new kind
15:29 RealBadAngel beeches
15:30 celeron55 that is what i wanted to see, in case you don't remember
15:30 RealBadAngel which are slightly bigger than defaults
15:30 RealBadAngel i made beeches to be a step
15:31 RealBadAngel 3-4 nodes higher
15:31 RealBadAngel and 1-2 in radius bigger
15:31 RealBadAngel with some branches
15:32 RealBadAngel but we are hittin a wall, emerge thread
15:33 RealBadAngel we pass there our calculations and game is freezing
15:40 PilzAdam Could someone review this: https://github.com/PilzAdam/minetest/commit/da0e83b01cd51d0b12742bbac34ef349cb4be599 ?
15:41 RealBadAngel http://realbadangel.pl/actual.png
15:41 RealBadAngel celeron55, here you go with screenshot WITHOUT default trees
15:42 RealBadAngel those smallest ones are beeches. new default ones
15:43 PilzAdam RealBadAngel, the small ones are nice
15:43 RealBadAngel also some stupid timber mods proof ;)
15:44 RealBadAngel they do have branches
15:44 celeron55 i'd like to have only reasonably sized trees in the vanilla game until later
15:45 celeron55 and those trees are way too varied; in nature trees form forests with large areas of fairly similar species
15:46 celeron55 those at the bottom right of the image seem like a good maximum size; they seem generally a bit smaller than large cliffs which is preferable
15:47 RealBadAngel heh, lemme find one picture for ya
15:47 celeron55 also, there REALLY needs to be more varied leaf colors
15:48 RealBadAngel let it be this one
15:48 RealBadAngel http://www.rlrouse.com/pic-of-the-day/sequoia-tree.jpg
15:49 RealBadAngel do you still think sequoia model is too small?
15:49 celeron55 it should be so rare it wouldn't have gotten to that picture even if you tried to find one
15:49 RealBadAngel t is an evergreen, long-lived, monoecious tree living for up to 2,000 years, and is commonly considered the tallest tree in the world, reaching up to 115.5 m (378 ft) in height and 7 m (23 ft) diameter at the base.
15:50 celeron55 over the years i have found that it gives no benefit to enlarge things more and more
15:50 celeron55 for example in the map generator it easily makes very boring landscapes that nobody wants to traverse through
15:51 RealBadAngel yup, but thx to voxel world
15:51 celeron55 in trees, they just oversize everything and it's not fun to build or do anything anymore because making something larger than them would take so much work
15:51 RealBadAngel we need some size to make models look different
15:51 celeron55 the thing in games with this large voxels is that it is easy to build large things with some detail
15:52 celeron55 if you take that away, you take all that is good away
15:52 RealBadAngel oh no
15:52 RealBadAngel i have to disagree
15:52 RealBadAngel same as with 16x16 textures
15:53 celeron55 that i have known forever, and you will have so hard time trying to convince me otherwise it's easier to make a fork of MT
15:53 RealBadAngel its damn hard to make good texture with such low level of details
15:53 RealBadAngel no, we dont trying you to convince
15:54 RealBadAngel all others went already this way
15:55 celeron55 anyway; i think i'll merge this mod selection now and then go do something else
15:55 RealBadAngel models are subject to change
15:55 celeron55 by the way, how much do you play the game?
15:56 * VanessaE wanders in
15:56 RealBadAngel im not said all are perfect
15:56 RealBadAngel but it seems im the only one trying
15:56 celeron55 by the way, how much do you play the game?
15:57 RealBadAngel last month almost nothing, EU server is down
15:57 celeron55 i kind of respect two kinds of people when it comes to minetest: people who think like me, and people who play the game a lot while also developing
15:57 RealBadAngel and i cannot play on US servers :(
15:57 dzho why, too much lag?
15:58 RealBadAngel 10-20 seconds lag when tryin to place a node
15:58 RealBadAngel impossible to play
15:58 dzho yeah
15:58 RealBadAngel i played skyblock with that lag
15:58 RealBadAngel real hardcore
15:59 RealBadAngel if oldcoder wont re-run EU server soon
15:59 RealBadAngel im goin to buy new VPS to run the game for EU players like me
16:00 RealBadAngel and that, celeron55, should be your answer
16:01 RealBadAngel im doing somethin here, but WANT to play it also
16:01 celeron55 make sure to use some time thinking what would actually be a thing you would like to play
16:01 celeron55 often stretching technical things is NOT it
16:01 celeron55 even usually
16:02 RealBadAngel but with the trees...
