Disable f16 on Aarch64 without neon for llvm < 20.1.1

[tgross35]

This check was added unconditionally in c51b229 ("Disable f16 on Aarch64 without neon") and reverted in 4a8d357 ("Revert "Disable f16 on Aarch64 without neon"") since it did not fail in Rust's build. However, it is still possible to hit this crash if using LLVM 19 built with assertions, so disable the type conditionally based on version here.

Note that for these builds, a similar patch is needed in the build script for compiler-builtins since it does not yet use cfg(target_has_reliable_f16) (hopefully to be resolved in the near future).

Report: #139276 (comment)
Original LLVM issue: llvm/llvm-project#129394

r? @cuviper

[tgross35]

Note that for Fedora, you will also need to add a similar patch here https://github.com/rust-lang/compiler-builtins/blob/674910e0fa6f0fb2cc055f4f7051ff0eb53c7735/compiler-builtins/configure.rs#L91 that checks target.features. This can't really be done in r-l/r since we don't have access to the LLVM version when building builtins. Now that builtins is a subtree I am planning to switch it to use cfg(target_has_eliable_f16), but that of course won't be in time for the version you are building.

[cuviper]

I assume this is leftover from a local test?

[tgross35]

🙃 yup, was double checking the features filter and apparently forgot to save

[cuviper]

Note that for Fedora, you will also need to add a similar patch here [...]

Ok, makes sense.

Now that builtins is a subtree I am planning to switch it to use cfg(target_has_eliable_f16), but that of course won't be in time for the version you are building.

Right, and so this PR won't be enough on its own for master either, but it may still make sense to land it in anticipation of that subtree change.

[rust-log-analyzer]

The job aarch64-gnu-llvm-19-2 failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

              f64
note: required by a bound in `TestableFloat`
   --> coretests/tests/num/flt2dec/mod.rs:178:22
    |
178 | trait TestableFloat: DecodableFloat + fmt::Display {
    |                      ^^^^^^^^^^^^^^ required by this bound in `TestableFloat`

[RUSTC-TIMING] alloctests test:true 15.706
[RUSTC-TIMING] corebenches test:true 8.455
[RUSTC-TIMING] alloctests test:true 16.320
error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/decimal.rs:18:42
   |
18 |         let actual = dec.try_fast_path::<f16>();
   |                          -------------   ^^^ the trait `RawFloat` is not implemented for `f16`
   |                          |
   |                          required by a bound introduced by this call
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64
note: required by a bound in `Decimal::try_fast_path`
  --> /checkout/library/core/src/num/dec2flt/decimal.rs:54:29
   |
54 |     pub fn try_fast_path<F: RawFloat>(&self) -> Option<F> {
   |                             ^^^^^^^^ required by this bound in `Decimal::try_fast_path`

error[E0599]: no method named `integer_decode` found for type `f16` in the current scope
##[error] --> coretests/tests/num/dec2flt/float.rs:7:33
  |
7 |     assert_eq!(3.14159265359f16.integer_decode(), (1608, -9, 1));
  |                                 ^^^^^^^^^^^^^^ method not found in `f16`

error[E0599]: no method named `integer_decode` found for type `f16` in the current scope
##[error] --> coretests/tests/num/dec2flt/float.rs:8:35
  |
8 |     assert_eq!((-8573.5918555f16).integer_decode(), (1072, 3, -1));
  |                                   ^^^^^^^^^^^^^^ method not found in `f16`

error[E0599]: no method named `integer_decode` found for type `f16` in the current scope
##[error]  --> coretests/tests/num/dec2flt/float.rs:10:32
   |
10 |     assert_eq!(2f16.powf(14.0).integer_decode(), (1 << 10, 4, 1));
   |                                ^^^^^^^^^^^^^^ method not found in `f16`

error[E0599]: no method named `integer_decode` found for type `f16` in the current scope
##[error]  --> coretests/tests/num/dec2flt/float.rs:11:21
   |
11 |     assert_eq!(0f16.integer_decode(), (0, -25, 1));
   |                     ^^^^^^^^^^^^^^ method not found in `f16`

