Minetest logo

IRC log for #minetest-dev, 2017-03-15

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

All times shown according to UTC.

Time Nick Message
01:09 paramat joined #minetest-dev
01:21 Lunatrius joined #minetest-dev
01:36 AntumDeluge joined #minetest-dev
02:00 paramat #5388 now tested, sorts out the map gen limit mess
02:00 ShadowBot https://github.com/minetest/minetest/issues/5388 -- Map gen limit: Rewrite by paramat
02:12 paramat sofar, considering the controversy, shall i now merge #5327 ? i'mnot sure. looks like 4 devs support and 2 object. it has 2 approvals
02:12 ShadowBot https://github.com/minetest/minetest/issues/5327 -- Fix various problems with sneaking by sfan5
02:15 paramat thought i'd ask before merging since its controversial
02:18 paramat is your +1 a code review too?
02:31 paramat for now i'll wait and look at the code myself
02:35 XtremeHacker joined #minetest-dev
02:37 sofar I've tested that patch for a few days now in various parcours
02:37 sofar so it's a firm +1 from testing and idea wise (sneak ladder for all I care never existed)
02:38 sofar code wise I scanned it pretty thoroughly, but, there are far better core reviewers out there than me :)
02:41 Taoki joined #minetest-dev
03:42 VanessaE 2017-03-15 04:42:20: ERROR[Main]: ServerError: AsyncErr: ServerThread::run Lua: Runtime error from mod '' in callback node_on_dig(): /usr/share/minetest/builtin/game/item.lua:499: attempt to index local 'def' (a nil value)
03:42 VanessaE all I did was dig an unknown node.  by hand.
03:45 VanessaE http://pastebin.ubuntu.com/24180642/
03:47 VanessaE https://github.com/minetest/minetest/commit/d785456b3fa35faf47cb972fde9e8668382c5e22#diff-83bb4564c6fbf230569bdeb2becc32b9L454
03:47 VanessaE this may be the breaking change
03:58 sofar oh, dug an unknown node?
03:58 VanessaE yeah.
03:59 VanessaE [03-14 23:49] <cheapie> VanessaE: For now, changing line 499 to "if def and def.after_dig_node then" should fix it.
03:59 sofar wow you got pretty far in that routine
04:00 sofar no, line 507 will make it crash too
04:00 sofar I'll see if I can come up with a fix in a bit after doing a review
04:00 VanessaE ok.
04:01 VanessaE RBA once advocated for a sort of "graduated" check for stuff like that.  his contention was that `if a.b.c.d` should automatically check a, then a.b, etc.
04:04 DI3HARD139 joined #minetest-dev
04:04 VanessaE this would appear to be the cause of the other recent crashes, by the way.  ones where I thought it was some random mod in my install.
04:07 sofar you know, I think lua could actually do that
04:08 VanessaE oh?
04:14 sofar #5398
04:14 ShadowBot https://github.com/minetest/minetest/issues/5398 -- Fix two nul deref if digging unknown nodes. by sofar
04:15 VanessaE shouldn't those be "nil ref"?
04:15 VanessaE (and, lgtm)
04:15 VanessaE er nil deref*
04:15 sofar shrug
04:19 VanessaE thanks again
04:20 Hunterz joined #minetest-dev
04:29 twoelk left #minetest-dev
04:34 octacian joined #minetest-dev
04:42 octacian sofar: BTW, I just updated game#1636 to return the stack from book_on_use, you might wanna take a look at https://github.com/minetest/minetest_game/pull/1636#discussion_r106084878
04:42 ShadowBot https://github.com/minetest/minetest_game/issues/1636 -- Fix books backwards compatibility issues by octacian
04:44 octacian sofar: in game#1597 do you mean that I should move keys entirely to a new file or just as I did there?
04:44 ShadowBot https://github.com/minetest/minetest_game/issues/1597 -- Allow skeleton keys to be stacked by octacian
04:45 sofar one commit for the code change, one commit for the function move to a different file
04:45 sofar now it's just a giant add/remove patch and I can't see the functional changes
05:00 octacian joined #minetest-dev
05:00 octacian joined #minetest-dev
05:02 sofar octacian: I have a patch pending to make the key use left-click as well
05:02 octacian ok
05:02 sofar maybe you can push on top of my pr?