16:02 RealBadAngel ask who DONT like them
16:02 VanessaE celeron55: the reason the trees seem to be too...mixed.. is because I'm not that good at defining biomes
16:02 celeron55 i like them, but way more thought should be spent in thinking how to make them fit with the world and gameplay
16:03 RealBadAngel we did
16:03 RealBadAngel thats because there are beeches
16:03 RealBadAngel easy to plant, easy to chop
16:04 doserj joined #minetest-dev
16:04 RealBadAngel but hands off from hundreds years olds oaks
16:04 RealBadAngel as in real life :)
16:04 celeron55 i think they simply don't look nice combined with the world currently, but don't know what it would need
16:04 VanessaE celeron55: make the map generator create larger, taller mountains/hills
16:05 celeron55 doserj: i merged your stuff; hope that it works 8)
16:05 VanessaE that would solve a lot of the problem I think
16:05 doserj oh :)
16:05 RealBadAngel i think you have never tried to PLAY with them
16:05 RealBadAngel did you?
16:05 celeron55 VanessaE: such are boring to play with our current generators; requires a lot of work to make them work
16:05 * VanessaE shrugs
16:05 RealBadAngel celeron55, please answer, have you tried to play with them?
16:05 celeron55 i have tried more map generating stuff than you can imagine
16:06 celeron55 bored myself to death many many times
16:06 RealBadAngel but the trees
16:06 RealBadAngel yes or no
16:06 celeron55 yes and no
16:06 RealBadAngel try to walk around them
16:06 celeron55 the answer is "a bit, but i don't play otherwise really at all either"
16:07 RealBadAngel if you have never did, please dont talk bout stuff you have never experienced
16:07 RealBadAngel HOWGH
16:08 doserj celeron55, ah, the mod_selection stuff. I have also a (now proper?) fix for the liquid update issue, if you want to have a look.
16:09 doserj (saving nodes to update in the mapblock)
16:09 celeron55 now i hope somebody else takes control if this chaos and go do something else ->A
16:09 celeron55 -A
16:09 VanessaE celeron55: under what conditions is the new server list/multiplayer tab supposed to save where you've connected to?
16:10 doserj Also, are people interested in having new_style_water work? I have a fix for that, but it may need changes to minetest_game to use the proper textures
16:11 RealBadAngel celeron55, as you can see all of us gettin specialized
16:12 hmmmm joined #minetest-dev
16:12 RealBadAngel hi hmmm
16:12 hmmmm hi
16:12 Progers hi
16:13 VanessaE good morning (etc), hmmmm.
16:13 hmmmm hi
16:13 doserj (by "I have a fix" i mean: "it may need some more work to get the details right, but I am only going to o it if people are interested)
16:13 Progers celeron55 you are on windows 7?
16:13 hmmmm lel
16:13 hmmmm i am pretty sure he'd shoot himself if he had to live with windows
16:14 Progers hehe)
16:16 Progers i think he is on ubuntu
16:17 hmmmm no, he's not
16:17 Progers Hm
16:17 hmmmm but what does it have to do with development of the core engine
16:17 Progers and how do you know?
16:18 hmmmm i've seen a screenshot of his desktop before
16:18 Progers show me
16:18 hmmmm it's one of those awesomewm arch l33t h4x0r sort of setups
16:18 Progers the screen shot
16:19 hmmmm you really want to know a lot about celeron don't you
16:19 hmmmm i don't have the screenshot on-hand....
16:19 Progers yes :D
16:19 hmmmm who would?>
16:19 hmmmm and again; what does it have to do with development of the minetest core?
16:23 PilzAdam celeron55, the mod selection handles empty modpacks like mods; this means that you can enable/disable the "Add-on" thing at the top of the list if you have no mods installed, wich produces a warning at startup
16:28 doserj lol, I didn't think of that case...
16:36 iqualfragile joined #minetest-dev
16:38 rubenwardy joined #minetest-dev
16:40 doserj for anyone who is interested: https://github.com/doserj/minetest/tree/new_style_water
16:41 doserj basically works but still has a slight lighting glitch
16:42 VanessaE screenshot?
17:01 doserj joined #minetest-dev
17:03 doserj damn internet. lighting glitch is fixed. screennshot of working version: http://imgur.com/3wAKDuw
17:04 VanessaE doserj: how does this differ from the effect caused by using shaders?
17:04 * VanessaE is comparing current git against the screenshot
17:05 VanessaE or yours lowers the water level without using shaders I guess?
17:05 Calinou joined #minetest-dev
17:07 doserj yes
17:07 VanessaE ah
17:07 doserj i didn't test this version with shaders on...
