Time Nick Message 01:06 ANAND rubenwardy, re: #8736, Minimum tier should at least be getting 30 FPS. No one deserves less than 30. 01:06 ShadowBot https://github.com/minetest/minetest/issues/8736 -- Define and document minimum and recommended requirements 07:07 ANAND Also, you missed the (arguably) most important factor: screen resolution 07:08 ANAND I can turn all the dials to max. and still get 60 FPS, if my res. is 640x480 :P 07:37 cheapie ANAND: Heck, I can play MT with playable frame rates on a ViRGE/GX if I use a resolution of like 100x100 :P 09:28 ANAND cheapie: Exactly! That's why screen res. is an important (if not the most important) part of performance tests. 09:28 Krock well so travis cannot create files? 09:28 Krock Test assertion failed: world_config.updateConfigFile(path.c_str()) == true 10:04 p_gimeno maybe a permissions issue? 10:09 Krock git clone issue 10:09 Krock the directory is no longer created as soon world.mt is removed 10:10 Krock fixed that now 11:45 Krock ANAND: what should I change the comment to? 11:45 Krock it's correct 11:49 Krock sfan5: dynamic_cast: 0 occurrences. c-style cast: 12 matches in l_object.cpp 11:49 sfan5 well then don't 12:12 red-045 DS-minetest: you can't just load the server sent files into the same VFS as client CSM files? 12:13 DS-minetest red? Hi! o/ 12:13 red-045 Hello 12:14 DS-minetest I though, the "virtual filesystem" of minetest just makes real paths out of something like mymod:folder/file.lua 12:14 red-045 it doesn't, at least it didn't 12:15 red-045 The way it was implemented in the initial PR it was a mapping of all those paths to real paths 12:16 red-045 so that it wouldn't try to validate filesystem paths like server-sided mod security 12:16 sfan5 it should be turned into a VFS imo 12:16 red-045 like a real VFS? 12:16 DS-minetest sscsms should be kept in ram, right? 12:16 red-045 considering right now it's more a mapping 12:16 red-045 more of a* 12:17 sfan5 a mapping is not a vfs, so yes that's what I'm saying 12:17 red-045 A real VFS would be ideal 12:17 DS-minetest it would be possible to map the file buffers to the paths 12:17 red-045 DS-minetest: well there was some controversy about that, but I would go with yes 12:17 DS-minetest lua can load from buffers 12:18 red-045 that's what I did initially 12:18 red-045 but there was some concerns with wasting memory iirc 12:19 DS-minetest huh? 12:19 sfan5 you can always add an option to keep the vfs contents compressed 12:19 DS-minetest btw. should the mod files be deleted after mod loading? 12:20 red-045 well I can't answer that with a yes or no 12:20 red-045 depends on how you want to structure this code 12:20 celeron55 most files are compressed anyway 12:20 red-045 and how you want people to use CSM 14:05 red-045 on the topic of running lua in a sandboxed process it's a good idea, but not exactly easy 14:09 DS-minetest what are the difficulties? 14:12 red-045 how you implement communications with the main process 14:12 red-045 and I suppose making it cross-platform 14:44 rubenwardy Krock: added patch comments, now writing tests for noclip 14:47 rubenwardy oh so, Minetest doesn't support negative element positions? 14:47 rubenwardy for GUIs 14:47 rubenwardy makes testing a bit painful 14:47 rubenwardy hmmm 14:49 rubenwardy damn it 14:49 rubenwardy it limits it to 0,0 in container 14:49 rubenwardy that's so stupid 14:49 rubenwardy who did that 14:50 rubenwardy IT WAS ME 14:52 Krock https://i.kym-cdn.com/entries/icons/mobile/000/023/397/C-658VsXoAo3ovC.jpg 14:54 rubenwardy why does firefox lag when selecting things, it's so annoying 14:55 Krock perfectly smooth here, with about 1/3rd of your CPU's performance 14:57 rubenwardy it's weird 14:57 rubenwardy what OS are you on? 14:57 rubenwardy also, this is on lua_api.txt 14:57 Krock bun2 18.04.2 15:00 rubenwardy Krock: https://user-images.githubusercontent.com/2122943/62413517-b9381f00-b607-11e9-8286-45d20a3f75df.png 15:01 Krock so I can see the dropdown, checkbox, scrollbar, list and button text do not work properly 15:01 Krock ninja. already seen that 15:01 rubenwardy wait 15:01 rubenwardy I forgot to add the styles for those 15:01 rubenwardy but the item image button isn't clipping correctly on the left 15:02 Krock there was a PR recently to fix that 15:02 Krock #8652 15:02 rubenwardy updated: https://user-images.githubusercontent.com/2122943/62413545-1a5ff280-b608-11e9-87de-d3c620fd60d1.png 15:02 ShadowBot https://github.com/minetest/minetest/issues/8652 -- Fix missing item images clipping in formspecs by pyrollo 15:03 DS-minetest the text on the left still isn't clipped 15:04 DS-minetest (with #8740, the clipping of all elements would work fine) 15:04 ShadowBot https://github.com/minetest/minetest/issues/8740 -- Formspec draw order by DS-Minetest 15:04 rubenwardy am I forgetting elements? 