error[E0599]: no method named `integer_decode` found for type `f16` in the current scope
##[error]  --> coretests/tests/num/dec2flt/float.rs:12:24
   |
12 |     assert_eq!((-0f16).integer_decode(), (0, -25, -1));
   |                        ^^^^^^^^^^^^^^ method not found in `f16`

error[E0599]: no method named `integer_decode` found for type `f16` in the current scope
##[error]  --> coretests/tests/num/dec2flt/float.rs:13:30
   |
13 |     assert_eq!(f16::INFINITY.integer_decode(), (1 << 10, 6, 1));
   |                              ^^^^^^^^^^^^^^ method not found in `f16`

error[E0599]: no method named `integer_decode` found for type `f16` in the current scope
##[error]  --> coretests/tests/num/dec2flt/float.rs:14:34
   |
14 |     assert_eq!(f16::NEG_INFINITY.integer_decode(), (1 << 10, 6, -1));
   |                                  ^^^^^^^^^^^^^^ method not found in `f16`

error[E0599]: no method named `integer_decode` found for type `f16` in the current scope
##[error]  --> coretests/tests/num/dec2flt/float.rs:18:43
   |
18 |     let (nan_m, nan_p, _nan_s) = f16::NAN.integer_decode();
   |                                           ^^^^^^^^^^^^^^ method not found in `f16`

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:62:17
   |
62 |     assert_eq!(<f16 as RawFloat>::INFINITY, f16::INFINITY);
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:63:17
   |
63 |     assert_eq!(<f16 as RawFloat>::NEG_INFINITY, -f16::INFINITY);
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:64:17
   |
64 |     assert_eq!(<f16 as RawFloat>::NAN.to_bits(), f16::NAN.to_bits());
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:65:17
   |
65 |     assert_eq!(<f16 as RawFloat>::NEG_NAN.to_bits(), (-f16::NAN).to_bits());
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:66:17
   |
66 |     assert_eq!(<f16 as RawFloat>::SIG_BITS, 10);
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:67:17
   |
67 |     assert_eq!(<f16 as RawFloat>::MIN_EXPONENT_ROUND_TO_EVEN, -22);
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:68:17
   |
68 |     assert_eq!(<f16 as RawFloat>::MAX_EXPONENT_ROUND_TO_EVEN, 5);
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:69:17
   |
69 |     assert_eq!(<f16 as RawFloat>::MIN_EXPONENT_FAST_PATH, -4);
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:70:17
   |
70 |     assert_eq!(<f16 as RawFloat>::MAX_EXPONENT_FAST_PATH, 4);
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:71:17
   |
71 |     assert_eq!(<f16 as RawFloat>::MAX_EXPONENT_DISGUISED_FAST_PATH, 7);
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:72:17
   |
72 |     assert_eq!(<f16 as RawFloat>::EXP_MIN, -14);
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:73:17
   |
73 |     assert_eq!(<f16 as RawFloat>::EXP_SAT, 0x1f);
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:74:17
   |
74 |     assert_eq!(<f16 as RawFloat>::SMALLEST_POWER_OF_TEN, -27);
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:75:17
   |
75 |     assert_eq!(<f16 as RawFloat>::LARGEST_POWER_OF_TEN, 4);
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/float.rs:76:17
   |
76 |     assert_eq!(<f16 as RawFloat>::MAX_MANTISSA_FAST_PATH, 2048);
   |                 ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]  --> coretests/tests/num/dec2flt/lemire.rs:6:30
   |
6  |     let fp = compute_float::<f16>(q, w);
   |                              ^^^ the trait `RawFloat` is not implemented for `f16`
   |
   = help: the following other types implement trait `RawFloat`:
             f32
             f64
note: required by a bound in `compute_float`
  --> /checkout/library/core/src/num/dec2flt/lemire.rs:27:25
   |
27 | pub fn compute_float<F: RawFloat>(q: i64, mut w: u64) -> BiasedFp {
   |                         ^^^^^^^^ required by this bound in `compute_float`

