5b7d05b775
docs: change wording from truncate to discard fractional bits
...
Truncation might be interpreted to mean rounding towards zero, which
is not what happens in negative two's-complement fixed-point numbers.
So "extra fractional bits are truncated" is better as "extra
fractional bits are discarded, which rounds towards −∞."
2020-05-10 18:43:22 +02:00
961d36aa2a
update half dependency to 1.6
2020-05-10 10:49:09 +02:00
c13ab2e324
add wrapping_next_power_of_two
2020-05-10 01:13:24 +02:00
dfb59cefc0
add leading_ones and trailing_ones
2020-05-10 01:09:41 +02:00
609f75575f
fmt nit in doc example
2020-05-08 15:55:49 +02:00
2b4e051ded
add some tests
2020-05-08 15:43:01 +02:00
e3fe9b7f9c
implement LosslessTryFrom for lossless integer <-> float combinations
2020-05-08 15:19:09 +02:00
c2847491cb
implement LosslessTryFrom for lossless fixed <-> float combinations
2020-05-08 14:23:23 +02:00
c196899bcc
implement LosslessTryFrom for all fixed<U0> -> integer combinations
2020-05-08 13:31:50 +02:00
234dbfecdc
implement LosslessTryFrom for all integer -> fixed combinations
2020-05-08 13:20:47 +02:00
bc5e01120f
add LosslessTry{From,Into} and immplement them for Fixed <-> Fixed
2020-05-08 12:41:35 +02:00
f23ddc6616
improve first line of module docs
2020-05-08 00:51:56 +02:00
eebb76e7b0
fix typos in tests and release notes
2020-05-08 00:37:14 +02:00
232decb692
add names to some constants
2020-05-07 23:48:48 +02:00
37167fddd0
add golden ratio and its conjugate to constants
2020-05-07 23:36:35 +02:00
eb7e369775
document impl blocks
2020-05-03 17:56:04 +02:00
ff0874a1f4
require rustc 1.43.0
...
Will probably bump requirement to 1.44.0 before releasing 1.0.0 so
that {from,to}_{le,be,ne}_bytes can be const fn.
2020-05-03 16:08:11 +02:00
48a763c01f
tweak wording on version 2 plan
2020-05-02 21:22:46 +02:00
eb51b85db5
remove reimplementation of wrapping_signum in Wrapping::signum
2020-05-02 21:19:44 +02:00
879861b875
remove deprecated items
2020-05-01 18:28:17 +02:00
079a13ca1b
version 0.5.6
2020-05-01 18:15:40 +02:00
86b542e7f3
add Lossy{From,Into} to prelude
2020-05-01 17:40:53 +02:00
672973c99f
Revert "remove deprecated items"
...
This reverts commit f67aed0078
2020-04-18 19:13:08 +02:00
74f0e4f575
add checked versions of signum
2020-04-18 18:15:22 +02:00
339bd1f5aa
specify version 2 when talking about const generics plan
2020-04-18 02:29:04 +02:00
f67aed0078
remove deprecated items
2020-04-18 02:21:17 +02:00
f8f9bd69ba
add some coverage
2020-04-17 22:11:57 +02:00
8d8fde75c5
remove unused functions
2020-04-17 21:48:25 +02:00
a47367fa6b
make reasoning clearer in log10
2020-04-17 21:31:13 +02:00
ed49d69e18
test constant LOG2_10 and LOG10_2 with f16 and bf16
2020-04-17 17:21:37 +02:00
800ed3a58c
az::StaticCast is deprecated
2020-04-17 17:20:25 +02:00
9bb95c5a5c
remove unused macro parameter
2020-04-17 00:01:00 +02:00
97d4e1a326
version 0.5.5
2020-04-16 23:29:50 +02:00
289f4b1697
add constants in consts module as associated constants
2020-04-16 23:08:29 +02:00
68c0fa6f8d
include generator in consts.rs
2020-04-16 23:05:53 +02:00
5db3b17284
fix comparison of e.g. I0F16 with +0.5
2020-04-16 21:19:30 +02:00
4c25f002a1
update copyright date
2020-04-13 20:11:39 +02:00
36ef9cae03
use macro to avoid repetitions in log10 tests
2020-04-13 18:12:56 +02:00
f038d17921
some log10 cleanups
2020-04-13 05:23:29 +02:00
de69767a3c
fmt
2020-04-13 04:28:23 +02:00
b0f6deab9e
some log10 optimizations
2020-04-13 04:27:36 +02:00
a32b56ab43
add int_log10 and checked_int_log10
2020-04-13 03:17:21 +02:00
919519fea1
doc nits
2020-04-13 00:27:42 +02:00
e8f3ff2c67
replace {min,max}_value() by {MIN,MAX} in examples
...
There are hidden import lines
use std::primitive;
that should be removed once MIN and MAX are supported by the minimum
supported rustc.
2020-04-11 19:23:58 +02:00
47287ab98f
replace {int,frac}_nbits() by {INT,FRAC}_NBITS
2020-04-11 18:50:26 +02:00
e1febebb31
replace {min,max}_value() by {MIN,MAX}
2020-04-11 18:34:52 +02:00
91bd73bca4
add int_log2 and checked_int_log2
2020-04-11 17:44:09 +02:00
060b58af13
doc: use [`true`] instead of [`true`][`bool`]
...
This is because link markdown is not interpreted in search results, so
at least bool is not displayed.