17:10 doserj that lowers the water level by twice the amount, as one could guess :)
17:10 VanessaE I'd say that needs...tweaked.
17:10 VanessaE :-)
17:11 doserj tweaking is not the problem. The problem is deciding what is the right behaviour
17:13 doserj another issue: how should it behave if two different liquid sources are on top of each other. What should the liquid level of the lower one be?
17:13 VanessaE full height.
17:13 VanessaE otherwise it looks weird with a gap between the two
17:14 VanessaE or, lower the bottom surface of the upper liquid
17:14 doserj probably: should a face still be drawn?
17:14 VanessaE yes.  what if there were two translucent liquids atop one another?
17:14 Calinou oh no, transparency sorting discussion
17:14 VanessaE a boundary layer should be visible then, if possible.
17:18 doserj and then we come to the issue i was originally trying to solve today: flowing liquids don't show their bottom face currently.
17:21 doserj http://imgur.com/bWrFLrW
17:21 VanessaE ah yeah I've seen that effect.
17:33 Calinou without shaders, they do
17:37 doserj what?
17:45 celeron55 hmmmm: did you see my request?
17:45 hmmmm what no
17:46 hmmmm let me check the channel logs
17:46 celeron55 15:00 UTC today
17:49 hmmmm okay
17:49 hmmmm yeah no problem, just a little changing
17:49 hmmmm how about we abbreviate "vanilla6" as "v6"?  8)
17:50 celeron55 i think no; using a long name encourages other people to use long names too
17:50 hmmmm vanilla implies we're boring though
17:50 celeron55 v6 sounds like "version 6" and the next guy is going to think they are just version numbers
17:52 celeron55 well, use v6 then
17:52 celeron55 i don't see how that is "less boring" than vanilla6 though
17:52 darkrose builtin6, default6, doesn't have to be 'vanilla'
17:53 celeron55 as long as if i make an experimental mapgen, i can call it "celeron55_mountains_1" or something
17:53 PilzAdam what about 6.gen?
17:53 celeron55 PilzAdam: eh, no? ".gen" is completely redundant
17:54 hmmmm i'd prefer either vanilla6 or v6, really would like v6 since it could be interpreted as a version or vanilla depend on how you see it
17:54 hmmmm it sounds cooler too
17:54 hmmmm and i don't have to change any of the class names in the code
17:54 celeron55 < celeron55> well, use v6 then
17:54 hmmmm yaaaaaaaaaaaah
17:56 celeron55 we probably want to reserve the possibility of some day having a large collection of unique native generators
17:56 celeron55 might not happen, but i don't want to be changing any file formats if it happens
17:57 celeron55 the order of importance in getting things right the first time goes like 1) file formats, 2) lua interfaces, 3) internal interfaces, 4) network protocol
17:57 celeron55 or, well, 3 and 4 might go the other way around
17:57 celeron55 depends
17:57 celeron55 point is, file formats are very persistent
18:03 hmmmm no i totally agree
18:04 hmmmm for some reason i have a feeling that the 'large collection of unique native generators' is going to consist of 3 liner changes
18:04 hmmmm well, you can't stop people.
18:06 hmmmm can you do a switch() statement on std::strings?  probably not, right?
18:07 hmmmm this is going to require a new way to 'register' mapgens i think
18:08 hmmmm you know, to come up with a neat and elegant way to do this, it probably won't be a 5 minute fix.. for now, are numbers okay?
18:09 darkrose hash it and switch the hash
18:09 rubenwardy1 joined #minetest-dev
18:09 hmmmm would do that if there was a convenient hash function laying around
18:10 hmmmm alright, i need peoples' opinions then.  would you like a "third party mapgen" to have to edit some static array in the code?
18:13 SiegeWizard i think that the point of having a mod-able game it's that, being able to mod it, without the need of touching the source code of the game
18:14 SiegeWizard so the asnwer is: no, I would prefer if there could exists third-party-mapgens without the need of touching any part of the source
18:14 hmmmm how would i do this without resorting to function pointers?  for (int i = 0; i != ARRAYLEN(mapgen_list); i++) { if (mgname == mapgen_list[i].name) { mapgen = mapgen_list[i].create(); printf("using %s mapgen\n", mapgen->getName(); ); } }
18:15 doserj SiegeWizard: we are talking about c++ mapgens. you don't want to write a mapgen in lua...
18:16 SiegeWizard i know
18:16 SiegeWizard my previous coment still aplies
18:16 SiegeWizard mapgen in lua will be very slow
18:18 hmmmm the core mapgen IS game code.