15:04 rubenwardy because I went down my list 15:05 rubenwardy wait, the list only includes elements which have setNoclipEnabled(bool) methods in C++ 15:07 Krock even if, those can be added in a later PR as well 15:07 Krock what counts is that the current supported elements work fine 15:07 rubenwardy yeah 15:07 rubenwardy those are all tested 15:07 rubenwardy using the two formspec pages, and tracked using that todo list 15:13 rubenwardy I've rebased #7895 15:13 ShadowBot https://github.com/minetest/minetest/issues/7895 -- Add luacheck to check builtin by rubenwardy 15:21 rubenwardy merging trivial bug fix(?) in 15m: https://github.com/rubenwardy/minetest/commit/41ad0f4f202b47daab0841a56bf27c3990670344 15:39 rubenwardy huh, that was weird 15:39 rubenwardy my dedi wasn't responding to me, but IRC remained attached 15:39 rubenwardy but everyone timed out on the MT server 15:53 rubenwardy yay, green ticks 15:58 rubenwardy Krock: would you like to check #7895 again, and how should it be merged? 15:58 ShadowBot https://github.com/minetest/minetest/issues/7895 -- Add luacheck to check builtin by rubenwardy 15:59 rubenwardy I suggest something like: Add luacheck configuration, Fix Luacheck warnings in non-mainmenu builtin, Fix Luacheck warnings in mainmenu builtin 15:59 rubenwardy although, not sure on the value of that 16:07 rubenwardy I'll merge #8383 later tonight 16:07 ShadowBot https://github.com/minetest/minetest/issues/8383 -- Add formspec styling using a style[] tag by rubenwardy 16:08 rubenwardy the next big PR is scrolling_container 16:08 rubenwardy on the formspec block list 16:08 DS-minetest \o/ 16:08 DS-minetest but I've outsourced some parts 16:09 rubenwardy lol? 16:09 DS-minetest the reorder PR and the bgcolor PR 16:09 DS-minetest they are both from scroll_container and should be merged before 16:09 rubenwardy oh really? 16:09 DS-minetest (the scroll_container PR is too bloated) 16:10 DS-minetest yes 16:10 rubenwardy oh, were they previously contained by scroll_container? 16:10 DS-minetest yes 16:11 DS-minetest I did write that in the title comments 16:14 rubenwardy I thought it was the other way around - they depended on scroll 16:15 DS-minetest no, scroll ships all that changes, which is why it is so big 16:16 DS-minetest (I will rebase scroll_container after reorder is merged) 16:22 ANAND #8271 is ready for review again 16:22 ShadowBot https://github.com/minetest/minetest/issues/8271 -- Fix breath statbar scaling; defer breath_bar hiding by one second by ClobberXD 16:24 Krock testing 8271 16:33 ANAND Krock: Sorry, that was an unintentional line comment 16:33 Krock what? 16:35 ANAND My line comment in 8742 16:35 Krock ah. I rephrased the comment now anyway. I hope it's better than before 16:36 GreenDimond I asked this yesterday but it was sleeping hours for most of the devs. I modified GenericCAO::setChildrenVisible() to always do obj->setVisible(true) (just for testing), but the attachment still doesnt draw. Is there another check somewhere that overrides the visibility of the object according to the camera mode? 16:36 Krock the parent needs to be visible too 16:36 GreenDimond You might be the one that knows the most about this, Krock ;P 16:36 Krock ninjas 16:36 GreenDimond :0 16:37 Krock already tried that earlier btw 16:37 GreenDimond so I need to figure out how to not need the parent visible 16:37 Krock the attached node in front was drawn but the collision box never changed its position unless I switchde to 2nd/3rd perspective 16:38 Krock weird stuff 16:39 GreenDimond Should(tm) be simple enough to work around 16:39 Krock thought the same for the attachment fix PR and it took me days 16:40 Sokomine rubenwardy: at which position is #8680 on the list? it would allow for rpg-like chats with npc 16:40 ShadowBot https://github.com/minetest/minetest/issues/8680 -- Formspec: Add HTML-like element by kilbith 16:40 Krock everything was already sent as wanted, but the order to attach the stuff was completely messed up 16:40 rubenwardy Near the bottom lol 16:40 rubenwardy You can already have RPG like chats 16:41 rubenwardy Hypertext is massively overkill for that 16:41 Krock chats aren't that input friendly like buttons 16:42 rubenwardy To left align the text in a button, you can use