error[E0599]: no associated item named `INFINITE_POWER` found for type `f16` in the current scope
##[error]  --> coretests/tests/num/dec2flt/lemire.rs:53:45
   |
53 |     assert_eq!(compute_float16(4, 6), (f16::INFINITE_POWER - 1, 851));
   |                                             ^^^^^^^^^^^^^^ associated item not found in `f16`
   |
help: there is an associated constant `INFINITY` with a similar name
   |
53 -     assert_eq!(compute_float16(4, 6), (f16::INFINITE_POWER - 1, 851));
53 +     assert_eq!(compute_float16(4, 6), (f16::INFINITY - 1, 851));
   |

error[E0599]: no associated item named `INFINITE_POWER` found for type `f16` in the current scope
##[error]  --> coretests/tests/num/dec2flt/lemire.rs:54:45
   |
54 |     assert_eq!(compute_float16(4, 7), (f16::INFINITE_POWER, 0)); // infinity
   |                                             ^^^^^^^^^^^^^^ associated item not found in `f16`
   |
help: there is an associated constant `INFINITY` with a similar name
   |
54 -     assert_eq!(compute_float16(4, 7), (f16::INFINITE_POWER, 0)); // infinity
54 +     assert_eq!(compute_float16(4, 7), (f16::INFINITY, 0)); // infinity
   |

error[E0599]: no associated item named `INFINITE_POWER` found for type `f16` in the current scope
##[error]  --> coretests/tests/num/dec2flt/lemire.rs:55:47
   |
55 |     assert_eq!(compute_float16(2, 655), (f16::INFINITE_POWER - 1, 1023));
   |                                               ^^^^^^^^^^^^^^ associated item not found in `f16`
   |
help: there is an associated constant `INFINITY` with a similar name
   |
55 -     assert_eq!(compute_float16(2, 655), (f16::INFINITE_POWER - 1, 1023));
55 +     assert_eq!(compute_float16(2, 655), (f16::INFINITY - 1, 1023));
   |

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]   --> coretests/tests/num/dec2flt/parse.rs:14:30
    |
14  |         assert_eq!(dec2flt::<f16>(s), Err(pfe_invalid()));
    |                              ^^^ the trait `RawFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `RawFloat`:
              f32
              f64
note: required by a bound in `dec2flt`
   --> /checkout/library/core/src/num/dec2flt/mod.rs:259:19
    |
259 | pub fn dec2flt<F: RawFloat>(s: &str) -> Result<F, ParseFloatError> {
    |                   ^^^^^^^^ required by this bound in `dec2flt`