05:02 octacian I assume the conflict?
05:02 sofar yeah, it changes on_place to on_use
05:02 octacian So.. 3 commits?
05:02 octacian wait, nvm
05:02 octacian >_< brain isn't working great, it's late lel
05:02 octacian I'll see what  Ican do.
05:03 sofar only 2 for you :)
05:03 octacian tomorrow tho
05:03 sofar no sweat, I'm not rushing tonight
05:03 sofar reviewing some minetest-mods open stuff atm
05:03 octacian ah, I see
05:04 Thomas-S sofar, could you maybe do me a favor and also review game#1580? Thanks in advance!
05:04 ShadowBot https://github.com/minetest/minetest_game/issues/1580 -- Change doors to default.can_interact_with_node() by Thomas--S
05:05 octacian Personally I've been working on mods a lot lately - haven't had time to update my PRs
05:35 cheapie In regards to #5393, while sending client-side mods to the client is an excellent idea, the user should *not* be forced to accept any particular mod, nor should they be prevented by the server from running whichever client-side mod they want to load, as no server should have the authority to decide what I can or can't run on my own computer.
05:35 ShadowBot https://github.com/minetest/minetest/issues/5393 -- CSM: Ability to inject from server-side mod
05:36 cheapie As far as security/anti-cheat is concerned, well, anything critical to that shouldn't be running (exclusively) client-side anyway.
05:38 lumidify joined #minetest-dev
05:57 VanessaE cheapie: my contention is that the server should send all code the client should need to run.  If the user wants to block it ala javascript in a web browser, I see no problem with that, but the server should be able to refuse any messages/data coming back from mods that it didn't send to the client.
06:06 Hunterz joined #minetest-dev
06:18 sofar VanessaE: I have no indication that people are not in violent agreement with you
06:19 sofar I also have no problems with developers goofing around with client side mods on their own, without a server sending out said code
06:19 sofar in practice, almost nobody will end up doing that
06:20 sofar everyone will rely on what the server sends
06:20 VanessaE good
06:20 VanessaE that's exactly what I would aim for.
06:21 VanessaE one problem MC users got into, as I understand it, is incompatibility between servers and the mods they need installed on the client.  server A wants mod B, server C also does, but server C relies on some old version of mod B for, oh, say legacy reasons
06:22 VanessaE what do MC users do in that situation?
06:23 VanessaE (serious question, as I've never used it)
06:23 sofar I was quite deep in bukkit modding, but I gave up on client-side modding for MC
06:23 sofar it's just not worth it
06:23 sofar so much crapware
06:23 VanessaE heh
06:32 sofar plus every dang server or mod maker wants you to use a different crap tool to patch minecraft.jar
06:32 sofar I quickly realized after patching 2-3 times that I never wanted any of it
06:35 sofar I still believe that client-side mods have a very limited function coming to minetest
06:35 sofar but the ones they do have will help a lot to offload things from the server
06:36 VanessaE now, tell me something...
06:36 VanessaE why such a hard-line stance on client-side mobs?
06:36 sofar what part is hard-line?
06:36 sofar oh, the mob thing
06:36 VanessaE [03-13 13:17] <sofar> oh fuck no
06:36 VanessaE [03-13 13:17] <sofar> mobs will never be client side
06:37 sofar yeah, come on, think about it
06:37 sofar first of all, you can't just eliminate mobs from the server and entirely put them on the client
06:37 VanessaE I had figured there would be some cooperation between the two
06:38 VanessaE damage prediction and the like
06:38 sofar second, no matter how hard you try to offload tasks to the client, it's just always better lag-wise on the server
06:39 sofar what I can think of is something I coded up in entity_ai myself
06:39 sofar basically, the server sends mobs out for goals that are ~ 2.0 seconds away
06:39 sofar and in between we don't really care
06:40 sofar but tbh you almost don't need csm for that
06:40 sofar although it would allow lua-coded pathfinding on the client
06:40 sofar and that's kinda nice
06:40 sofar but the server still would have to discover whether those paths are legal
06:40 sofar e.g. the mobs don't go though a door
06:40 VanessaE consider pipeworks as another example where I'd think the client could help - namely the animations of the items flowing through tubes.  when the server lags, they float out of the tubes because the client can't stop them from moving due to lag.