18:19 hmmmm i would prefer for it to be completely encapsulated into one source file only, but that's simply impossible
18:19 hmmmm so there'll be a mapgen_list array the user adds the mapgen name to and the address of a create function that'll return an instance of itself
18:20 celeron55 hmmmm: well umm... the mapgens should be just stored in a std::map<std::string, a function to get an instance of the mapgen>
18:20 SiegeWizard mmmmm
18:20 celeron55 string being the name
18:20 celeron55 or something like that i guess
18:20 hmmmm how do you statically define a std::map
18:20 hmmmm you can't!
18:20 celeron55 statically?
18:21 celeron55 oh, yes; you can't
18:21 hmmmm mmm.
18:21 celeron55 well, just make a plain C array of struct{name, that func}s
18:21 hmmmm alright alright so i guess what we do is have Server::Server() call a function which is where we'd have the user write a line of code adding their mapgen to the mapgen list
18:21 celeron55 a simple linear search doesn't hurt in that case
18:21 SiegeWizard maybe would be easier to make a create funtion that accepts as a param the name and calls the proper funtion
18:21 hmmmm that's what I already did, see above
18:21 hmmmm but i was afraid it wasn't C++y enough
18:22 SiegeWizard instead having all that create funtions in an array
18:22 celeron55 hmmmm: you just described how most C++ programs register their internals to abstraction layers :P
18:22 hmmmm keep in mind that minetest is my first foray into C++, i've never done C++ before
18:22 rubenwardy1 will the map gen be a dll plug in?
18:22 hmmmm no.
18:23 SiegeWizard hmmmm another thing
18:23 celeron55 but seriosuly, i don't really care how this is done now, and whether the C++ side even actually supports anything else than v6; i just want the file format to be more final at this point
18:24 SiegeWizard wouldnt it be faster to use: printf("using %s mapgen\n", mapgen_list[i].name ); ?
18:24 hmmmm lol
18:24 celeron55 :D shut up, that's the most useless advice ever
18:24 hmmmm oh that's true, i could just do that.  i was sort of thinking in the back of my mind that the getName() would return a more complete, descriptive version
18:25 hmmmm while "mgname" would just be the internal name
18:25 hmmmm but there doesn't really need to be a difference as long as it's not ridiculous
18:25 hmmmm i don't know yet.
18:25 hmmmm that's something i'll figure out later
18:26 SiegeWizard i know it's just an stupid detail
18:26 SiegeWizard but thought i could tell it
18:26 darkrose heh, "v6" getName() spits out a 4KB debate on why vanilla is/isn't boring
18:27 hmmmm shit you're right
18:27 hmmmm alright, getName() won't exist
18:27 hmmmm blunder of the century
18:32 celeron55 anyway, tell me when the map_meta format has a string as a mapgen name (instead of a numeric version) and i'll probably be lazy enough to not try to find more things to do
18:33 celeron55 8)
18:33 hmmmm in like 5 minutes
18:37 SiegeWizard i wonder why do you choose to have map generator identificators to be non-numerical if that makes you unable to use switch
18:38 hmmmm he just explained why above
18:38 hmmmm alright works
18:38 hmmmm pushing
18:38 SiegeWizard isn't it a purely aesthetic change that makes the code a bit more complex?
18:39 hmmmm yes, but this aesthetic change is one that changes the minds of others
18:40 SiegeWizard I don't get the point sorry
18:40 hmmmm https://github.com/kwolekr/minetest/commit/6899e643c043593218dee495f420d3647ca86b33
18:40 hmmmm I'll make it more elegant later on
18:41 hmmmm don't worry about it now
18:43 PilzAdam joined #minetest-dev
18:45 doserj SiegeWizard: one point is that this way nobody has to decide which map generator gets which number assigned.
18:46 celeron55 the exact thing as was done for nodes back when 0.4 was designed
18:46 celeron55 0.3 uses plain integer enumerations
18:47 celeron55 i could've gone for modders guessing some probably free numbers... or just use strings
18:47 hmmmm best idea:  have them use GUIDs
18:47 hmmmm ..just kidding
18:47 celeron55 internally, the engine still uses numbers for nodes, but they are assigned dynamically at runtime
18:47 Calinou how 'bout public-private keys
18:47 Calinou ...kidding again
18:48 Calinou why not just use numbers?
18:50 sfan5 <doserj> SiegeWizard: one point is that this way nobody has to decide which map generator gets which number assigned.