error[E0277]: the trait bound `f16: RawFloat` is not satisfied
##[error]   --> coretests/tests/num/dec2flt/parse.rs:34:31
    |
34  |                     dec2flt::<f16>(&input),
    |                               ^^^ the trait `RawFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `RawFloat`:
              f32
              f64
note: required by a bound in `dec2flt`
   --> /checkout/library/core/src/num/dec2flt/mod.rs:259:19
    |
259 | pub fn dec2flt<F: RawFloat>(s: &str) -> Result<F, ParseFloatError> {
    |                   ^^^^^^^^ required by this bound in `dec2flt`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]  --> coretests/tests/num/flt2dec/random.rs:92:23
   |
92 |         decode_finite(x)
---
             f64
note: required by a bound in `num::flt2dec::random::decode_finite`
  --> coretests/tests/num/flt2dec/random.rs:10:25
   |
10 | pub fn decode_finite<T: DecodableFloat>(v: T) -> Decoded {
   |                         ^^^^^^^^^^^^^^ required by this bound in `decode_finite`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/random.rs:135:27
    |
135 |             decode_finite(x)
---
              f64
note: required by a bound in `num::flt2dec::random::decode_finite`
   --> coretests/tests/num/flt2dec/random.rs:10:25
    |
10  | pub fn decode_finite<T: DecodableFloat>(v: T) -> Decoded {
    |                         ^^^^^^^^^^^^^^ required by this bound in `decode_finite`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:250:23
    |
38  |         let (buf, k) = $f(&decode_finite($v), &mut buf);
    |                            ------------- required by a bound introduced by this call
...
250 |     check_shortest!(f(0.1f16) => b"1", 0);
    |                       ^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::decode_finite`
   --> coretests/tests/num/flt2dec/mod.rs:16:25
    |
16  | pub fn decode_finite<T: DecodableFloat>(v: T) -> Decoded {
    |                         ^^^^^^^^^^^^^^ required by this bound in `decode_finite`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:255:23
    |
38  |         let (buf, k) = $f(&decode_finite($v), &mut buf);
    |                            ------------- required by a bound introduced by this call
...
255 |     check_shortest!(f(1.0f16/3.0) => b"3333", 0);
    |                       ^^^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::decode_finite`
   --> coretests/tests/num/flt2dec/mod.rs:16:25
    |
16  | pub fn decode_finite<T: DecodableFloat>(v: T) -> Decoded {
    |                         ^^^^^^^^^^^^^^ required by this bound in `decode_finite`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:260:23
    |
38  |         let (buf, k) = $f(&decode_finite($v), &mut buf);
    |                            ------------- required by a bound introduced by this call
...
260 |     check_shortest!(f(3.14f16) => b"314", 1);
    |                       ^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::decode_finite`
   --> coretests/tests/num/flt2dec/mod.rs:16:25
    |
16  | pub fn decode_finite<T: DecodableFloat>(v: T) -> Decoded {
    |                         ^^^^^^^^^^^^^^ required by this bound in `decode_finite`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:265:23
    |
38  |         let (buf, k) = $f(&decode_finite($v), &mut buf);
    |                            ------------- required by a bound introduced by this call
...
265 |     check_shortest!(f(3.1415e4f16) => b"3141", 5);
    |                       ^^^^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::decode_finite`
   --> coretests/tests/num/flt2dec/mod.rs:16:25
    |
16  | pub fn decode_finite<T: DecodableFloat>(v: T) -> Decoded {
    |                         ^^^^^^^^^^^^^^ required by this bound in `decode_finite`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:271:23
    |
38  |         let (buf, k) = $f(&decode_finite($v), &mut buf);
    |                            ------------- required by a bound introduced by this call
...
271 |     check_shortest!(f(ldexp_f16(1.0, 5)) => b"32", 2);
    |                       ^^^^^^^^^^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::decode_finite`
   --> coretests/tests/num/flt2dec/mod.rs:16:25
    |
16  | pub fn decode_finite<T: DecodableFloat>(v: T) -> Decoded {
    |                         ^^^^^^^^^^^^^^ required by this bound in `decode_finite`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:276:23
    |
38  |         let (buf, k) = $f(&decode_finite($v), &mut buf);
    |                            ------------- required by a bound introduced by this call
...
276 |     check_shortest!(f(f16::MAX) => b"655", 5);
    |                       ^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::decode_finite`