06:41 sofar that's a good example
06:41 VanessaE client-side pathfinding might be useful for that, now that you mention it
06:41 sofar clear start and stop state
06:41 nerzhul joined #minetest-dev
06:41 sofar so the client can just "remove" the entity when it's at the right point
06:41 VanessaE carts have a similar problem to tubed items - they tend to float right off of the track during lag
06:42 sofar yes but at some point the server needs to know where the damn cart is
06:42 VanessaE yep, I know
06:43 sofar plus it all needs to add up or else anticheat fails
06:43 VanessaE server says "move in direction <vector> at speed <velocity>" but it can't tell the client "stop the entity if it crosses within <coords>"  (of course, this doesn't need CSM)
06:44 sofar well the server could predict a curve ahead
06:44 sofar but what if the rider stops the car?
06:44 VanessaE that would be tricky to solve, I'll grant you
06:44 sofar also current carts are not physical, which I really don't like
06:44 VanessaE not physical?
06:44 sofar one quirk and you fly through stones
06:44 VanessaE oh yeah
06:45 VanessaE or they drop off of the track on hills sometimes (or used to)
06:45 sofar I'd rather see them physical and buggy
06:45 sofar it's fine if they derail even
06:45 VanessaE so maybe for carts and pipeworks, the server needs to be able to tell the client "here's a list of vectors and velocities.  play it out and delete the entity when done"
06:46 VanessaE (of course the server will have to play the list as well)
06:46 VanessaE point being to decouple server responsiveness from client smooth animation
06:46 VanessaE (or network responsiveness)
06:48 sofar see the problem with all this is
06:48 sofar now we've reduced the scope of mods on the client to ... just about a shell of what mobs actually need
06:49 sofar and we're not even talking smart stuff yet
06:49 sofar scriptable mobs
06:49 sofar npcs that can perform more complex functions
06:50 sofar plus, when it comes to complexity, you're talking about the most complex thing to solve
06:51 sofar it has synchronization issues
06:51 sofar it has cheating problems
06:51 VanessaE hrm
06:51 sofar lag is a problem
06:51 VanessaE I can see your point.
06:51 sofar lol
06:51 sofar I'm not saying it's not possible, but, wow, it's going to be a looooong journey before we can even start on that
06:51 VanessaE that's fine too
06:51 VanessaE I'm not saying "let's dump it all into the API right away"
06:52 VanessaE just more a matter of "let's have a good plan that'll still work 5 years down the road" etc
06:52 VanessaE (even if the whole thing is done in 6 months and stale in a year :) )
06:52 sofar personally, I have far more important features on my radar that I'd like to see progress
06:53 sofar node setmeta, we need that
06:53 sofar facetext, we need that
06:53 VanessaE yeah.
06:53 nerzhul lag will be reduced when client mods will be really used by modders to port their functions to client instead of server, example a craft guide has nothing to do server side :)
06:53 sofar colored itemstacks, we need that
06:53 sofar I need to work on entity_ai, so we can get rid off all the buggy mob mods
06:53 sofar and get truly scriptable AI mobs
06:53 VanessaE insert reference here to XKCD "14 standards" cartoon :P
06:54 nerzhul sofar, i have a pure C++ part not perfect on my fork, based on mobs redo
06:54 sofar nerzhul: excellent example
06:54 sofar nerzhul: forget mobs redo
06:54 sofar it's kids' stuff
06:54 sofar it's not even intelligent code
06:54 nerzhul sofar, yes i know, i rewrote many parts
06:54 sofar my entity_ai solves the problem at a much more fundamental level
06:54 sofar splitting code from decoration
06:55 sofar making code pluggable
06:55 nerzhul sofar, look at that: https://gitlab.com/epixelgame/epixel/blob/master/src/contrib/creature.cpp
06:55 sofar fixing the stupid memory problems that most lua mobs have since they stuff kilobytes of code into instanced objects (dumb)
06:56 sofar nerzhul: again, it's cool it's C++ but it's no different from a mob template
06:56 sofar we need *pluggable* and *scriptable* mobs
