You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
What happened was that the following function uses the compiler built-in __builtin_readcyclecounter: https://github.com/zhaokg/Rbeast/blob/master/R/src/abc_timer.h#L101 __builtin_readcyclecounter is not implemented in GCC when targeting AArch64, but Clang does so M1 Mac linked OK.
Essentially they check if the built-in exists and if not fallback to inline ASM, and since we know it's AArch64, it should be relatively safe.
I'm happy to open a PR to fix this.
The text was updated successfully, but these errors were encountered:
First, really appreciate your input, which is extremely insightful and helpful. To properly acknowledge you, would you mind opening a PR and fix it? I will merge it so some proper credits can be given to you.
Once again, thanks a lot for sharing your expertise.
While installing this package on a AArch64 Linux machine, I got the following error at link time:
What happened was that the following function uses the compiler built-in
__builtin_readcyclecounter
:https://github.com/zhaokg/Rbeast/blob/master/R/src/abc_timer.h#L101
__builtin_readcyclecounter
is not implemented in GCC when targeting AArch64, but Clang does so M1 Mac linked OK.I propose we handle the
cpu_ARM64
case following the patch to a similar issue that resolved it for MariaDB: https://bugs.freebsd.org/bugzilla/attachment.cgi?id=216668&action=diffEssentially they check if the built-in exists and if not fallback to inline ASM, and since we know it's AArch64, it should be relatively safe.
I'm happy to open a PR to fix this.
The text was updated successfully, but these errors were encountered: