Time Nick Message 07:06 [MatrxMT] Doesn't minetest enable -ffast-math? 07:33 sfan5 it does not 08:19 Krock repetitivestrain: https://github.com/luanti-org/luanti/pull/9682 08:19 Krock !title 11:49 Krock Could someone on Windows please check whether the dynamic shadows work? Because my sole Windows PC does apparently not support them. https://i.postimg.cc/Pf9dfm7b/luanti-5-13-0-dev-26aab6e-comparison.gif 11:51 sfan5 you could try mesa software rendering 11:51 Krock oh right. 11:54 Krock heh. With software rendering I get 12 FPS (+ 100% uplift) and working dynamic shadows. Intel drivers on Windows surely are special. 11:56 Krock will merge #16368 and #16363 in 15 minutes 11:56 ShadowBot https://github.com/luanti-org/luanti/issues/16368 -- Update credits for 5.13.0 by SmallJoker 11:56 ShadowBot https://github.com/luanti-org/luanti/issues/16363 -- Work around #16221 by updating parent chains by appgurueu 12:31 Krock Pushed 5.13.0 to my fork for testing. waiting for the buildbot to produce the Windows binaries .... 12:54 Krock Would someone please be so nice to merge https://github.com/luanti-org/docs.luanti.org/pull/252 ? I don't have permissions to do so. 12:55 sfan5 yes 12:58 Krock Thanks. 13:03 Krock stable-5 updated. Missing: signed Android builds (whoever has the keys at hand), macOS builds ( ping SFENCE ), package maintainer information 13:03 MTDiscord opening an issue on f-droid, could someone tag 5.13.0 on https://github.com/luanti-org/luanti_android_deps 13:04 Krock does the repo need and update or can it be tagged as-is to HEAD? 13:04 Krock s/and/any 13:04 MTDiscord tagged as-is to HEAD, yeah 13:04 MTDiscord that's what should correspond to 5.13.0 when f-droid builds it for android 13:04 Krock alright. will do. 13:06 Krock @rollerozxa done 13:07 sfan5 soo did we import the weblate translations? 13:07 Krock SFENCE: would you please be so nice to build, test and upload the macOS builds for 5.13.0 / commit 5ab66da when you get the chance? Thanks. 13:07 MTDiscord thx 13:07 MTDiscord https://gitlab.com/fdroid/fdroiddata/-/issues/3596 13:08 Krock sfan5: they were last imported on July 12, which seemed late at that time. Let's hope they're complete enough :3 13:09 Krock at least the recent commits made no changes to the gettext strings 13:09 sfan5 sounds fine 13:10 Krock but now that you mention it - there's a few more recent languages that need fixing as they decided to translate LANG_CODE yet again incorrectly 13:16 rubenwardy can't you mark strings as readonly using a fla 13:16 rubenwardy +g 13:17 Krock yes. they are read-only. perhaps nobody fixed those prior to that lock 13:17 Krock new languages will have an empty string instead 13:23 MTDiscord once all the builds are in releases, i can do https://github.com/luanti-org/luanti-org.github.io/pull/348/files again for the main site 14:39 Krock (to any forum moderator) - Please update the pinned topics in https://forum.luanti.org/viewforum.php?f=18 14:51 sfan5 done 14:51 sfan5 > Vectors passed to C++ API functions may no longer have nil components. This change can reveal logic issues within mods. 14:51 sfan5 IIRC this is not true. we downgraded that to a warning 15:21 Krock right. The second sentence is technically still true, but the first implies that it would cause an error 15:58 pgimeno are there command-line options not shown in luanti --help ? specifically I'm looking for a way to specify a seed for unit tests 16:00 MTDiscord you could provide a config (via flags) with fixed_map_seed 16:03 pgimeno I mean seed as in Catch seed, not map seed 16:04 sfan5 the catch options are not exposed afaik 16:05 pgimeno hm, I am unfamiliar with Catch, could I get a hint on where to specify the seed? (what file to edit) 16:08 pgimeno is it src/unittest/test.cpp ? 16:40 pgimeno btw I'm getting a warning that `stat` (of type HudElementStat) may be used uninitialized in an inlined function in enum_string.h, is that expected? 16:47 pgimeno here's the full report (happens during LTO): http://www.formauri.es/personal/pgimeno/pastes/full-error-info.txt 16:58 pgimeno sorry about the question marks, these are ‘’ quotes that gcc emits and for some reason they don't paste well through VNC 17:03 Krock pgimeno: int result_int should not be assigned there at all. It does however not matter, because the return value does indicate whether the value was actually written 17:04 Krock the generic type `string_to_enum` version should only assign `result` if `ret == true`, and not initialize `result_int` 17:06 pgimeno sorry, I haven't looked into that code at all, just reported it; I'm looking into the unit test that failed on ARM Mac and I think I have a verdict 18:22 pgimeno @luatic http://www.formauri.es/personal/pgimeno/pastes/better-fix-for-16365.txt 18:24 pgimeno the patch also renames test_rotation_degrees to test_rotation_radians as that better reflects the accepted arguments 18:44 pgimeno the seed 4072041937 generated an angle of 1.57016981 which is about 0.0006 radians off of pi/2 18:45 pgimeno that's probably the test that caused the failure 18:47 pgimeno s/an angle/a pitch angle/ 18:48 Desour pgimeno: in case u still need a way to set the catch seed: https://github.com/luanti-org/luanti/commit/d68dc2bd21b312f3aff365799b43e32a22455d15 18:49 Desour (was part of #16365 reproduction steps) 18:49 ShadowBot https://github.com/luanti-org/luanti/issues/16365 -- Unittest `getRotationRadians` in `test_irr_matrix4.h` is flakey 18:51 pgimeno Desour: thanks, I actually found that later