Time Nick Message 05:20 GreenDimond paramat: Why do collisionboxes still work if the camera is at 0,0,0? 05:20 GreenDimond And does #8015 actually fix the bug or is that PR broke too? 05:20 ShadowBot https://github.com/minetest/minetest/issues/8015 -- Camera: Code improvements, use Raycast by SmallJoker 05:31 GreenDimond nvm on that last question 05:31 GreenDimond But I still dont get how the collision boxes work if the camera is causing the bug 05:32 paramat hi, answered in the thread 05:33 GreenDimond paramat: Yeah I saw, but why does the camera affect collisionboxes? 05:36 GreenDimond If the collisionbox was working why did it only "appear" to work 05:45 paramat core devs: can we decide to now only merge blockers, bugfixes and very trivial PRs? (i don't like calling it a 'feature freeze' because some of the blockers may be features). i think we should not merge PRs unessential for 5.0.0 as they can and will cause new issues and bugs, probably as fast as we can fix them. nerzhul rubenwardy sfan5 sofar 05:45 GreenDimond D: 05:45 GreenDimond Jordach's set_sky PR is almost ready though 05:45 paramat i'll discuss later in the thread, bye 06:38 luk3yx Would it be a good idea to implement server-sent CSMs as a CSM? 06:48 sofar that's the whole idea, currently 08:21 nerzhul paramat, we should just ensure what is in the list. Please ensure the 5.0.0 milestone contains only the correct PR and issues 08:21 nerzhul but yes stop that features, i started to delay many PR to ensure we can release in january 08:22 luk3yx If I make a SSCSM CSM, should I PR upstream MT with it? 08:43 Krock nerzhul: see discussion above in #7034. The PR resulted in a bug 08:43 ShadowBot https://github.com/minetest/minetest/issues/7034 -- Fix missing warningstream (or similar problem) by HybridDog 08:51 Krock vector3d operator/(const T v) const { T i=(T)1.0/v; return vector3d(X * i, Y * i, Z * i); } 08:51 Krock wth irrlicht 08:52 Krock this way divisions are always resulting in 0 when T isn't floating point 08:53 Krock well, or 1 if T=1 10:35 nerzhul Krock: erf 10:46 p_gimeno division as multiplication has issues, that's the cause of #3943 too 10:46 ShadowBot https://github.com/minetest/minetest/issues/3943 -- testStreamRead and testBufReader failures 10:48 p_gimeno with -ffast-math, which is active in release mode, that optimization will be made by the compiler anyway, no need to have it expressly in the code 12:20 nerzhul i will release a release apk with the following fixes: #8012 12:20 ShadowBot https://github.com/minetest/minetest/issues/8012 -- Minetest 0.4.17.1-3 (android only) 12:21 nerzhul i do a test on my phone which will be sufficient 12:21 nerzhul and publish on the play store 12:23 rubenwardy have you fixed the INVALID_ENUM yet? 12:23 nerzhul i just backported some fixes, see the issue 12:24 nerzhul as i miss some time to backport more fixes i just release with those fixes and we can re-release with more fixes after if needed 12:25 rubenwardy maybe it's worth starting the build from scratch (ie: make clean and delete dep folder) and then comparing commands with stujones1? 12:26 nerzhul i always do that 12:26 nerzhul and i do make clean_all not juts make clean 12:26 nerzhul :p 12:27 nerzhul note on the play store we are in 0.4.17.2 12:47 rubenwardy pushing trivial bug fix in 10: https://github.com/rubenwardy/minetest/commit/7d7ccf5c0f3d8689519629282d72b1fea8f4bf30 19:43 nerzhul strange i cannot compile libiconv on my machine 19:43 nerzhul (for android) 19:43 nerzhul i get 19:43 nerzhul "ERROR: Failed to create toolchain." 20:08 stujones11 nerzhul: you need to use ndk-r17c it looks like you are using 18 20:08 nerzhul yeah i don't know what's the problem with 18, maybe arm drop ? 20:09 stujones11 I have a working branch here https://github.com/stujones11/minetest/tree/android-ndk-18 20:10 stujones11 only it requires we up the min target to api 16, still WIP 20:22 nerzhul okay if you can provide a pr it can be nice (on stalbe-0.4 and another for master) 20:26 stujones11 I would like for it to still work for api 14 & 15 but I don't suppose it will effect many useable devices, anyway. 20:34 stujones11 There is no need to be using the latest ndk, really 20:34 stujones11 It only becomes a problem if left for too long then we have to update for some reason 20:34 stujones11 like c++11 20:54 Rama31 Has anyone profiled the client (gprof)? I want to know my it runs so slow on a Raspberry Pi (1 FPS) but cannot get gprof to work on a cross-compiler configuration. Is it wasting time in Irrlicht rendering or on OpenGL? 23:59 paramat merging #7943 in a minute 23:59 ShadowBot https://github.com/minetest/minetest/issues/7943 -- Extend pitch fly mode to swimming by random-geek