a label and an invisible button 16:42 Sokomine how, rubenwardy? i know of no way that doesn't look awfully bad. current formspecs fail miserably at multiline-text 16:42 rubenwardy Invisible buttons created using FS styling 16:42 rubenwardy Not sure whether text areas go behind buttons 16:43 Sokomine hmm. is the size of the buttens sufficiently adjustable? keep in mind that the length of the texts may vary greatly. possible responses might range from "no" to a several lines long reply 16:43 DS-minetest text areas are normal elements, buttons, too, so ordering them should work 16:44 Sokomine #8680 isn't perfect because there's no identation (like in tables) - but it is almost perfect. "links" can be clicked, and multi-line text looks good 16:44 ShadowBot https://github.com/minetest/minetest/issues/8680 -- Formspec: Add HTML-like element by kilbith 16:45 Sokomine i don't see a chance of solving this mod-side with textareas. how should the mod know in advance which size to assign to a textarea? 16:47 rubenwardy Can you share a screenshot of what you mean by RPG like dialog? 16:47 Sokomine moment please 16:51 GreenDimond Krock: Do you know where the check is for player visibility is off he top of your head? 16:51 GreenDimond check for* 16:56 Krock camera.cpp and game.cpp 16:56 Krock both for CAMERA_MODE_* 16:56 Sokomine rubenwardy: https://forum.minetest.net/download/file.php?mode=view&id=19525 16:56 GreenDimond Thanks 16:57 ANAND game.cpp:L2877 16:58 GreenDimond ? 16:58 GreenDimond That line is self attachment 16:59 Krock using a debug build? not bad 17:00 GreenDimond Krock: Maybe you misunderstood.. I changed the function run from l_camera.cpp and game.cpp itself. (setChildrenVisible). It always sets visible to true. 17:01 GreenDimond Regardless of what is passed 17:01 GreenDimond Just for testing because I couldnt get it to display 17:01 GreenDimond And it still didnt draw 17:01 Krock https://github.com/minetest/minetest/blob/master/src/client/game.cpp#L2877 17:01 ANAND ^ 17:01 Krock the parent needs to be visible too 17:02 GreenDimond Is that an irrlicht thing or something? 17:02 Krock I suppose 17:02 ANAND Yes 17:02 GreenDimond Arg 17:02 Krock the function there will call node->setVisible(x) at some point 17:03 Krock maybe there's an option to not depend the visibility on parents? I didn't look into the docs for that functionality 17:06 ANAND See http://irrlicht.sourceforge.net/docu/classirr_1_1scene_1_1_i_scene_node.html#a2e3a88fe87d11caa7986a203afe6838c 17:07 ANAND and http://irrlicht.sourceforge.net/docu/classirr_1_1scene_1_1_i_scene_node.html#a1aef4f0feccebcd0d38373beadbc1827 17:07 ANAND setVisible and isVisible docs 17:07 GreenDimond Yeah im reading it rn 17:08 ANAND Both of them suggest that the parent has to be visible in order for the children to be visible 17:08 GreenDimond Perhaps I can manually add the node to the renderer 17:09 ANAND If this is about displaying player's attachments in first-person mode, one could attach the children to a dummy node instead of the player maybe? 17:09 GreenDimond I feel like thats a hack 17:09 ANAND It is :( 17:09 GreenDimond But if nothing else works, that would 17:10 GreenDimond Let me try manually adding the node to the renderer 17:10 ANAND But that dummy node might need to be attached to the player in turn. o_O 17:10 GreenDimond yuck x2 17:10 ANAND Stupid idea 17:11 ANAND Hmm, what if the player is visible? 17:11 GreenDimond Then you have a face in your face :) 17:12 ANAND The camera should be moved to where the "eyes" are, ofc 17:12 GreenDimond that is also a hack 17:12 GreenDimond One which shouldnt have to happen 17:12 GreenDimond And would mess up a lot of stuff 17:12 ANAND But that's how a camera is supposed to work ;) 17:12 ANAND But yea, it'd break some stuff 17:13 GreenDimond > thats how its supposed to work 17:13 GreenDimond Well, yes.. but actually no 17:13 GreenDimond ;p 17:13 ANAND That's literally the definition of first-person mode :) 17:14 ANAND Well, not literally 17:14 GreenDimond You arent supposed to work around the local mesh, you are supposed to hide it ;p 17:15 ANAND Noooo :'( 17:16 ANAND I want to be able to see my body when I look down :D 17:16 Lone_Wolf ^ 17:17 ANAND Camera offset can even be set from the Lua API, btw 17:17 GreenDimond Yes 17:21 ANAND Imma rotate myself by 90 degrees and align my head with a pillow 17:21 ANAND G'nite ;) 17:24 GreenDimond Hey it almost worked! 17:24 GreenDimond In first person mode it appears but doesnt update position x) 17:24 GreenDimond Progress! 