Time Nick Message 01:53 hmmmm okay, approved 02:00 yunfan so anyone can explain me the lag on local server? 02:01 yunfan i want to know which reason caused so much lag 02:06 est31 celeron55, we could add a pagage to the ppa to override the memory using the ldpreload trick. server owners who want to use it could install it. 02:06 est31 but I don't think it should be included in official packages, because I've heard that luajit's garbage collector doesn't handle such large piles, so the limit is there for two reasons, solving one doesnt help alot. 03:07 paramat ok rebased will merge 3418 after checks 03:10 yunfan hey, could minetest.register_chatcommand accept chatcommand that take paramaters? 03:11 yunfan like /sethome blah 03:15 paramat yes i think so, that's a subject for #minetest channel not here. now merging 3418 03:18 yunfan what? so i should go to #minetest for mod relavant questions? 03:18 yunfan i thought that was for player :[ 03:25 yunfan ok there is any question, is the lighting problem caused by the irr engine or just a configure problem? paramat 03:27 paramat yes mod stuff goes to #minetest. which lighting problem? we have several 03:28 paramat now merged, a kick-ass commit 03:31 yunfan paramat: the light effects looks too weak in night 03:32 yunfan campared to other block build game i have saw 03:35 paramat light spread is limited to 16 nodes, as in MC 03:35 paramat that's engine design, not irrlicht 03:44 yunfan then why i feel scene much lighter in MC than MT? 04:04 kaeza kinda important(ish): https://github.com/minetest/minetest/pull/3433 04:31 sofar well bummer, I just learned the hard way that you can't minetest.register_craftitem after the server has finished initializing 04:35 sofar I had hopes I could use it to make books appear in the inventory with custom descriptions for each book 04:42 sofar ohh wait, maybe not entirely 04:43 sofar oh wow I think that actually worked, somehow 04:52 sofar if only the server would send new craftitems to the clients 04:55 hmmmm please don't do this 04:55 sofar well, the alternative is that all the books are written before the server is started 04:56 hmmmm the alternative is subtle race conditions 04:56 sofar otherwise all in-game books have the same description 04:56 sofar a race condition? 04:56 hmmmm maybe you can't really notice it because of the current threading setup, but registering things after mod initialization is bad 04:57 hmmmm in fact, i should make the register_* functions error if it's called outside of initialization 04:57 sofar yeah, I realize after I wrote the code that its not going to work 04:57 sofar the core can't handle it 04:57 sofar but it exposes something I'd like to see changed 04:58 sofar "prefer" 04:58 sofar let's say I get that it's not meant to be right now 04:58 sofar but it severely makes my plans for having an interesting subgame with lots of books pretty much... impossible 04:58 sofar since every book will be ... identical from the inventory view 04:59 sofar unless I make all books ahead of time, in which case it's ... well, that's just a cop-out ;^) 05:04 sofar hmmm: actually it's close to working: a second client connected to that server was able to grab the item from the creative inventory and sees the new name 05:04 sofar remember it's a crafitem, not a node 05:05 sofar hmmmm: would it be possible to make some sort of a "child craftitem" ? one that is mostly identical to the client but has a different metadata, description ? 05:06 hmmmm don't think so 05:06 sofar that would allow things like named gear items, enchanted weapons, books written during runtime, etc.. 05:06 hmmmm minetest is an open source game, so.. 05:11 sofar you know what, I think this could all be implemented in metadata 05:11 sofar have a custom_description entry in metadata 05:11 sofar client displays the custom_description in the metadata if present, otherwise just the regular description 05:12 sofar and then the books mod can handle the content 05:12 sofar that way you could literally rename anything in game 05:12 sofar and no new registration would ever take place 05:12 sofar since it's all just part of the metadata 05:20 sofar https://forum.minetest.net/viewtopic.php?f=5&t=13629 looks like I wasn't the only person with this thought 05:48 sofar ahh yes, that would totally work, the client creates the tooltips every time the formspec is opened 06:04 est31 in fact, i should make the register_* functions error if it's called outside of initialization 06:04 sofar reading backlog? 06:04 est31 ^ please do that 06:04 sofar agreed 06:04 est31 yes sofar 06:05 sofar but, help me modify https://github.com/minetest/minetest/blob/master/src/guiFormSpecMenu.cpp#L1489 too so it deserializes the itemstack metadata, fetches a custom_description field and displays that as a tooltip too 06:05 sofar lol 06:05 sofar I think it's all possible 06:05 sofar and it wouldn't be a protocol change 06:06 sofar custom craftitem names! 