18:50 sfan5 ^ Calinou
18:52 doserj i can see several different map generators in the near future: PilzAdam's flat map generator, a very simple one that only generates air (useful for skyblock maps), and I may actually try to write a skygrid map generator.
18:52 doserj and there is hmmm's v7, of course
18:53 hmmmm actually you can generate flat maps right now with v6
18:53 sfan5 a skygrid mapgen should be easy
18:53 SiegeWizard okey so you want them just to have to assign a name and a funtion?
18:53 hmmmm you just need to set np_mgv6_terrain_base.scale = 0 and np_mgv6_terrain_higher.scale = 0
18:53 hmmmm :-)
18:54 hmmmm this is what i'm afraid of.... people are going to start making mapgens for things that you could already do with an existing mapgen that just needs a settings tweak
18:54 sfan5 SiegeWizard: actually a ame and a class
18:54 sfan5 *name
18:56 Calinou I segfault when clicking "connect" when no server is selected (and the list is empty, too)
18:56 Calinou :P
18:57 Calinou "Delete" button does that too
18:57 sfan5 the functions should check the length of the list
18:59 PilzAdam doserj, my flat mapgnerator does not create air only; it creates a flat map with 2 dirt layers and stone
18:59 Calinou I still segfualt
18:59 doserj PilzAdam: I know, i was listing different map generators
19:00 sfan5 Calinou: segfualt  is that a new type of exception? ;)
19:00 PilzAdam oh, I thought you described mine :-)
19:01 doserj I thought it was self-explanatory  :)
19:02 darkrose SIG_SPELLCHECK
19:02 sfan5 haha
19:06 hmmmm i'm going to have to make up some "presets" for V6 mapgen params
19:06 hmmmm one would be "superflat", another would be "plains", another would be "plains with mountains", another would be "large features", another would be "waterworld", another would be "no water", etc.
19:07 hmmmm "lots of desert"
19:07 PilzAdam super flat: http://forum.minetest.net/viewtopic.php?id=4412
19:08 hmmmm none of these presets i'm talking about require any change of the source code whatsoever
19:08 sfan5 thats good
19:09 sfan5 celeron55: i made some modification to the cross-compiling script (like curl support and only-download-files-if-they-don't-exist or manual download if automatic download fails) /msg if you're interested
19:12 celeron55 it doesn't really hurt at all if people try making map generators (which i actually doubt); it's actually a good thing because they'll be introduced to the C++ side that way
19:12 thexyz sfan5: pull request
19:12 celeron55 there is a lot of space for different generators that are not just variations of existing ones
19:12 celeron55 for example the 0.3.1 generator is almost completely different in every possible way
19:13 celeron55 it shares like nothing with v6
19:14 thexyz sfan5: 23:12   thexyz: sfan5: pull request
19:14 thexyz in case you missed that part
19:15 sfan5 i didn't miss it, your bouncer was so nice to tell me
19:15 sfan5 <thexyz> [01:12:33] sfan5: pull request
19:16 thexyz great, now do that ^^
19:16 sfan5 i'll do that
19:24 celeron55 thexyz: do you wish to complain about https://github.com/celeron55/minetest/commits/kwolekr_mg_r1 before merging
19:24 sfan5 celeron55: is it ok if i add a dropbox link to buildbot, you can't find libvorbis-1.3.2-[dev,dll].7z anywhere  i had to convert an .rpm package to get it
19:26 celeron55 maybe thexyz could host them? i would if i had a proper host but currently i don't
19:27 thexyz celeron55: how can I complain about hmmmm's code? it's beatiful
19:27 thexyz yes, I can
19:27 thexyz sfan5: PM me those files
19:27 rubenwardy so you are adding map gen modifications?
19:27 rubenwardy (but not with plugins)
19:28 hmmmm the way i'm going to have it set up, in the future if somebody writes a simple dll/so loader then it's possible to really have a plugin
19:28 hmmmm they'd just need to call EmergeManager::registerMapgen() on DllLoad
19:28 hmmmm DllMain or whatever
19:28 sfan5 what about the mac users, there needs to be a dylib loader for that case
19:29 hmmmm doesn't mac use .so?