17:42 rubenwardy merging #8383 in 30 17:42 ShadowBot https://github.com/minetest/minetest/issues/8383 -- Add formspec styling using a style[] tag by rubenwardy 17:45 Krock 30 potatoes 17:45 * DS-minetest has successfully run a sscsm 17:46 DS-minetest (with cpcsm script) 17:47 Krock nore, rubenwardy, sofar: Would you be available for a meeting in a few minutes? 17:47 rubenwardy I'm currently making lasagne 17:47 Krock okay. chances for a meeting this week dropped to 0% 17:48 rubenwardy I can pop in and out though 17:48 rubenwardy so go ahead 17:50 GreenDimond So the place where irrlicht decides to draw entities or not is in OnRegisterSceneNode() .. If I understand correctly, registering my own OnRegister for GenericCAO doing the same function without checking for IsVisible should in theory fix the problem? 17:53 Krock perhaps? 17:53 Krock try & error is a learning strategy as well 17:54 GreenDimond True 18:55 GreenDimond So it looks like I have to make my own scene node to do onregisterscenenode 19:13 paramat core devs, please consider and comment in #8745 19:13 ShadowBot https://github.com/minetest/minetest/issues/8745 -- Add a modernised Mapgen v61? 19:14 rubenwardy don't we need v8-v60 first? 19:14 GreenDimond ^ 19:14 GreenDimond also are you gonna be working on this and watershed at the same time? 19:19 paramat lol 19:19 paramat it won't affect how i work on watershed 19:25 Sokomine regarding the formspec pr with the html links: rubenwardy had equally spaced answers in mind. those can be done with textarea[]. but that won't help me at all. possible answers to select from may vary a lot in length 19:26 Sokomine paramat: i need to take a look at the biomes for the lakes. seems they work pretty straightforward...but i wonder what happens if a mod changes anything? nothing i guess? sometimes a way to pass modified heightmap, biomemap and in particular voxelmanip up to the next mod would be very helpful 19:40 paramat you could use the 'get biome data at point' API? 19:41 paramat that gives you biome, heat and humidity values. or yes you could use the biomemap for per-node biome information 19:43 paramat oh erm, this is not dev talk, you can private message me in IRC if you want 19:43 paramat well, the 2nd bit is dev talk 19:51 paramat merging game#2433 19:51 ShadowBot https://github.com/minetest/minetest_game/issues/2433 -- Make the creative mod hand dig 'dig_immediate' nodes fast by paramat 20:02 Sokomine yes...i'm sure the "there's a lake here"-information could be of intrest for other mods. so the biomemap ought to be changed and passed on. same with other data. but then we need some order in which mods work on the data 20:51 p_gimeno ~tell paramat about 8745: YESSS PLEASEEE! http://irc.minetest.net/minetest/2019-03-27#i_5519275 20:51 ShadowBot p_gimeno: O.K. 21:20 paramat =) 21:26 paramat Sokomine ok i see what you mean. correct, the mapgen objects (heatmap, humidmap, biomemap, heightmap, voxelmanip) are not modified or modifiable by mods. however, heat, humidity and biome are very rarely changed by a mod ... 21:28 paramat ... also, for mapgen objects other than the voxelmanip, often multiple mods will want to access the original data instead of the modified data. so it's probably best to leave them alone and leave it to mods to pass modified data ... 21:31 paramat ... however, the voxelmanip is rather different and it may be good to modify it and pass it on, as suggested in an issue 21:34 Sokomine passing the voxelmanip data on would be a great help, yes. still...even biomemap might have been changed. just imagine a new biome type "lake shore" or "house" beeing added 21:39 Sokomine modas just placing a structure may not need to change the heightmap. but mgvillages flattens the area. the origianl heightmap will be wrong afterwards. trouble is there's no real solution on which mod ought to come in which order. passing voxelmanip on might speed things up a lot; passing the other data on might help with consistency 21:55 DS-minetest will there be styling properties to customize everything of a formspec element? 21:55 DS-minetest eg. for labels of elements 21:56 DS-minetest and for the real text 21:56 DS-minetest and the one background and the other 21:56 DS-minetest of the texts 22:11 paramat i don't think the biomemap should contain anything other than the IDs of the registered biomes 22:38 Sokomine hmm. maybe you're right. something else placed there needs information on a diffrent level. the original biome is still there in theory (although human settlers do change a biome a lot)