easy-cast

This version is a substantial revision of easy-cast, introducing generics over rounding modes via the new Rounding trait.

Simple usage via the Conv, Cast, ConvApprox and CastApprox traits is largely unchanged. Usage with other rounding modes should be moved to the new ConvTo and CastTo traits (e.g. i32::conv_to(Nearest, 5.3), x.cast_to(Floor)).

External implementations of conversions should be moved to the ConvExact and ConvTo traits.

Added:

• Rounding trait with Exact, Approx rounding modes (#58)
• Trunc, Nearest, Floor, Ceil rounding modes (#61)
• ConvExact trait as an implementation helper for conversions which are never inexact (#58, #65)
• ConvTo trait for implementation and usage of conversions with a specific rounding mode (#58, #60, #63, #65)
• CastTo trait for casting with a specific rounding mode (#58, #60, #63)
• RangeError as a new error type (#56)

Changed:

• Added an associated Error type to the Conv and Cast traits (#62)
• Rename impl_via_trivial! to impl_via_identity! (#68)
• Implement Exact f64 → f32 conversion (#59)
• Implement Approx and Nearest int-to-float conversions (#59, #63)
• u128::MAX and other too-large values no longer coerce to f32::INFINITY under Exact rounding, as used by Conv and Cast (#67)
• Truncating-conversions (Trunc) no longer require the std or libm feature (#66)
• Implement core::error::Error on no-std (#57)
• Enabled the doc_cfg feature (#66)

Removed:

• Result type-def (#56)
• ConvFloat, CastFloat traits; this functionality remains available through the new rounding modes (#61)

Full Changelog: 0.6.1...0.7.0

This version is a substantial revision of easy-cast, introducing generics over rounding modes via the new Rounding trait.

Simple usage via the Conv, Cast, ConvApprox and CastApprox traits is largely unchanged. Usage with other rounding modes should be moved to the new ConvTo and CastTo traits (e.g. i32::conv_to(Nearest, 5.3), x.cast_to(Floor)).

External implementations of conversions should be moved to the ConvExact and ConvTo traits.

Added:

• Rounding trait with Exact, Approx rounding modes (#58)
• Trunc, Nearest, Floor, Ceil rounding modes (#61)
• ConvExact trait as an implementation helper for conversions which are never inexact (#58, #65)
• ConvTo trait for implementation and usage of conversions with a specific rounding mode (#58, #60, #63, #65)
• CastTo trait for casting with a specific rounding mode (#58, #60, #63)
• RangeError as a new error type (#56)

Changed:

• Added an associated Error type to the Conv and Cast traits (#62)
• Rename impl_via_trivial! to impl_via_identity! (#68)
• Implement Exact f64 → f32 conversion (#59)
• Implement Approx and Nearest int-to-float conversions (#59, #63)
• u128::MAX and other too-large values no longer coerce to f32::INFINITY under Exact rounding, as used by Conv and Cast (#67)
• Truncating-conversions (Trunc) no longer require the std or libm feature (#66)
• Implement core::error::Error on no-std (#57)
• Enabled the doc_cfg feature (#66)

Removed:

• Result type-def (#56)
• ConvFloat, CastFloat traits; this functionality remains available through the new rounding modes (#61)

Added:

• Add pub macros impl_via_from!, impl_via_trivial! (#54)

Full Changelog: 0.6.0...0.6.1

Added:

• Add pub macros impl_via_from!, impl_via_trivial! (#54)

This release contains some breaking changes which should be easy/trivial to accommodate.

