| Time |
Nick |
Message |
| 03:16 |
|
mrcheese joined #luanti-dev |
| 04:00 |
|
MTDiscord joined #luanti-dev |
| 04:18 |
|
rubenwardy joined #luanti-dev |
| 04:40 |
|
Farooq joined #luanti-dev |
| 04:47 |
|
Farooq joined #luanti-dev |
| 05:29 |
|
Farooq joined #luanti-dev |
| 05:41 |
|
[MatrxMT] joined #luanti-dev |
| 05:53 |
|
[MatrxMT] joined #luanti-dev |
| 06:01 |
|
Farooq joined #luanti-dev |
| 06:03 |
|
Farooq joined #luanti-dev |
| 06:28 |
|
Farooq joined #luanti-dev |
| 06:31 |
|
Farooq joined #luanti-dev |
| 06:50 |
|
Farooq joined #luanti-dev |
| 06:52 |
|
Farooq joined #luanti-dev |
| 07:37 |
|
YuGiOhJCJ joined #luanti-dev |
| 09:03 |
|
YuGiOhJCJ joined #luanti-dev |
| 09:52 |
|
basxto joined #luanti-dev |
| 09:56 |
|
pgimeno joined #luanti-dev |
| 11:25 |
|
YuGiOhJCJ joined #luanti-dev |
| 12:02 |
|
YuGiOhJCJ joined #luanti-dev |
| 12:36 |
MTDiscord |
<paradust> Is it intentional that asserts are always enabled in irr/ (but not src/) ? |
| 12:37 |
MTDiscord |
<paradust> for optimized release builds |
| 12:37 |
sfan5 |
no idea |
| 12:37 |
MTDiscord |
<paradust> if you don't know who would 😮 |
| 12:40 |
MTDiscord |
<paradust> that appears to be the current behavior. CMakeLists.txt in src/ includes -DNDEBUG in CMAKE_CXX_FLAGS_RELEASE, but irr/src does not include it. |
| 12:43 |
sfan5 |
if you look at what the we did before https://github.com/minetest/irrlicht/blob/7173c2c62997b6416f17b90f9a50bff11fef1c4c/include/irrTypes.h#L68-L82 - https://github.com/minetest/irrlicht/blob/7173c2c62997b6416f17b90f9a50bff11fef1c4c/src/CMakeLists.txt#L23 |
| 12:43 |
sfan5 |
then it's apparent that asserts were only enabled in explicit debug builds |
| 12:45 |
MTDiscord |
<paradust> It looks like irrlicht preferred to have its own custom debug flag _DEBUG, rather than using NDEBUG |
| 12:45 |
sfan5 |
(context: we replaced the IRR_DEBUG_BREAK_IF with normal asserts at some point) |
| 12:46 |
sfan5 |
however I can't confirm what you said. I added assert(asdnlkjsdgoijsefbiuef); to some irrlicht file and there's no compile error |
| 12:46 |
sfan5 |
using CMAKE_BUILD_TYPE=RelWithDebInfo |
| 12:46 |
MTDiscord |
<paradust> On linux? |
| 12:47 |
sfan5 |
yes |
| 12:47 |
sfan5 |
what it runs: /usr/bin/clang++ -DIRR_PREFER_SDL_GL_HEADER -D_IRR_COMPILE_WITH_JOYSTICK_EVENTS_ -D_IRR_COMPILE_WITH_OGLES2_ -D_IRR_COMPILE_WITH_OPENGL_ -D_IRR_COMPILE_WITH_SDL_DEVICE_ -D_IRR_POSIX_API_ -I... -I... -I... -I... -march=native -pipe -O2 -g -DNDEBUG -std=gnu++17 -fPIC -Wall -pipe -Werror=mismatched-tags -Wnull-dereference -MD -MT |
| 12:47 |
sfan5 |
irr/src/CMakeFiles/IRRMESHOBJ.dir/WeightBuffer.cpp.o -MF CMakeFiles/IRRMESHOBJ.dir/WeightBuffer.cpp.o.d -o CMakeFiles/IRRMESHOBJ.dir/WeightBuffer.cpp.o -c .../src/WeightBuffer.cpp |
| 12:48 |
sfan5 |
probably CMAKE_CXX_FLAGS_* is global including for subprojects? |
| 12:50 |
MTDiscord |
<paradust> I think it may only happen with Release, not RelWithDebInfo |
| 12:50 |
MTDiscord |
<paradust> but i'll check, one sec |
| 12:50 |
|
jonadab joined #luanti-dev |
| 12:55 |
MTDiscord |
<paradust> yea confirmed. It only happens with Release, because irr/src/CMakeLists.txt overrides CMAKE_CXX_FLAGS_RELEASE, not CMAKE_CXX_FLAGS_RELWITHDEBINFO |
| 12:58 |
sfan5 |
we should stop doing that then |
| 13:00 |
MTDiscord |
<ggepic64> > (...) -D_IRR_POSIXAPI -I... -I... -I... -I... -march=native (...) Why is this command stuttering ? XD |
| 14:47 |
|
Farooq joined #luanti-dev |
| 15:14 |
|
[MatrxMT] joined #luanti-dev |
| 15:20 |
|
Farooq joined #luanti-dev |
| 15:21 |
|
[MatrxMT] joined #luanti-dev |
| 15:44 |
|
Panquesito7 joined #luanti-dev |
| 17:20 |
|
Farooq joined #luanti-dev |
| 19:03 |
|
Desour joined #luanti-dev |
| 19:34 |
|
Sendein joined #luanti-dev |
| 19:53 |
|
celeron55 joined #luanti-dev |
| 21:23 |
|
Farooq joined #luanti-dev |
| 21:40 |
MTDiscord |
<herowl> When will #17359 be merged? |
| 21:40 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/17359 -- HUD inventory: allow separate background by CodingMantisUA |
| 21:40 |
MTDiscord |
<herowl> It's got an approval under one approval rule |
| 21:42 |
sfan5 |
whoever did the one and only review should do it |
