Time Nick Message 00:14 rubenwardy so, my preference for #7226 is to have it merged and work on it in future PRs 00:14 ShadowBot https://github.com/minetest/minetest/issues/7226 -- Add online content library by rubenwardy 00:15 rubenwardy future PRs will be loading screenshots and supporting tags 00:15 rubenwardy among other things 00:17 rubenwardy plus having the bulk of it in will mean any issues can be found well before the release 00:41 paramat ok with me 00:46 rubenwardy pushing trivial patch to master in ~5mins: https://github.com/rubenwardy/minetest/commit/574dab5c1195f7ce30cf9b59950a229a953fd59f 00:46 rubenwardy !title 00:46 ShadowBot Add CMakeDoxy* to .gitignore · rubenwardy/minetest@574dab5 · GitHub 00:52 rubenwardy done 10:33 Wuzzy forum.minetest.net is broken (SQL ERROR) 11:18 Fixer can confirm 12:26 ashtrayoz3 So, what would be a good way to get someone like @VanessaE or @tenplus1 to give feedback on #7231 12:26 ShadowBot https://github.com/minetest/minetest/issues/7231 -- Add "not tool" function to extended drop tables, document how extended drop tables work. by ashtrayoz 12:26 ashtrayoz3 *? 12:53 shivajiva I'm sure both will offer an opinion when they read the logs :) 16:48 VanessaE I have no opinion :P 17:26 Krock ashtrayoz3, git checkout master && git reset --hard HEAD~1 && git pull upstream 17:27 Krock your master branch is salted. > "Documentation for preserve_metadata.". Only works if the merge commit is the head commit, however 18:33 JDCodeIt Working on #7235 found the place extra renderers are being created ContentFeatures::updateTextures - but bazaar thing is during debugging the instruction pointer is jumping all over - seems to only be in the main thread during this time. Cleaning and rebuilding all. 18:33 ShadowBot https://github.com/minetest/minetest/issues/7235 -- Memory is not freed when exiting a world back to the main menu 18:34 JDCodeIt shdsrc->getShader( should return a reference to existing shader if it exists, or create a new one with the material_type and draw type. But seems randomly some garbage is going into the function and causing a new shader to create. 18:59 JDCodeIt The stack is corrupted when this structure is copied: tdef[j] = tiledef[j]; 19:30 JDCodeIt Rebuild fixed that... the reason materials keep getting added each menu-game loop is that shdsrc->getShader( is using a different shdsrc each cycle - so the material is not found, then adds it - maybe this pointer needs to be kept above "Client" to stay alive during these loops. 19:37 JDCodeIt ClientLauncher creates and instance of Game the_game which calls connectToServer which uses new to create a Client which then creates a new shdsrc - which doesn't remember any of the previous material_type/drawtype combinations 19:38 rubenwardy it might be a good idea to summarise this in an issue, not everyone is in this channel 19:38 rubenwardy I believe numberZero was the one to write or at least last touch that code 19:38 rubenwardy so they may be able to help - they're only on Github though 19:38 JDCodeIt hoping for some discussion here, but of course will add to the issue 19:41 JDCodeIt Will summarize on Github maybe with a proposed solution 20:04 paramat fixed it! #7249 that was tough 20:04 ShadowBot https://github.com/minetest/minetest/issues/7249 -- Cavegen: Fix variable typo that broke mgvalleys large cave distribution by paramat 20:29 Krock paramat, cavegen.h L137 unused "max_stone_y" 20:30 Krock i.e. initialized to any garbage value each time 20:33 paramat i see 20:34 paramat ok that explains all then, thanks 20:35 paramat 5 year old bug fixed 20:37 Krock Nice. Will it be noticeable - and what did the bug do? 20:39 Krock ah. I see. intersections didn't happen 20:40 paramat see comments in #7244 20:40 ShadowBot https://github.com/minetest/minetest/issues/7244 -- Mgvalleys: Use shared tunnel / cavern code instead of internal by paramat 20:40 paramat tested mgv7 and is fine, so i think mergeable 20:41 paramat i mean the fix is mergeable 20:41 paramat (7249) 20:41 paramat i'm not sure if the bug affected any other mapgen than mgvalleys, funny if it did 20:42 paramat "MT caves broken since 2013, nobody notices" 20:43 Krock Would you mind removing the unused variable from cavegen.h too? 20:43 paramat oh yes 20:43 paramat i mean, i don't :) 20:43 paramat (mind) 20:44 Krock hehe :) due the garbage values it always dodged when people tried to observe this bug 20:44 Krock nice catch 20:51 paramat done 20:55 paramat thanks 20:55 JDCodeIt Issue #7235 will take some restructuring. The shadersource for whatever reason needs to have some GameGlobalShaderConstantSetterFactory attached to it, which has a reference to Client, which won't exist yet if we raise shadersource creation above existence of Game in the_game 20:55 ShadowBot https://github.com/minetest/minetest/issues/7235 -- Memory is not freed when exiting a world back to the main menu 20:55 paramat so large caves will now occur at all levels and won't all be flat, much more vertical motion 20:56 paramat will merge #7249 in 20 mins 20:56 ShadowBot https://github.com/minetest/minetest/issues/7249 -- Cavegen: Fix variable typo that broke mgvalleys large cave distribution by paramat 21:17 JDCodeIt I hacked a test of #7235 - but it didn't work - still adds renderers with the same shader pointer - back to the drawing board 21:17 ShadowBot https://github.com/minetest/minetest/issues/7235 -- Memory is not freed when exiting a world back to the main menu 21:24 paramat merging 7249 21:29 paramat done 21:30 paramat #7244 is ready 21:30 ShadowBot https://github.com/minetest/minetest/issues/7244 -- Mgvalleys: Use shared tunnel / cavern code instead of internal by paramat 21:35 * paramat is at 701 commits total 21:38 * rubenwardy is at 70 commits total 21:39 * Krock is at 7.01 commits total 21:39 rubenwardy although I've only added just under half as many lines as you, paramat 21:39 rubenwardy wait, maths 21:39 rubenwardy *quarter 21:40 rubenwardy ah, you 21:40 rubenwardy ah, you're counting MTG 21:40 rubenwardy MTG don't count ;) 21:41 rubenwardy well, it does. Hopefully we'll have more games available soon so it'll be less on a pedistal 22:34 Fixer MTG should be first class citizen in this room 23:01 JDCodeIt The duplicate shaders are being created in a subsequent call to rebuildShaders() - still looking for this resource holdover.