Fixed:

• Fix unsafe code in NonZero conversion on input values which wrap to zero (#52)
• Fix: allow overflow of large u128 values to f32::INFINITY (#51)

Changed:

• Adjust ConvApprox<f64> for f32 to behave more like *_f64 as f32: too-large values round to infinity (instead of a range error) and rounding uses roundTiesToEven. Unlike as f32, NAN still yields Err(Error::Range). (#47)
• Change Conv<f32> for f64 to trap NAN (#48)

Removed:

• Remove support for Saturating, Wrapping (#49)
• Remove assert_nonzero feature flag (#52)

Full Changelog: 0.5.5...0.6.0

Fixed:

• Fix unsafe code in NonZero conversion on input values which wrap to zero (#52)
• Fix: allow overflow of large u128 values to f32::INFINITY (#51)

Changed:

• Adjust ConvApprox<f64> for f32 to behave more like *_f64 as f32: too-large values round to infinity (instead of a range error) and rounding uses roundTiesToEven. Unlike as f32, NAN still yields Err(Error::Range). (#47)
• Change Conv<f32> for f64 to trap NAN (#48)

Removed:

• Remove support for Saturating, Wrapping (#49)
• Remove assert_nonzero feature flag (#52)

Support core::range types (MSRV = 1.96.0) by @dhardy in #44

Full Changelog: 0.5.4...0.5.5

• Bump MSRV to 1.96.0 and use Edition 2024 (#44)
• Support std::range types (#44)

What's Changed

• Bump MSRV to 1.79.0
• Support Conv, Cast for core::ops::Range and other range types (#38)
• Support Conv, Cast for core::num::{Saturating, Wrapping} (#38)
• Support Conv, Cast for core::num::NonZero (#39)

Full Changelog: 0.5.3...0.5.4

• Bump MSRV to 1.79.0
• Support Conv, Cast for core::ops::Range and other range types (#38)
• Support Conv, Cast for core::num::{Saturating, Wrapping} (#38)
• Support Conv, Cast for core::num::NonZero (#39)

What's Changed

• Bump MSRV to 1.60.0 and use Edition 2021 (#30)

Full Changelog: 0.5.2...0.5.3

• Bump MSRV to 1.60.0 and use Edition 2021 (#30)

• Support cast to self, useful in generics (#28)

• Support cast to self, useful in generics (#28)

• Document feature assert_digits and imply by always_assert (#26)

• Document feature assert_digits and imply by always_assert (#26)

• Reorganise code (#20)
• Bump MSRV to 1.53.0 (#21)
• Add default implementations of conv methods over equivalent try_conv
  variant to facilitate custom implementations (#21)
• Add ConvApprox and CastApprox, supporting approximate conversions
  with unspecified rounding, implemented for f64 → f32 and all type
  conversions supported by Conv (#21)
• Support Conv and ConvFloat for arrays and tuples (#21)
• Remove impl<T> Conv<T> for T (#21)
• Add easy_cast::Result type alias (#22)
• Move traits into new easy_cast::traits public module (#23)

• Reorganise code (#20)
• Bump MSRV to 1.53.0 (#21)
• Add default implementations of conv methods over equivalent try_conv variant to facilitate custom implementations (#21)
• Add ConvApprox and CastApprox, supporting approximate conversions with unspecified rounding, implemented for f64 → f32 and all type conversions supported by Conv (#21)
• Support Conv and ConvFloat for arrays and tuples (#21)
• Remove impl<T> Conv<T> for T (#21)
• Add easy_cast::Result type alias (#22)
• Move traits into new easy_cast::traits public module (#23)

Fix negative int to float cast digits check

• Fix negative int to float digits check (#18)

• Unify some macros via impl_int_generic (#16)
• Improve error messages in asserts (#17)

• Fix i16::conv(1usize) (#15)
• Update README (#15)

• Fix conv(0) from int to float (#14)

• Add try_conv and try_cast methods (#12)
• Add try_conv_nearest etc. (#12)
• Removed Conv<f64> for f32 (#12)
• Replaced assert_range and assert_non_neg with assert_int (#12)
• MSRV is 1.32.0 (#12)

• Add conv_trunc / cast_trunc (#11)
• Explicitly support Rust 1.36.0 (and potentially older; #10)
• Support no_std (#10)
• Fix rounding for floor on negative values (#10)

• Add feature flags controlling assert behaviour
• Remove restrictions on isize/usize (#6)
• Fix bad revert checks (#6)