06:06 est31 Name can already be customized, can't it? 06:06 est31 I would agree to the idea 06:06 sofar "description" is what shows in the tooltip 06:06 est31 just wondering what name it should have 06:07 sofar item name would declare a new name, or change the itemstack to a different item 06:07 est31 "custom_description" is good but its always better to get an overview over other used names 06:08 sofar right, that was a quick pitch 06:08 sofar we could just re-use "description" 06:08 sofar which would be even more clear to the modder 06:08 sofar naming items would become very easy - we could simply implement name tags 06:09 sofar punch a nametag in the air to attach a name to that nametag, then craft a nametag and a craftitem in the craft grid to transfer that to the craftitem 06:10 est31 well game designers could 06:10 est31 minetest shuould be more general 06:10 sofar right, I intend to use it to make named books appear with their name in book cases 06:10 sofar or to enchant armor and weapons 06:11 sofar I'm specifically not interested in naming mobs/entities 06:11 sofar or even nodes 06:11 sofar (which doesn't make sense to me atm) 06:11 est31 well yeah 06:11 est31 minetest tries to be generic 06:11 sofar I did read #1118 06:11 est31 making stuff easy equally for all 06:12 sofar I don't think it applies 06:12 est31 that doesnt apply 06:12 sofar one could "on_place" transfer the metadata to a node 06:12 sofar anyway, I think that's orthogonal 06:13 est31 the whole concept of how meta set nodedef is done is wrong IMO 06:13 est31 totally broken 06:13 sofar let's ignore that for this discussion ;) 06:14 est31 not a bad feature, but it should be done differently 06:14 est31 more efficiently 06:14 sofar not sure what's more efficiently than having the client parse the ItemStack metadata when they look at it in a formspec? 06:15 est31 well thats okay I think for our purposes 06:15 est31 so go on, implement a PR I agree to the general idea 06:15 est31 I now try to find out where the engine uses item metadata names 06:15 est31 bc lua_api.txt is silent about it 06:16 sofar well this is where I have an issue 06:16 sofar afaics nothing is using the ItemStack metadata in core 06:16 sofar so I have no examples how to use it 06:16 sofar plus 06:16 sofar there could be modules using it wrongly, since it's coded up to be a string 06:16 sofar and nobody forces serialization of a nice table 06:17 est31 "wrongly"? 06:17 sofar true 06:17 est31 what precisely do you mean 06:17 sofar according to the definition I've seen, ItemStack metadata is "just a string" 06:17 sofar so mods could put just that in there 06:18 est31 ah 06:18 sofar anyway, it's probably moot since if deserialization of metadata fails, it's obviously not serialized metadata and we can fall back to displaying the old description field 06:18 est31 well, then its a bit harder 06:18 est31 we'd have to have some nodedef flag 06:19 est31 and the flag encodes whether to interpret the stack metadata as table or not 06:19 est31 defaulting to "don't interpret it as table" 06:19 sofar right 06:19 sofar but let's see if we can cook something up that just ignores that for now 06:20 sofar I'd be interesting to see if I can actually cook something up 06:20 sofar (freakishly dangerous think to say at 10pm in the evening for me) 06:20 est31 yeah, for simplcicities sake you can just assume the whole metadata contains the description 06:21 est31 if its != "", then it should override 06:21 est31 later on that can be improved, but good for a first prototype 06:21 sofar for a test case, that's indeed enough 06:28 est31 merging in 10 minutes: https://github.com/est31/minetest/commit/9a5a538e8d1981fb3c2dd69ca3d37e4f9c426cf5 06:41 est31 pushed 06:41 est31 and bye again 07:21 asl97 can any core developers take a look at https://github.com/minetest/minetest/pull/3417 and finally merge this simple fix? 08:16 hmmmm wtf is this https://github.com/minetest/minetest/commit/a78dd7f2b6b0e1fefdbaa1ae21b722dd4459e4f4#diff-5c9fad38a1e2b7a0227fd3f5282dcc09R954 08:16 hmmmm who reviewed this shit? 08:17 hmmmm noise-based ore is now broken, thanks. 