06:57 nerzhul sofar, a right and good API is mangos emulator API
06:57 nerzhul i like it very much
06:57 sofar I can do that in entity_ai
06:57 sofar I just need to add a spawnpoint driver
06:57 sofar and a waypoint driver
06:57 sofar and then you can do that with entity_ai
06:57 sofar btw drivers can be registered in a separate mod in entity_ai
06:58 sofar same with pathfinders
06:58 nerzhul sofar, exactly what the mangos API does, and it has castspell you add your itmers , you set go to pos, attack, cast spell, wait, speak :)
06:58 sofar I've coded quests in mangos
06:58 nerzhul sofar, i coded all ulduar, ICC
06:59 nerzhul beofre they were in SD2
06:59 sofar offtopic, but, lmao, I run a BC server at my house
06:59 nerzhul https://github.com/nerzhul/MangosFX/tree/master/src/bindings/ScriptDev2/scripts/northrend
06:59 sofar also entity_ai has arbitrary animation support
06:59 nerzhul i was youg :p
06:59 nerzhul and also i was a C++ noob at this moment ^^
06:59 sofar so if you want a mob to have special animations? no problem
06:59 sofar special code drivers? no problem
07:00 sofar special state transitions? no problem
07:00 sofar npc that changes model into a demon mob? it can be done
07:00 sofar npc that you can trade with that then builds your house? totally doable
07:00 nerzhul sofar, an event api on a npc class is the good way to do it i think, like in SD2
07:00 sofar no need to code up low level bits, I have all those
07:01 sofar just code your "npc trade formspec" and hook it to an event
07:01 sofar and code the state transitions
07:01 nerzhul sofar, go then
07:01 sofar nobody is looking at entity_ai, so, I'll pick it up when I finish what I'm doing now
07:01 nerzhul i hope your entity AI is pure C++
07:02 sofar of course not
07:02 sofar prototype it in lua
07:02 sofar then make it solid and test
07:02 nerzhul prototype in lua is good, after convert it :)
07:02 sofar and then optimize in C++
07:02 nerzhul i need to spawn 2000 mobs :D
07:02 nerzhul but if they where in one place on client => massive UI lag :p
07:04 sofar anyway, I'm a pumpkin
07:04 sofar *poof*
07:04 * sofar is off to bed
07:04 nerzhul sofar, merging 5398
08:18 Fritigern joined #minetest-dev
08:25 Lunatrius joined #minetest-dev
08:30 red-001 joined #minetest-dev
08:44 nrzkt joined #minetest-dev
08:46 proller joined #minetest-dev
10:15 kilbith joined #minetest-dev
10:37 nrzkt joined #minetest-dev
10:50 YuGiOhJCJ joined #minetest-dev
11:03 Fixer joined #minetest-dev
11:03 Player_2 joined #minetest-dev
11:41 proller joined #minetest-dev
12:28 xerox123 joined #minetest-dev
12:41 lisac joined #minetest-dev
12:47 red-001 joined #minetest-dev
13:24 proller__ joined #minetest-dev
14:17 Darcidride joined #minetest-dev
14:41 Wayward_One #5400
14:41 ShadowBot https://github.com/minetest/minetest/issues/5400 -- (Android) Fix undefined references by Wayward1
14:42 nrzkt Wayward_One: another time i missed android, i need to add an android builder on travis at a point to show that :)
14:42 nrzkt thanks for the point, i'm merging it now as it's fairly trivial
14:43 Wayward_One :) that would be useful
14:43 Wayward_One np, glad to gelp
14:43 Wayward_One *help
14:45 QwertyDragon joined #minetest-dev
14:57 octacian joined #minetest-dev
14:57 octacian joined #minetest-dev
15:10 Hunterz joined #minetest-dev
15:49 OldCoder joined #minetest-dev
15:50 AnotherBrick joined #minetest-dev
15:57 paramat joined #minetest-dev
16:39 proller__ joined #minetest-dev
16:48 Hunterz joined #minetest-dev
16:56 kilbith joined #minetest-dev
17:01 paramat #5382 is ready for review
17:01 ShadowBot https://github.com/minetest/minetest/issues/5382 -- Mgv7: Add optional giant caverns at depth by paramat
17:17 cheapie [00:57:38] <VanessaE> cheapie: my contention is that the server should send all code the client should need to run.  If the user wants to block it ala javascript in a web browser, I see no problem with that, but the server should be able to refuse any messages/data coming back from mods that it didn't send to the client.
17:17 cheapie ^ That's pretty much exactly what I meant, yeah.
