migrate to embedded-hal=1.0.0.rc1
#25
Conversation
rursprung
force-pushed
the
migrate-to-eh1.rc1
branch
2 times, most recently
from
2e23e9d to
844f01c
Compare
also ensure that `done()` is called on all mocks (otherwise the tests now fail). this also unearthed 1-2 bugs where the expectations were wrong. since e-h-m 0.10.0 requires rust 1.63 the MSRV for this crate has been raised accordingly as we now cannot test on older releases anymore. it might still work there however.
rursprung
force-pushed
the
migrate-to-eh1.rc1
branch
from
844f01c to
7a4e5b3
Compare
|
the newly emitted warning seems to be a false-positive: rust-lang/rust#76053 |
To provide a little more explanation: The This crate uses The new thing is that |
as alpha.10 of embedded-hal has added `SetDutyCycle` as the equivalent / successor of the old `PwmPin` support for it can now be added here. since e-h 1.0.0 will be released at the end of this year it's safe to just directly migrate to the RC1 instead of adding it as an optional support (requiring another breaking change in a later step which would then remove e-h 0.2 support).
a new warning was caused by the `?` operator as it contains `#[allow(unreachable_code)]`, which clashes with the previous `forbid` rule. as `forbid` does not allow `deny` to override it this would cause a compile error, however due to a compiler bug this was previously not reported at all and has now been changed to a warning in a recent(-ish) version of rustc and will eventually become a compile error. with `deny` it's still possible to use `allow`, thus there's no warning see also rust-lang/rust#76053 thanks @jannic
rursprung
force-pushed
the
migrate-to-eh1.rc1
branch
from
7a4e5b3 to
204de95
Compare
see the individual commit messages for more details
this resolves #3