Message393167
I experienced this with Python 3.8 when building it as part of LibreOffice (see <https://git.libreoffice.org/core/+/b0a4b49a88aacfbc127965c0c6fe50a065eb3a0f%5E%21> "external/python3: Clang 13 trunk implements --print-multiarch now", quoting from its commit message below) but from the sources it looks like it would still be an issue with the latest cpython main branch:
Clang 13 trunk implements --print-multiarch now since <https://github.com/llvm/llvm-project/commit/a921d2d2fb46b898794091e7410426c518a4f0cc> "[Driver] Add -print-multiarch", which causes an issue when building with such a compiler on macOS:
> checking build system type... x86_64-apple-darwin19.6.0
> checking host system type... x86_64-apple-darwin19.6.0
[...]
> checking for the platform triplet based on compiler characteristics... darwin
configure: error: internal configure error for the platform triplet, please file a bug report
as configure.ac computes PLATFORM_TRIPLET as "darwin", and instead of computing MULTIARCH as empty (as `$CC --print-multiarch` used to just print
> clang: error: unsupported option '--print-multiarch'
> clang: error: no input files
to stderr), it now computes it as e.g. "x86_64-apple-darwin19.6.0" (or whatever -target is explicitly set to in $CC), so the check that they have equal values if they are both nonempty fails now when building against Clang 13 trunk. (This does not yet appear to be an issue with any Apple Clang version, though.) |
|
Date |
User |
Action |
Args |
2021-05-07 06:45:46 | sberg | set | recipients:
+ sberg |
2021-05-07 06:45:46 | sberg | set | messageid: <1620369946.97.0.428718248395.issue44065@roundup.psfhosted.org> |
2021-05-07 06:45:46 | sberg | link | issue44065 messages |
2021-05-07 06:45:46 | sberg | create | |
|