17:19 VanessaE I'm glad we all agree on that :)
17:22 Krock joined #minetest-dev
17:23 sofar violently
17:24 sofar but, as I said before, it's up to the server operators to implement that part
17:24 sofar (we'll make the tools and interfaces for them to do so, of course)
17:31 VanessaE sofar: imho, server and client should both default client-receives-mod-code.  i.e. blocking needs to be an explicit action by the user.
17:32 red-001 agreed
17:32 VanessaE because it's easier to tell the paranoid to disable a feature than to get the average user to enable it
17:32 red-001 there should be a setting like enable_mod_auto_download
17:32 red-001 set by default to true
17:32 sofar VanessaE: I don't think anyone disagrees with that?
17:32 VanessaE enable_remote_content
17:33 VanessaE sofar: ok, just making sure.  I kinda got the impression that some folks here would prefer opt-in.
17:33 sofar no, if we can't enable that by default then I consider CSM failed
17:33 VanessaE ok good
17:33 red-001 also when possible mods supplied by the server should over-ride the ones supplied by the client
17:34 VanessaE agreed.
17:49 sofar just like worldmods override non-worldmods
17:49 VanessaE yes
17:58 paramat Krock were you not invited to be a dev? still considering?
17:59 Krock paramat, please check out the minetest team again. I'm already a dev ;)
18:00 paramat kewl
18:01 paramat feel free to alter the 'one approval' 'two approvals' labels on PRs you approve
18:06 Hijiri what do people think about including a monoid-like interface in the lua API for physics overrides, armor values etc.? The current interface just begs to break compatibility between mods
18:07 Hijiri With a monoidish interface mods will behave the same as they do now when they are operating alone, and when operating together they will at worst create balance problems
18:07 Hijiri rather than stepping  on each others' toes
18:07 Hijiri of course the actual API doesn't have to mention monoids or anything
18:09 Hijiri The interface would have something like contribute_value("modname:contribution_key", 0.5) for adding effects, and uncontribute_value("modname:contribution_key") to remove an effect that was added with the key
18:09 Hijiri To calculate the actual physics override, or armor value, or whatever, all the contributions from every mod is multiplied together and applied
18:10 Hijiri If you only have one mod using it, it would be as if it used the current override interface
18:11 Hijiri If you have two mods doing it, their values are combined gracefully together rather than having something like a race condition where only the last mod to apply the thing gets the effect
18:12 Hijiri maybe I'll post a Github issue later
18:12 Hijiri since I have class
18:16 paramat Krock here's the correct way to merge PRs from terminal https://gist.github.com/paramat/fdb80b6f03923d444af261a789894215
18:17 Krock paramat, thanks for the introduction but I already know how to prevent the merge commits with my git gui ;)
18:18 Krock I'm basically doing the same - just using a graphical interface
18:19 Fixer hey, I suddenly realised, if there are so much prosneak, and it still needs fixing, why not introducing something like a rope (already available in some mods) for fast travel
18:22 sofar Krock: ah, good, then I can merge some stuff :)
18:23 paramat ok good
18:23 sofar curl -L http:/..../1234.patch | git am
18:23 paramat Fixer i've suggested modders work on alternatives
18:23 sofar but, squash/rebase buttons on github are fantastic, too
18:25 paramat if reasonable such a mod could be added to game
18:25 Fixer paramat: there is already rope mod somewhere on vanessas servers (for climbing)
18:26 paramat just not 6n/s speed unless magickal or mechanical
18:26 VanessaE Fixer: castles mod provides one, and vines provides another.
18:27 VanessaE (both suck)
18:27 VanessaE castles *modpack* which I don't yet use, provides what looks like a pretty good rope block
18:27 VanessaE might be a good idea to coordinate with FaceDeer about that.
18:27 paramat i'll test the sneak PR and might +1 it later
18:28 juhdanad joined #minetest-dev
18:31 juhdanad Hi! Could you explain me how can you store mutiple things in an itemstack metadata?
18:39 Fixer paramat: i have intriguing question, what if "get biome info" can get status of the terrain? like land/sea/river/dungeon ?
18:39 Fixer imagine river protector mod, blocks any building on rivers for example
18:39 Fixer just based on generated structures
18:45 juhdanad To be more specific: how can I push a 'color' property to an itemstack?