   --> coretests/tests/num/flt2dec/mod.rs:16:25
    |
16  | pub fn decode_finite<T: DecodableFloat>(v: T) -> Decoded {
    |                         ^^^^^^^^^^^^^^ required by this bound in `decode_finite`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:281:23
    |
38  |         let (buf, k) = $f(&decode_finite($v), &mut buf);
    |                            ------------- required by a bound introduced by this call
...
281 |     check_shortest!(f(f16::MIN_POSITIVE) => b"6104", -4);
    |                       ^^^^^^^^^^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::decode_finite`
   --> coretests/tests/num/flt2dec/mod.rs:16:25
    |
16  | pub fn decode_finite<T: DecodableFloat>(v: T) -> Decoded {
    |                         ^^^^^^^^^^^^^^ required by this bound in `decode_finite`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:287:23
    |
38  |         let (buf, k) = $f(&decode_finite($v), &mut buf);
    |                            ------------- required by a bound introduced by this call
...
287 |     check_shortest!(f(minf16) => b"6", -7);
    |                       ^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::decode_finite`
   --> coretests/tests/num/flt2dec/mod.rs:16:25
    |
16  | pub fn decode_finite<T: DecodableFloat>(v: T) -> Decoded {
    |                         ^^^^^^^^^^^^^^ required by this bound in `decode_finite`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:297:20
    |
223 |         check_exact(|d, b, k| $f(d, b, k), $v, stringify!($v), $buf, $exp)
    |         ----------- required by a bound introduced by this call
...
297 |     check_exact!(f(0.1f16)            => b"999755859375     ", -1);
    |                    ^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `check_exact`
   --> coretests/tests/num/flt2dec/mod.rs:98:8
    |
96  | fn check_exact<F, T>(mut f: F, v: T, vstr: &str, expected: &[u8], expectedk: i16)
    |    ----------- required by a bound in this function
97  | where
98  |     T: DecodableFloat,
    |        ^^^^^^^^^^^^^^ required by this bound in `check_exact`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:298:20
    |
223 |         check_exact(|d, b, k| $f(d, b, k), $v, stringify!($v), $buf, $exp)
    |         ----------- required by a bound introduced by this call
...
298 |     check_exact!(f(0.5f16)            => b"5                ", 0);
    |                    ^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `check_exact`
   --> coretests/tests/num/flt2dec/mod.rs:98:8
    |
96  | fn check_exact<F, T>(mut f: F, v: T, vstr: &str, expected: &[u8], expectedk: i16)
    |    ----------- required by a bound in this function
97  | where
98  |     T: DecodableFloat,
    |        ^^^^^^^^^^^^^^ required by this bound in `check_exact`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:299:20
    |
223 |         check_exact(|d, b, k| $f(d, b, k), $v, stringify!($v), $buf, $exp)
    |         ----------- required by a bound introduced by this call
...
299 |     check_exact!(f(1.0f16/3.0)        => b"333251953125     ", 0);
    |                    ^^^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `check_exact`
   --> coretests/tests/num/flt2dec/mod.rs:98:8
    |
96  | fn check_exact<F, T>(mut f: F, v: T, vstr: &str, expected: &[u8], expectedk: i16)
    |    ----------- required by a bound in this function
97  | where
98  |     T: DecodableFloat,
    |        ^^^^^^^^^^^^^^ required by this bound in `check_exact`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:300:20
    |
223 |         check_exact(|d, b, k| $f(d, b, k), $v, stringify!($v), $buf, $exp)
    |         ----------- required by a bound introduced by this call
...
300 |     check_exact!(f(3.141f16)          => b"3140625          ", 1);
    |                    ^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `check_exact`
   --> coretests/tests/num/flt2dec/mod.rs:98:8
    |
96  | fn check_exact<F, T>(mut f: F, v: T, vstr: &str, expected: &[u8], expectedk: i16)
    |    ----------- required by a bound in this function
97  | where
98  |     T: DecodableFloat,
    |        ^^^^^^^^^^^^^^ required by this bound in `check_exact`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:301:20
    |
223 |         check_exact(|d, b, k| $f(d, b, k), $v, stringify!($v), $buf, $exp)
    |         ----------- required by a bound introduced by this call
...
301 |     check_exact!(f(3.141e4f16)        => b"31408            ", 5);
    |                    ^^^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `check_exact`
   --> coretests/tests/num/flt2dec/mod.rs:98:8
    |
96  | fn check_exact<F, T>(mut f: F, v: T, vstr: &str, expected: &[u8], expectedk: i16)
