Martin Kinkelin
Martin Kinkelin
I'm pretty sure this was fixed in https://github.com/dlang/dmd/pull/16248.
(That's also CI-tested: https://github.com/dlang/dmd/pull/16287)
Without looking at the diff yet, I really hate all of this logic. I'm a fan of LDC normalizing this crap at CMake time, with its own renamed `libldc_rt.*` (no...
> This is mostly passing along the complexity for determining the correct suffixes/paths to the user. I guess this is overall easier then trying to reimplement all the logic from...
> Finishing my rant about what is ideal for Gentoo, I think we could tone down the complexity a little. > […] > While I do want to have more...
Alright, and thanks for taking the time to investigate the situation with other package managers/distros, for a nice overview. :+1:
I'll rebase this to re-verify CI (FreeBSD might be broken with its current LLVM 15 and need a trimmed-down #4903).
Okay, I've cherry-picked 2 commits from #4903 (was required for FreeBSD as expected), and don't newly add the conpiler-rt libdir to the .conf when using `LDC_INSTALL_LLVM_RUNTIME_LIBS=ON` - that isn't needed...
Well we don't tamper with those compiler-rt build settings. I guess you are supposed to link the shared sanitizer variant, which we probably don't bundle (but you can use the...
Is there an upstream test for this?