18:46 paramat Fixer that would be complex, it would have to analyse the noise parameters for the specific mapgen
18:58 red-001 joined #minetest-dev
19:05 nerzhul joined #minetest-dev
19:10 Krock #4925 became superflous due to #5320 right?
19:10 ShadowBot https://github.com/minetest/minetest/issues/4925 -- Fix block drawing glitch at block boundaries by Rogier-5
19:10 ShadowBot https://github.com/minetest/minetest/issues/5320 -- Allow server side occlusion culling. by lhofhansl
19:14 Fixer don't think so
19:22 lisac joined #minetest-dev
19:35 Krock Merging works fine as always :)
19:37 proller__ joined #minetest-dev
19:37 Krock It it desired to announce merges with >= 2 approvals here too?
19:41 sfan5 yes
19:42 Krock alright, will do that in the future
19:52 Warr1024 joined #minetest-dev
19:52 paramat yeah because if 2 devs do merges simultaneously it could cause a problem
19:53 paramat nope 5320 did not fix 4925
19:55 paramat but we now have a setting prevent that bug happening too close to the player #4686
19:55 ShadowBot https://github.com/minetest/minetest/issues/4686 -- Remove optimization that causes underwater and cave rendering glitches. by lhofhansl
20:02 red-001 joined #minetest-dev
20:04 Tmanyo joined #minetest-dev
20:21 twoelk joined #minetest-dev
20:48 kaeza joined #minetest-dev
20:53 paramat joined #minetest-dev
20:56 octacian_ joined #minetest-dev
21:02 Fixer nerzhul: from forum: client-side skins?
21:02 nerzhul Fixer, it was requested but does those needs to be sent to server ?
21:02 Fixer no idea
21:03 nerzhul a texture pack client side okay, a skin sent to server no, due to some who can send nude/nazi etc symbols using this
21:03 twoelk what woeld client side skins be?
21:03 nerzhul i prefer a server owner formspec permitting to show it and guarantee correct usage from players
21:04 nerzhul nore, Zeno` sfan5 paramat can you look at #5396 please ? Krock already looked at it and i fixed the mentioned point, it's jsut a simple portage to client
21:04 ShadowBot https://github.com/minetest/minetest/issues/5396 -- [CSM] Add ModStorageAPI by nerzhul
21:04 twoelk if I change the textures client side only I can see them, so again, what would clm skins be?
21:05 twoelk csm
21:05 nerzhul if you want like in some games see al players nude it's your problem, but sharing it with other players is a bit risky
21:05 nerzhul we have child in our userbase
21:06 twoelk like I change all players to pandas but only I can see that?
21:06 nerzhul y
21:07 twoelk I don"t quite see the use
21:07 nerzhul merging #5311
21:07 ShadowBot https://github.com/minetest/minetest/issues/5311 -- lua_api: Document minetest.features by SmallJoker
21:07 twoelk I would rather have a better way to communicate my skin wish to the server owner
21:08 twoelk but the skin as such stays  served from the server owner
21:10 twoelk actually I'm not really sure what csm is - it looks to me as if two concepts are mixed at the moment
21:10 nerzhul CSM is a pure client side modding
21:10 nerzhul atm you cannot do many things because some more handlers are required
21:11 twoelk the first beeing mods provided by the server but run in part on the client
21:11 nerzhul #5399 is a pure CSM interesting usage
21:11 ShadowBot https://github.com/minetest/minetest/issues/5399 -- [CSM] Add minimap API modifiers by nerzhul
21:11 twoelk the second some mod I install locally and only I have use of
21:11 nerzhul for first phase, we will permit to customize client without changing the game rules, which are owned by server
21:11 nerzhul it's the safer way to start, after thinking properly, extend the CSM with some communication to serve,r but after having a real client API
21:12 red-001 could someone review #5113 ?
21:12 ShadowBot https://github.com/minetest/minetest/issues/5113 -- [CSM] Add function to get the name of the player and functions to colour chat. by red-001
21:12 nerzhul red-001, can you drop the player part as we should change this to a pure client luaobjectref ?
21:13 red-001 so remove the name getter function?