08:19 hmmmm BlockMen can fix it 08:38 est31 hmmmm, https://github.com/minetest/minetest/pull/3428 08:38 est31 three devs approved it, including me, and including a mapgen dev. 08:39 est31 mistakes are human, thanks for pointing this out. 08:39 est31 will push a fix for it 08:39 est31 technically blockmen didnt even voice approval 08:40 est31 he just merged it, he could argue that he only saw the two votes and then merged the PR 08:44 est31 this one PTAL: https://github.com/est31/minetest/commit/49b3fbc7eae31d8618af4f2447523a3b6aa4f7a 08:45 est31 (merging it in 10 minutes) 08:49 est31 asl97, I have already approved your PR, now a second dev has to approve it too 08:55 hmmmm this is an incredibly obvious mistake 08:55 hmmmm est, did you actually test your fix? 08:56 est31 hmmmm, how can I test it? 08:56 hmmmm also the commit message is misleading, it's not a "typo". this is a blatant error. 08:56 hmmmm see? wow you don't even test things before you commit them 08:56 hmmmm that's all I ask for 08:56 est31 I wrote "Fix threshold type" 08:56 est31 type not typo :) 08:57 hmmmm oh, nevermind 08:57 hmmmm "a recent commit to fix a typo has changed.." 08:57 hmmmm why not just put in the commit id? 08:57 est31 ok 08:58 hmmmm so anyway why did you miss this blatant error? 08:59 est31 https://github.com/est31/minetest/commit/51e8c2b27786c050f0271eeeaed5eea17d62f0a0 08:59 est31 well, idk, I've looked at it, didnt really check for the type 08:59 est31 I confirmed that it actually checked for the old value 08:59 est31 thought that it also warns when the deprecated way is used 09:00 hmmmm maybe some lua script unit tests should be added in 09:00 hmmmm register_* things can be as simple as running an embedded string and checking that the data structure is filled out as expected 09:01 est31 hrmm how can I test it now 09:02 est31 it builds, that works 09:02 hmmmm print out the ore structure built and verify it is what you specified in a mod 09:03 est31 lol 09:03 est31 https://github.com/est31/minetest/commit/a78dd7f2b6b0e1fefdbaa1ae21b722dd4459e4f4#diff-394cf121b2cc4085d04fc6487fdfadbfL102 09:03 est31 fixed a typo but didnt see the other obvious grammar mistake 09:04 hmmmm you, me, and others (kahrl too I think) all looked at that comment a zillion times and we always missed it 09:04 est31 but wont touch that comment, you promised to give a better one remember? 09:04 hmmmm yeah I guess. 09:04 hmmmm I'll do that 09:13 est31 okay its even tested now 09:13 est31 merging 15:43 sofar est31: https://github.com/minetest/minetest/issues/3435 15:43 est31 thx# 15:43 sofar I think we can iron this out... 15:48 est31 next step: add a flag to nodedef 17:43 paramat i missed the 'int nthresh' mistake, sorry =s 17:44 paramat i was the main approver of that PR 18:25 paramat nore sfan any comments? https://github.com/minetest/minetest_game/pull/735 'Default: Slightly reduce alpha of water post effect colour' 18:30 paramat next i would like to add a "mapgen_air" alias so that games can use a core mapgen but place vacuum or custom atmosphere nodes instead of air 18:34 sofar could minetest do this too: https://www.opengl.org/archives/resources/code/samples/mjktips/caustics/ ? 18:39 paramat maybe, as a shader. although (of course) personally i feel that's excessive realism :} 18:40 paramat RealBadAngel might be interested 18:42 sofar hahahaha 18:42 sofar well quake 2 had caustics 18:43 sofar I suppose you're right about excessive realism if quake 2 had that effect ;^P 18:46 paramat well i just dislike most shaders :) 18:48 est31 yeah that's paramat's taste... 18:48 est31 the water done by this guy looks really great 18:48 est31 lemme dig up the link 18:49 paramat talk to our shader guy RBA 18:49 sofar given the popularity of shaders for minecraft, I'd say I would give them a fair shot in minetest first 18:49 est31 https://www.youtube.com/watch?v=gVmtZITs3Ss 18:49 est31 and I dont even think he has caustics 18:49 est31 (cant really judge it, /me gfx noob) 18:52 Calinou well quake 2 had caustics 18:52 Calinou not that I know of 18:52 Calinou maybe some clients had, though 18:52 sofar right, I stand corrected, it had a screen wobble under water 18:52 est31 Calinou is quake expert 18:53 sofar it's been too long since I've played 18:53 sofar I was like half my age when that came out :^) 18:54 sofar well, close to 18:54 Calinou est31, yes, played all of them :p except Quake 4 18:54 Calinou sofar, Quake 3 had screen wobble, it's just a FOV change :P 18:54 Calinou easily doable in Minetest 18:55 sofar Calinou: I'd prefer better volumetric fog underwater and caustics 18:55 sofar the wobble always felt artificial, but a fov change may be nice 18:56 Calinou the problem is, we can't use modern OpenGL 18:56 Calinou and Irrlicht is not really made for that I guess, too 18:57 sofar but fog is already in the game... 19:18 Calinou volumetric fog probably isn't 19:22 Krock I was like half my age when that came out :^) 19:22 Krock Algebra told me that you're ~36 years old now 19:23 Krock err. this is #-dev 19:23 sofar close enough, 42 19:23 Krock ^^ 19:24 sofar right, help me implement https://github.com/minetest/minetest/issues/3435 ;^)