Martin Kinkelin
Martin Kinkelin
> And for what it’s worth, ldc is also able to link the example program even without a __forceinline or inline, which is actually weird as gcc/clang fails to link...
The closest thing to a C `inline` function is a function literal - that's codegen'd into each referencing object file, and only in those. With LDC at least, but I...
IIRC, a `pragma(lib)` in a compiled D module is mapped to a linker directive embedded in the resulting object file - but the compiler also keeps a list of all...
AFAICT, the job was re-run once (see `All Other Runs`), but the re-run got rescheduled too. I've noticed the same behavior for our aarch64 job (e.g., https://cirrus-ci.com/task/4993236477935616) in the last...
Only the CI-tested x86[_64] and AArch64 ABIs are working (i.e., compatible with C, and DMD's custom x86 ABI); there's a 64-bit RISC-V implementation in https://github.com/ldc-developers/ldc/blob/master/gen/abi/riscv64.cpp, but I have no idea...
There is a `CMAKE_INSTALL_LIBDIR` variable in https://github.com/ldc-developers/ldc/blob/master/runtime/CMakeLists.txt#L50 (overriding a user-specified one ATM), but it's only used for the rpath and the ldc2.conf file, not as install destination dir. The multilib...
I've had a quick stab at it: https://github.com/jellyfin/jellyfin-web/pull/7329
Such warnings are also seen when compiling the ldc2 unittest executable with LTO for Linux x86_64. It's non-fatal there though, e.g., from https://github.com/ldc-developers/ldc/actions/runs/9791860098/job/27036496349?pr=4693: ``` [542/3441] Generating bin/ldc2-unittest definition subprograms cannot...
AFAIK, this depends on @ibuclaw re-creating his Buildkite agents using the latest docker image definition: https://github.com/dlang/ci/pull/489#issuecomment-2218942102
We could also opt to go with MSVC for Win32 too. That would mean extending some `version(Win64)` paths in the runtimes to `version(Windows)`. Not sure how compatible this would be...