21:13 red-001 ok
21:13 nerzhul it's better to have a player object in the lua stack than creating many getters
21:14 red-001 yeah I was thinking the same thing
21:15 twoelk a pure client side mod would be one the server knows nothing about and doesn't care - I'm not sure that potential godd and bad has been studied yet
21:17 nerzhul don't burn steps, we need to add API, when there will be sufficent API client side we will have new things to develop to enhance it
21:18 nerzhul if i remember crafts are sent client side, now imagine we have an API which listen key bindings, you open your craftguide by pushing C on your keyboard, it's pure client side :)
21:19 twoelk I understand that part (version one of my statement above)
21:20 twoelk but client side skins would be version two
21:20 twoelk just like texture are at the moment
21:20 red-001 done
21:21 red-001 #5113
21:21 ShadowBot https://github.com/minetest/minetest/issues/5113 -- [CSM] Give CSM access to core.colorize by red-001
21:21 twoelk For example I could imagine a client side only mod to keep personal notes - such as my waypoints or other lists
21:22 nerzhul twoelk, yes it's an interesting idea
21:22 twoelk this might interfere with game design though
21:22 nerzhul atm mod storage is shared between servers, we need two storages client side at the end, one global and one per server
21:23 nerzhul twoelk, the mod principle is to make game enhanced. After maybe we should add a mod blacklist for permitting servers to disable some client mods, but i'm not sure it's a good idea
21:23 twoelk between servers?
21:23 red-001 I think we might want to consider limiting certain functions to server supplied mods
21:24 twoelk I could have a clientside minimap mod that queries my local cached map data only, never exposing itself to the server - close to cheating
21:25 twoelk it could highlight certain nodes
21:27 nerzhul twoelk, yes between servers, like for example keymapping, it's a personal user preference not depending on server
21:30 twoelk will the client side mod code be downloaded fresh everytime I join or will the cache be searched for allready downloaded versions?
21:31 twoelk I could for example manipulate cached content
21:31 nerzhul twoelk, sorry we are at early stages, don't search us to have answers to all your questions, CSM will be in heavy development
21:31 nerzhul first CSM controled by client, second extend to server or client peering
21:32 twoelk I have a feeling many decisions may have to be made quite early - especially about the scope and security
21:33 nerzhul for CSM security is priority, it's why first stage is pure client side
21:34 nerzhul yes it doesn't allow to do crazy things, but i permits to ensure a correct and secure user experience
21:34 nerzhul it's also why we don't added get_mod_path at this moment, because i don't see any interest to know the mod path as a mod is a runtime and if it needs to write things there will be mod storage
21:38 twoelk hm, about local scins example. how would that work? load all local scins at worldstart?
21:38 twoelk *skins
21:40 twoelk I wonder wether this policy doesn't clutter the client memory with media too much
22:06 paramat updated and tested #5388
22:06 ShadowBot https://github.com/minetest/minetest/issues/5388 -- Map gen limit: Rewrite by paramat
22:10 paramat will merge game#1579 game#1621 game#1635 game#1639 in a moment
22:10 ShadowBot https://github.com/minetest/minetest_game/issues/1579 -- Migrate sethome mod to player attributes. by sofar
22:10 ShadowBot https://github.com/minetest/minetest_game/issues/1621 -- sapling_on_place: Restore default behaviour by SmallJoker
22:10 ShadowBot https://github.com/minetest/minetest_game/issues/1635 -- Skeleton key: Change to use left-click (tool on_use) by sofar
22:10 ShadowBot https://github.com/minetest/minetest_game/issues/1639 -- Farming: Add override for default:dirt_with_rainforest_litter by paramat
22:14 red-001 joined #minetest-dev
22:26 Taoki joined #minetest-dev
22:26 paramat merging ..
22:27 paramat 1000 commits!
22:28 kaeza is it just me, or the topmost leaves node of mapgen-placed aspen trees just fails to decay?
22:28 twoelk|2 joined #minetest-dev
22:28 kaeza haven't tested aspen trees grown from saplings
22:28 paramat lol i managed to get number 1000
22:30 paramat i raised the trunk by 1 node in the schematic, ahh perhaps the already-placed aspens are as they were before, so the new leafdecay radius 2 is not large enough for them
22:30 red-001 I wonder who will get 1024
22:31 sofar paramat: lol, yes, that would do that
22:31 sofar kaeza: time for timberrrrr!
