zeroomega wrote:
> Thanks for the analysis. Per-target runtime dir strikes again! That setting
> keeps on creating issues (by no fault of its own) because it's not the
> default.
>
> Can we do something like this instead, then?
>
> ```
> config.substitutions.append(('%{lib}',
> '@LIBUNWIND_TESTING_INSTALL_PREFIX@/@LIBUNWIND_INSTALL_LIBRARY_DIR@'))
> ```
>
> We should fix this one forward since there's a complicated dependency chain
> on top of this (we switched to Github-hosted runners for macOS testing after
> this patch and they will start failing if we revert this). Hopefully it's not
> too complicated to fix forward.
That seems to fix the test:
Test run:
https://ci.chromium.org/ui/p/fuchsia/builders/toolchain.ci.shadow/clang-linux-x64/b8735236964550803665/overview
based on https://github.com/zeroomega/llvm-project/tree/libunwind_fix
https://github.com/zeroomega/llvm-project/commit/1e3b4a6a97b32cd1b501ad5256036967a0821327
https://github.com/llvm/llvm-project/pull/110171
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