19:29 hmmmm i should be able to dlopen() it the same
19:29 loggingbot_ joined #minetest-dev
19:29 Topic for #minetest-dev is now Minetest core development and maintenance. Chit-chat goes to #minetest. Consider this instead of /msg celeron55. http://irc.minetest.ru/minetest-dev/ http://dev.minetest.net/
19:29 thexyz done
19:30 thexyz oops, wrong channel
19:31 sfan5 thexyz: sfan.sf.funpic.de/libvorbis-1.3.3-pkgs.zip
19:31 PilzAdam celeron55, the settings should be added to minetest.conf.example
19:31 celeron55 what settings
19:32 PilzAdam https://github.com/celeron55/minetest/compare/master...kwolekr_mg_r1#L5R176
19:36 celeron55 doing
19:37 celeron55 hmm, mg_flags is bad
19:38 celeron55 it should expand the bits to a list of strings, or something
19:39 celeron55 whatever, i hope hmmmm makes this better before it's too late, merging now 8)
19:39 celeron55 we're going to end up with horrible hacks to support old parameter formats anyway i guess
19:40 * celeron55 is ah so positive
19:43 celeron55 "release early, release broken" or whatever they say
19:43 celeron55 8))))
19:43 PilzAdam thats what ruined Gothic 3
19:45 celeron55 technically this isn't a release though
19:45 celeron55 but in practice it is, because active users will get an early build of it an use it anyway
19:45 sapier joined #minetest-dev
19:45 celeron55 it's pushed
19:47 VanessaE oh boy.
19:47 VanessaE *looks at list of commits*
19:47 celeron55 thexyz: what is the state of the db abstraction layer
19:47 VanessaE I'm afraid to build this ;)
19:48 PilzAdam VanessaE, only takes 10 sec.
19:48 celeron55 i would be afraid too
19:48 thexyz celeron55: the same as before, it works
19:48 celeron55 master is somewhat experimental now
19:48 VanessaE PilzAdam: read "build" as in "build and deploy"
19:48 thexyz what's its state should be?
19:48 sfan5 celeron55, thexyz: https://github.com/celeron55/minetest/pull/433
19:48 celeron55 thexyz: compatible with hmmmm's changes?
19:48 celeron55 probably not perfectly
19:48 thexyz sfan5: meld ur in one commit
19:48 thexyz *it
19:49 thexyz celeron55: ok, I'll update it to work with latest master
19:49 thexyz tomorrow maybe
19:49 sfan5 ...ok
19:50 PilzAdam minetest crashes when no world is selected and you click on "Configure"
19:50 thexyz don't do this
19:50 hmmmm what
19:50 hmmmm the mg_flags things
19:51 hmmmm well in my defense that's only internal
19:52 hmmmm the user, if they'd like to tinker with it, will have to construct the correct flag values on their own, or wait until the configuration GUI gets finished
19:57 PilzAdam celeron55, you missed cloud_height
19:59 sfan5 https://github.com/celeron55/minetest/pull/433 is ready  for merging
20:11 thexyz btw, PilzAdam, better fill an issue
20:11 thexyz regarding that crash
20:12 celeron55 hmmmm: servers are configured from files
20:12 celeron55 and always will be
20:13 hmmmm people who run servers are smart enough, right?  can't we just give them a list of the values of each flag and tell them to put the "summation of all attributes you'd like"
20:13 thexyz msvc2010 build is broken
20:14 thexyz http://pastebin.com/raw.php?i=iLYPUuKb
20:14 celeron55 hmmmm: no; it's just dumb to not have clear values in a textual file
20:14 thexyz wait, i'm not really sure if it's minetest fault
20:14 celeron55 the reason why it is text in the first place is that it can be user-edited
20:14 hmmmm yeah man your SDK looks busted
20:17 celeron55 sfan5: testing
20:18 thexyz hmmmm: nope
20:18 hmmmm can you show a bit more of the compile?
20:20 thexyz it fails at #include <winsock2.h> #include <windows.h>
20:20 thexyz oops
20:20 thexyz at porting.cpp
20:20 hmmmm this is related to my changes i guess
20:21 thexyz relevant (probably) http://stackoverflow.com/questions/1372480/c-redefinition-header-files
20:22 hmmmm the only two places where winsock2 is included are jmutex.h and socket.cpp
20:22 hmmmm winsock2 really doesn't have a header guard for windows 7?
20:22 hmmmm you can't be serious
20:22 celeron55 sfan5: hmm, i guess thexyz should host openal_stripped.zip too
20:23 thexyz of course it has
20:23 hmmmm thexyz, try commenting out #include <windows.h> in porting.cpp:95
20:23 sfan5 celeron55: yeah
20:24 celeron55 ...and mingwm10.dll
20:25 celeron55 thexyz: throw these on the server too http://c55.me/random/2012-04/mingwm10.dll http://c55.me/random/2012-04/openal_stripped.zip
20:25 sapier https://github.com/celeron55/minetest/pull/391 any chance to get this into core?
