Stabilize vget_low_s8 on AArch64 #1284
Conversation
|
@Amanieu: no appropriate reviewer found, use r? to override |
| @@ -4989,6 +4989,10 @@ pub unsafe fn vget_high_f32(a: float32x4_t) -> float32x2_t { | |||
| #[target_feature(enable = "neon")] | |||
| #[cfg_attr(target_arch = "arm", target_feature(enable = "v7"))] | |||
| #[cfg_attr(test, assert_instr(nop))] | |||
| #[cfg_attr( | |||
| target_arch = "aarch64", | |||
| stable(feature = "neon_intrinsics", since = "1.59.0") | |||
There was a problem hiding this comment.
This probably shouldn't be 1.59.0 as it's probably not going to be backported.
There was a problem hiding this comment.
I tried that but rustc doesn't like it when I put different version numbers on the same feature name.
There was a problem hiding this comment.
I think you can use a different feature name such as vget_low_s8, then rustc won't complain. Feature names for stable features are mostly irrelevant anyway - those features are stable, they don't need feature annotation. Stabilizing things while changing their feature names was done before, see rust-lang/rust#86344 for instance (in this case, this stabilized MaybeUninit::write which back then used maybe_uninit_extra feature without stabilizing everything else that used maybe_uninit_extra feature).
This was missed when most of the AArch64 intrinsics were stabilized.
This was missed when most of the AArch64 intrinsics were stabilized.
cc rust-lang/rust#94083