I bumped llvm to 17.0..6 and osl and ipsc/oidn seem to build, but i have not yet had the time to do build/test blender against these libs
Think my longest libs upload was shortly over 14 hours, landing "at the same time" will likely be impossible, I suspect the best we can do is land on the same day and tell people not to commit…
I Vaguely remember @PrototypeNM1 fighting with this, and i'd like to get his opinion before making changes here
The CI environment are VM's they likely don't have a GPU available, I managed to get this working without a GPU by doing the following:
- get mesa3d-23.3.0-release-msvc.7z from https://github.co…
Also the assert in that function doesn't work, because STR_CONCAT will refuse to overflow the buffer.
We're on llvm15 currently since that's what the old patch had, given llvm bumps tend to be disruptive in the formatting department, do we want to bump to 16 or 17 so we have a little longer stable…
#115954 should have handled that one, you should be able to carry on with F5
I'm kinda stabbing in the dark here, do you have a screenshot of the debugger when this crash occurs?
can you try #115954 ? Tablet support would still be broken, but at least it wouldn't crash blender anymore.
Can you run it in the debugger and see where it crashes?
That's a no repro for me, when it crashes does it produce a crash log i can perhaps look at?
accepted, but this does need a better title/commit message when landing
All those commits are essentially 99% automated work, so i have no strong attachment to the work, while i'd like to understand what is happening here, i'd probably need @brecht to take a look, and his schedule is busy enough as is, and if he has time i'd rather have him work on the 4.1 lib update, so I'd say we revert the ones you flagged, so you and clement can move forward again. Once shader_builder gets enabled by default on the mac bot, I'll pick up the cmake work again and we can figure out whats happening.