20:25 celeron55 openal_stripped.zip should have oal_wrap.dll or whatever it was added too though
20:26 celeron55 actually, wrap_oal.dll i think
20:26 celeron55 (to actually have all that openal needs)
20:26 thexyz celeron55: done
20:27 thexyz hmmmm: nope
20:27 celeron55 hmm where i'd find that now
20:27 thexyz minetest.ru/bin
20:27 celeron55 download wrap_oal.dll from some shady google search result to there too
20:28 celeron55 then we have all that we will need
20:28 hmmmm shoot
20:28 hmmmm i have no clue to be honest
20:38 celeron55 sfan5: i don't have a suitable system to try this completely right now
20:39 PilzAdam the script works for me, but I get some errors while compiling
20:41 sfan5 PilzAdam: which errors?
20:41 PilzAdam https://gist.github.com/4589151
20:42 sfan5 do you have 64 bit?
20:43 PilzAdam yes
20:43 sfan5 it may not work correctly on 64bit systems
20:47 thexyz celeron55: i do have some d494267bc169604fac5e3679b9a97fed  wrap_oal.dll
20:48 celeron55 thexyz: that's the same as in the official 0.4.4 package
20:49 thexyz hmmmm: oh. there's also #include <windows.h> at line 27
20:49 thexyz commenting both seems to work
20:50 celeron55 it's funny how robustly windows handles all kinds of binaries... you can use the same old 32-bit x86 DLL for like tens of years
20:51 celeron55 on the other hand, they obviously cannot afford ever breaking any compatibility 8)
20:57 thexyz celeron55: this patch fixed the build issue i was talking about http://bpaste.net/show/71906/
20:57 thexyz funny header hell
20:57 sfan5 uh-oh!
20:58 sfan5 the new mapgen broke crosscompilng, its not an 64 bit issue PilzAdam
20:58 sfan5 *cross-compiling
20:59 sfan5 celeron55: i updated the pull request for auto-downloading of openal_stripped.zip and mingwm10.dll
20:59 sapier auto downloading binary code?
21:00 sfan5 yes
21:00 sapier I hope you do at least ask user if you're allowed to to malicious things like that?
21:01 sfan5 it is required to download libs for cross-compiling?
21:01 sfan5 -?
21:02 thexyz sapier: user? you mean, developer?
21:02 sapier where's the difference?
21:02 thexyz those guys who build win releases are smart enough to figure it themselves
21:03 sapier I don't like programms/scripts downloading anything without asking me
21:03 thexyz why?
21:03 thexyz what's the problem with downloading? it's not like downloaded code will be executed
21:03 sapier because any download can harm your system, if an applicatio doesn't even tell me its suspicious
21:03 thexyz how can it harm your system?
21:03 sapier a dll is binary code thus can be executed
21:04 sfan5 the downloaded lib will only be used for linking, as long as you don't run the resulting minetest.exe or minetestserver.exe othing can happen
21:04 sfan5 *libs
21:04 sfan5 *nothing
21:04 thexyz shall it be like >downloading lib[Y/n], >creating file [Y/n], >making dir[Y/n], >compiling file[Y/n] ?
21:04 thexyz damn, this conversation is pointless
21:05 sapier no it isn't it's a matter of trust
21:05 celeron55 eh
21:05 thexyz then don't use it
21:05 sapier any trustwothy systema asks even cpan does
21:05 celeron55 sapier: the whole point of buildwin32.sh is to DOWNLOAD stuff, and then build
21:05 celeron55 if you run it, you agree to what it does
21:06 sapier I wouldn't expect a "build" does download
21:06 sfan5 as long as you don't run the resulting executeables nothing can happen
21:06 sfan5 ..how are you going to cross-compile w/o the required libraries?
21:07 sapier I've already called a script I can't control what happens afterwards
21:07 thexyz celeron55: anything regarding headers issue?
21:09 sapier yes I know this seams paranoid ... but doesn't change the fact that downloading without asking is suspicious
21:10 sapier usually I download a build environment by myself
21:10 thexyz fine
21:10 thexyz don't run buildwin32.sh then
21:11 sapier i wouldn't mind about a download script either ... but build isn't prepare
21:11 celeron55 if you don't want anything unpredictable to happen, you shouldn't run any .sh without first reading it anyway
21:11 celeron55 unless it comes from a signed package from your package manager
21:11 sapier it's your decision how to do it of course
21:11 celeron55 yes it is, and this how we do it
21:12 sfan5 sapier: the build script doesn't download any libs if you alread downloaded them manually and put them in place
21:12 celeron55 you should take a look at all the tech going on on the web - it's all about automating everything and generously using all resources available 8)
21:13 sapier I should read any code to avoid that of course but why doen't cmake autoinstall all dependencys ?