22:32 paramat might have to raise the radius to 3 again for a few years
22:33 kaeza BTW, yesterday I noticed my cactus farm was not growing at all except for a single one. turns out I placed all cactus sideways :I
22:33 kaeza </offtopic>
22:33 nerzhul yeah CSM with minimap now works, jsut need to push and document :)
22:36 paramat merge complete
22:36 paramat down to 22, hope we can stay below 25
22:40 Fixer kaeza: you play minetest?
22:42 Fixer innocent question, why put into /clientmods but not in /client/mods ?
22:44 nerzhul if i remember the choice was done because of ~/.minetest/client
22:45 nerzhul to make mods independent of this already used namespace for storing persistent data
22:45 nerzhul sofar, i need you !
22:45 sofar I need you too nerzhul, but not like that :)
22:45 kaeza jeez, get a room you two
22:46 nerzhul sofar, ? :o
22:46 Fixer nevermind
22:46 sofar hahaha
22:47 sofar well tbh client/serverlist is kinda weird
22:47 sofar there's just one file in there
22:47 nerzhul sofar, can you review #5396 & #5399 for permit merge ?
22:47 ShadowBot https://github.com/minetest/minetest/issues/5396 -- [CSM] Add ModStorageAPI by nerzhul
22:47 ShadowBot https://github.com/minetest/minetest/issues/5399 -- [CSM] Add minimap API modifiers by nerzhul
22:48 sofar nerzhul: tonight (it's work-work time here atm)
22:48 TheReaperKing joined #minetest-dev
22:48 nerzhul sofar, work for minetest :p
22:48 nerzhul what time is it in your country ?
22:49 nerzhul sofar, i introduced core.ui namespace client side for ui coreside components, minimap reference is core.ui.minimap, what do you think about it ?
22:49 nerzhul as many global UI components are singleton (console, minimap, chat for example)
22:50 sofar 4pm here in west-cuckoostan
22:55 paramat wow mtgame has had 500 of those commits since 0.4.13
22:57 sofar nerzhul: honestly you can move the entire HUD to lua for all I care
22:58 sofar nerzhul: take it out of the server mod realm
22:58 sofar just make it client only
22:59 Fixer at the time i considered 0.4.12-13 pretty depressing time in mt history
22:59 nerzhul sofar, xD, crazy but not entierely possible
23:00 nerzhul Fixer, why ? not many commits ?
23:00 Fixer yes, stagnant development and very few devs
23:00 Fixer somewhere 0.4.14-0.4.15-dev it became great again
23:01 nerzhul Fixer, i looked at 0.4.12 it's before my network refactor
23:01 nerzhul look at this ugly thing : https://github.com/minetest/minetest/blob/0.4.12/src/client.cpp#L1465
23:01 nerzhul if i remember i started to develop between 0.4.12 and 0.4.13 on mt
23:02 nerzhul MT code has made many improvements, especially in terms of performance, semi C++11 containers, many memory copy removal (not all but many anoying)
23:02 nerzhul network packet handling performance using functions pointers instead of a huge if/elseif
23:02 nerzhul ABM vector
23:07 sofar 0.4.15 will be a lot smoother gameplay too
23:07 sofar 0.4.16* will be a lot smoother gameplay too
23:11 nerzhul sofar, yeah, many commits are improving server load
23:12 nerzhul + mod meta + player attributes, for better mod storage :)
23:13 nerzhul mod security
23:13 Fixer there will be jitter by default, but that can be solved before everybody dies
23:13 Fixer good releases indeed
23:13 Fixer very fun times
23:13 Fixer I feel like I'm back in 2011-2012
23:13 nerzhul no more breath cheat
23:14 nerzhul 2D sheet animations
23:15 nerzhul and disabled csm :p
23:17 kaeza 2D sheet animations?
23:18 kaeza oh you mean the old 0.3-style DM? :P
23:18 nerzhul don't know, ask to sfan5
23:25 paramat +1 for sneak PR, will merge later
23:27 red-001 could someone else review #5113 ?
23:27 ShadowBot https://github.com/minetest/minetest/issues/5113 -- [CSM] Give CSM access to core.colorize by red-001
23:27 red-001 all it does is move around some code
23:28 red-001 builtin/game -> builtin/common
23:32 Tmanyo joined #minetest-dev
23:35 Puka joined #minetest-dev

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