Time  Nick      Message
02:16 MTDiscord <wsor4035> all pages are converted to markdown (at least first pass, see issue 3 for more), looking for input on https://github.com/wsor4035/devwiki/issues/2
19:31 sfan5     Krock: the mapgen test seems to randomly fail https://github.com/minetest/minetest/actions/runs/12242046658/job/34148498773?pr=15531#step:6:639
19:34 sfan5     moving the "time of VBO upload" is like a magic performance wand
19:35 Krock     sfan5: nice. can we somehow seed our rand() with that result?
19:36 sfan5     is the map seed during the test random?
19:36 sfan5     i was thinking more like there being accidental UB somewhere
19:37 Krock     hmm. the biome selection is actually global and thus seed independent
19:37 sfan5     the "Randomness seeded to:" it prints is from catch2 and unrelated to our stuff btw
19:37 Krock     I rather meant the fail/pass characteristics of the test that could be an overengineered source of rancomness
19:37 Krock     *randomness
19:37 sfan5     oh haha
19:38 Krock     running valgrind on latest master.....
19:38 sfan5     i recommend ASan rather
19:40 Krock     ( 5 hours later )
19:41 Krock     irr/src/CGUIListBox.cpp:700:38: warning: ‘<anonymous>.irr::video::SColor::color’ may be used uninitialized [-Wmaybe-uninitialized]
19:43 Krock     that's actually true. The default constructor doesn't initialize color
19:44 Krock     so the compiler might in fact just return whatever was in RAX (since the constructor can be inlined, and thus omitted entirely.
19:47 sfan5     still looking for test results on #15531 btw
19:47 ShadowBot https://github.com/minetest/minetest/issues/15531 -- [no squash] Group sparse mesh buffers over entire scene for rendering by sfan5
19:48 sfan5     you can just merge it in, do what you do normally with luanti and see if any difference is noticable
19:48 sfan5     especially interesting if you have a scene with low fps
19:49 Krock     asan reported no leaks or UB on the mapgen unittest
19:50 Krock     sfan5: OpenGL/VBO.h:26:9: error: ‘size_t’ does not name a type << missing stddef include
19:52 Krock     no. the header lacks a pragma once. maybe that's it.
20:01 Krock     > Tracer caught signal 11: addr=0x2a000018 pc=0x7fc3a4fa07e8 sp=0x7fc39ea44d20    << that's an interesting AddressSanitizer output. Cannot catch it with gdb :/
20:06 Krock     https://www.imghost.net/ib/d2SJTwDD3mtXdOw_1733774753.png  (not to worry, this is from master. PR works.)