    |    ----------- required by a bound in this function
97  | where
98  |     T: DecodableFloat,
    |        ^^^^^^^^^^^^^^ required by this bound in `check_exact`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:302:20
    |
223 |         check_exact(|d, b, k| $f(d, b, k), $v, stringify!($v), $buf, $exp)
    |         ----------- required by a bound introduced by this call
...
302 |     check_exact!(f(f16::MAX)          => b"65504            ", 5);
    |                    ^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `check_exact`
   --> coretests/tests/num/flt2dec/mod.rs:98:8
    |
96  | fn check_exact<F, T>(mut f: F, v: T, vstr: &str, expected: &[u8], expectedk: i16)
    |    ----------- required by a bound in this function
97  | where
98  |     T: DecodableFloat,
    |        ^^^^^^^^^^^^^^ required by this bound in `check_exact`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:303:20
    |
223 |         check_exact(|d, b, k| $f(d, b, k), $v, stringify!($v), $buf, $exp)
    |         ----------- required by a bound introduced by this call
...
303 |     check_exact!(f(f16::MIN_POSITIVE) => b"6103515625       ", -4);
    |                    ^^^^^^^^^^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `check_exact`
   --> coretests/tests/num/flt2dec/mod.rs:98:8
    |
96  | fn check_exact<F, T>(mut f: F, v: T, vstr: &str, expected: &[u8], expectedk: i16)
    |    ----------- required by a bound in this function
97  | where
98  |     T: DecodableFloat,
    |        ^^^^^^^^^^^^^^ required by this bound in `check_exact`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:304:20
    |
223 |         check_exact(|d, b, k| $f(d, b, k), $v, stringify!($v), $buf, $exp)
    |         ----------- required by a bound introduced by this call
...
304 |     check_exact!(f(minf16)            => b"59604644775390625", -7);
    |                    ^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `check_exact`
   --> coretests/tests/num/flt2dec/mod.rs:98:8
    |
96  | fn check_exact<F, T>(mut f: F, v: T, vstr: &str, expected: &[u8], expectedk: i16)
    |    ----------- required by a bound in this function
97  | where
98  |     T: DecodableFloat,
    |        ^^^^^^^^^^^^^^ required by this bound in `check_exact`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:641:33
    |
641 |         assert_eq!(to_string(f, f16::MAX, Minus, 0), "65500");
    |                    ---------    ^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |                    |
    |                    required by a bound introduced by this call
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::to_shortest_str_test::to_string`
   --> coretests/tests/num/flt2dec/mod.rs:585:12
    |
583 |     fn to_string<T, F>(f: &mut F, v: T, sign: Sign, frac_digits: usize) -> String
    |        --------- required by a bound in this function
584 |     where
585 |         T: DecodableFloat,
    |            ^^^^^^^^^^^^^^ required by this bound in `to_string`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:642:33
    |
642 |         assert_eq!(to_string(f, f16::MAX, Minus, 1), "65500.0");
    |                    ---------    ^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |                    |
    |                    required by a bound introduced by this call
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::to_shortest_str_test::to_string`
   --> coretests/tests/num/flt2dec/mod.rs:585:12
    |
583 |     fn to_string<T, F>(f: &mut F, v: T, sign: Sign, frac_digits: usize) -> String
    |        --------- required by a bound in this function
584 |     where
585 |         T: DecodableFloat,
    |            ^^^^^^^^^^^^^^ required by this bound in `to_string`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:643:33
    |
643 |         assert_eq!(to_string(f, f16::MAX, Minus, 8), "65500.00000000");
    |                    ---------    ^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |                    |
    |                    required by a bound introduced by this call
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::to_shortest_str_test::to_string`
   --> coretests/tests/num/flt2dec/mod.rs:585:12
    |
583 |     fn to_string<T, F>(f: &mut F, v: T, sign: Sign, frac_digits: usize) -> String
    |        --------- required by a bound in this function
584 |     where
585 |         T: DecodableFloat,
    |            ^^^^^^^^^^^^^^ required by this bound in `to_string`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:646:33
    |
646 |         assert_eq!(to_string(f, minf16, Minus, 0), "0.00000006");
    |                    ---------    ^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |                    |