21:14 celeron55 eh
21:14 sfan5 thats not the point of cmake
21:14 sfan5 its a build system
21:14 celeron55 thexyz: just fix it, dunno
21:14 sapier exactly "build"
21:14 sfan5 and not a do-it-all-automagically-system
21:15 sapier you wouln't expect make to download anything and so I wouldn't expect build32... to download anything
21:15 sfan5 its a buildbot
21:15 celeron55 sapier: the script is in the util/buildbot directory
21:16 sfan5 bots do things automatic, so ...
21:16 thexyz celeron55: ok
21:16 sapier as I said asking is a matter of trust
21:17 sapier it's not a big deal to ask at begining of script if its allowed but it's a big problem if by some accident you download an infected binary and execute it
21:17 celeron55 sapier: does your irc client ask every time if you are sure to send the line?
21:17 celeron55 sapier: if not, why? it's a matter of trust
21:18 sapier pressing enter approves sending this line and I exactly know this line will be sent
21:18 celeron55 okay, now if you continue this, you will receive a kick
21:18 thexyz celeron55: also, anything regarding my travis suggestion? i already have working .travis.yml
21:18 celeron55 which will be approved by my nice press of enter
21:19 celeron55 thexyz: i can't care about everything
21:20 thexyz celeron55: i can't enable hook required to integrate with it
21:20 thexyz http://about.travis-ci.org/docs/user/getting-started/#Step-one%3A-Sign-in here, steps 1 and 2
21:21 sapier I assume this is how different oppinions are handled in here ... don't worry I'll leave myself.
21:21 celeron55 sapier: it's democracy
21:21 sapier left #minetest-dev
21:22 celeron55 "i am the only one with this opinion but it is the right one"
21:22 celeron55 8D
21:22 sfan5 :D
21:23 thexyz inb4 another "leaving" post?
21:23 celeron55 minetest drama
21:23 celeron55 your daily dose (tm)
21:25 sfan5 "i'm leaving minetest because the others are wrong and i am right"
21:26 thexyz that's what happened before
21:26 sfan5 yeah
21:27 celeron55 seriously, i'm already 100% used to this
21:27 celeron55 i'm just laughing :D
21:27 PilzAdam I everyone except you leaves you should think about it
21:29 celeron55 what is there to think
21:30 PilzAdam block the word "leave" in the forum?
21:30 PilzAdam :-)
21:30 celeron55 nothing, basically - people come and go, and some want to take some kind of a nudge before they feel confident in leaving
21:31 celeron55 in small and large time scales
21:35 celeron55 thexyz: i need to turn this on for minetest and minetest_game?
21:36 thexyz minetest only
21:36 thexyz how is it supposed to "build" minetest_game?
21:36 celeron55 uhm... guess not 8)
21:37 celeron55 it's now supposed to be set up
21:37 thexyz great
21:37 celeron55 (1 and 2, that is)
21:38 celeron55 hmm
21:38 celeron55 actually, maybe not; checking...
21:39 thexyz ok, now this thing should start building https://travis-ci.org/celeron55/minetest/builds/4291981
21:39 celeron55 i think it's set up
21:40 thexyz oh, btw, it'll spam your with emails starting from now
21:40 celeron55 and how do i turn that off
21:40 thexyz like >The build passed or >The build failed
21:41 celeron55 ...in no way?
21:41 celeron55 well this is bullshit
21:41 celeron55 complete, utter
21:41 thexyz it can be done!
21:42 thexyz like https://github.com/puppetlabs/puppet-webhooks/pull/9/files
21:42 celeron55 do that
21:42 celeron55 obviously
21:42 celeron55 it's not like i would have the slightest of interest to read any kind of travis spam
21:43 thexyz one moment
21:43 Taoki joined #minetest-dev
21:45 celeron55 i already hate the travis web interface
21:45 thexyz done
21:45 celeron55 this is so unresponsive without any load indicators or anything
21:46 celeron55 oh god... *closes page*
21:47 celeron55 it's worse than the minetest inventory with bad mapgen mods :-D
21:48 celeron55 hopefully the backend works
21:50 VanessaE hah
21:50 sfan5 good night
21:50 thexyz night, sfan5
22:00 thexyz passed https://travis-ci.org/celeron55/minetest/builds/4291981
22:49 rsiska joined #minetest-dev

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