    |                    required by a bound introduced by this call
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::to_shortest_str_test::to_string`
   --> coretests/tests/num/flt2dec/mod.rs:585:12
    |
583 |     fn to_string<T, F>(f: &mut F, v: T, sign: Sign, frac_digits: usize) -> String
    |        --------- required by a bound in this function
584 |     where
585 |         T: DecodableFloat,
    |            ^^^^^^^^^^^^^^ required by this bound in `to_string`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:647:33
    |
647 |         assert_eq!(to_string(f, minf16, Minus, 8), "0.00000006");
    |                    ---------    ^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |                    |
    |                    required by a bound introduced by this call
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::to_shortest_str_test::to_string`
   --> coretests/tests/num/flt2dec/mod.rs:585:12
    |
583 |     fn to_string<T, F>(f: &mut F, v: T, sign: Sign, frac_digits: usize) -> String
    |        --------- required by a bound in this function
584 |     where
585 |         T: DecodableFloat,
    |            ^^^^^^^^^^^^^^ required by this bound in `to_string`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:648:33
    |
648 |         assert_eq!(to_string(f, minf16, Minus, 9), "0.000000060");
    |                    ---------    ^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |                    |
    |                    required by a bound introduced by this call
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::to_shortest_str_test::to_string`
   --> coretests/tests/num/flt2dec/mod.rs:585:12
    |
583 |     fn to_string<T, F>(f: &mut F, v: T, sign: Sign, frac_digits: usize) -> String
    |        --------- required by a bound in this function
584 |     where
585 |         T: DecodableFloat,
    |            ^^^^^^^^^^^^^^ required by this bound in `to_string`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:765:33
    |
765 |         assert_eq!(to_string(f, f16::MAX, Minus, (-2, 2), false), "6.55e4");
    |                    ---------    ^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |                    |
    |                    required by a bound introduced by this call
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::to_shortest_exp_str_test::to_string`
   --> coretests/tests/num/flt2dec/mod.rs:695:12
    |
693 |     fn to_string<T, F>(f: &mut F, v: T, sign: Sign, exp_bounds: (i16, i16), upper: bool) -> String
    |        --------- required by a bound in this function
694 |     where
695 |         T: DecodableFloat,
    |            ^^^^^^^^^^^^^^ required by this bound in `to_string`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:766:33
    |
766 |         assert_eq!(to_string(f, f16::MAX, Minus, (-4, 4), false), "6.55e4");
    |                    ---------    ^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |                    |
    |                    required by a bound introduced by this call
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::to_shortest_exp_str_test::to_string`
   --> coretests/tests/num/flt2dec/mod.rs:695:12
    |
693 |     fn to_string<T, F>(f: &mut F, v: T, sign: Sign, exp_bounds: (i16, i16), upper: bool) -> String
    |        --------- required by a bound in this function
694 |     where
695 |         T: DecodableFloat,
    |            ^^^^^^^^^^^^^^ required by this bound in `to_string`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied
##[error]   --> coretests/tests/num/flt2dec/mod.rs:767:33
    |
767 |         assert_eq!(to_string(f, f16::MAX, Minus, (-5, 5), false), "65500");
    |                    ---------    ^^^^^^^^ the trait `DecodableFloat` is not implemented for `f16`
    |                    |
    |                    required by a bound introduced by this call
    |
    = help: the following other types implement trait `DecodableFloat`:
              f32
              f64
note: required by a bound in `num::flt2dec::to_shortest_exp_str_test::to_string`
   --> coretests/tests/num/flt2dec/mod.rs:695:12
    |
693 |     fn to_string<T, F>(f: &mut F, v: T, sign: Sign, exp_bounds: (i16, i16), upper: bool) -> String
    |        --------- required by a bound in this function
694 |     where
695 |         T: DecodableFloat,
    |            ^^^^^^^^^^^^^^ required by this bound in `to_string`

error[E0277]: the trait bound `f16: DecodableFloat` is not satisfied

[tgross35]

Oh, that's interesting; Cranelift now supports these types even though LLVM 19 doesn't. The situation here looks kind of weird; IIUC cranelift is building coretests but must be running it against LLVM-built core?