2020-04-11 16:54:22 +02:00
b41c726da5
test constant LOG2_10 and LOG10_2
...
Before this commit, the tests were skipped because the corresponding
standard library constants were unstable, but the values themselves
could easily be computed for the test. Once the minimum rustc
supported becomes 1.43.0, the tests should be changed to use the
constants instead of the computation.
2020-04-08 13:23:04 +02:00
b035ae9dbe
typo
2020-04-08 12:51:11 +02:00
8803e0786d
docs: link to F::is_{finite,nan} in impl {To,From}Fixed for F
...
Before this commit, all links were to the f64 documentation, even for
implementations for f32, f16 and bf16.
2020-03-20 01:59:51 +01:00
f86e59a204
document individual conversion traits
2020-03-05 15:30:08 +01:00
aeac17c0d8
fix failing cmp_parse test
...
While with the actual string
f ≤ string < f+Δ
we are using a parsed version, which could be actually larger than
string so that the test should only check
f ≤ parsed ≤ f+Δ
2020-02-24 14:29:34 +01:00
c6d8e14537
test that constants are truncated version of known correct strings
2020-02-24 14:16:23 +01:00
327bc05413
specify that <Err = PraseFixedError> in Fixed: FromStr constraint
2020-02-24 14:13:30 +01:00
524fc0c28f
version 0.5.4
2020-02-21 16:34:27 +01:00
97fdf52ac6
silence clippy
2020-02-21 15:57:56 +01:00
e0f3eb77b8
fix incorrect overflow handling for rem_euclid_int
2020-02-21 15:39:02 +01:00
e8f5a8238e
tweak Fixed doc
2020-02-13 15:57:12 +01:00
5ed4979048
version 0.5.3
2020-02-13 13:48:03 +01:00
4b889cbb1f
reorganize {,checked} {,Euclidean} remainders {,by integers}
...
Now the main logic is in the checked functions, which fail only if rhs
is zero, and the other functions all .expect("division by zero").
checked_rem now uses % instead of core checked_rem.
checked_rem_euclid now uses core rem_euclid.
checked_rem_int uses checked_rem for the most common case.
checked_rem_euclid_int uses checked_rem_euclid for the most common
case.
2020-02-13 12:04:49 +01:00
68f72f8b56
some fixes and tests
...
* rem_euclid no longer uses rhs.abs() which could overflow.
* Rem<Inner> and rem_euclid_int now allow for the signed case where
rhs is equal to the absolute value of the minimum fixed value.
* checked_rem_int, checked_rem_euclid and checked_rem_euclid_int now
check for zero then defer to Rem<Inner>, rem_euclid and
rem_euclid_int.
2020-02-13 01:39:02 +01:00
be49c12e93
typo
2020-02-12 23:58:44 +01:00
137d8989e6
mention fixed-sqrt in top-level doc
2020-02-12 23:56:26 +01:00
e3e17341f2
add {div,rem}_euclid_int to Wrapping
2020-02-12 23:51:31 +01:00
873bdfa128
fix bug in round_to_zero
2020-02-12 23:45:58 +01:00
ddf69cdca0
some doc tweaks
2020-02-12 23:18:24 +01:00
3ee5811556
fix remaining checked {div,rem}_euclid_int methods
...
Closes https://gitlab.com/tspiteri/fixed/issues/13
2020-02-12 18:40:04 +01:00
97128d2b9f
fix checked_rem_int and fix+deprecate {wrapping,overflowing}_rem_int
...
https://gitlab.com/tspiteri/fixed/issues/13
2020-02-12 17:32:34 +01:00
677ee97f6f
add more arithmetic tests
2020-02-12 16:59:10 +01:00
f8d2004c8f
fix Rem<Inner> for Fixed
...
https://gitlab.com/tspiteri/fixed/issues/13
2020-02-12 16:58:21 +01:00
f070f5e75a
more of remove Frac constraints for trait impls that do not need it
2020-02-12 16:44:26 +01:00
f6f4b9d174
add {div,rem}_euclid to Wrapping
2020-02-12 15:07:15 +01:00
bf7c2024e1
add {saturating,wrapping,overflowing}_div_euclid
2020-02-12 14:48:10 +01:00
43926daa62
add checked_{div,rem}_euclid
2020-02-12 14:16:19 +01:00
96e922d287
remove Frac constraints for trait impls that do not need it
2020-02-12 13:57:20 +01:00
6e93affa0d
add div_euclid and rem_euclid
2020-02-12 13:48:35 +01:00
c2fa7448c2
implement Rem and RemAssign, and add checked_rem
...
Closes https://gitlab.com/tspiteri/fixed/issues/12
2020-02-12 13:15:23 +01:00
2f2b05e33f
version 0.5.2
2020-02-02 21:02:14 +01:00
b0e3828e79
re-use inner serde impls for Wrapping types
2020-02-01 03:01:11 -08:00
7aa5b0844f
add serde impls for Wrapping types
2020-02-01 01:26:35 -08:00
ab0aa534c7
version 0.5.1
2019-12-23 18:15:31 +01:00
823916346f
add std feature
2019-12-23 17:39:41 +01:00
59bc0bfd26
link to core libs instead of std libs where possible
2019-12-23 17:22:21 +01:00
be35c826da
simplify overflow checks for multiplication and division
2019-12-07 22:40:09 +01:00
950f9051f7
version 0.5.0
2019-12-06 11:46:41 +01:00
ae0b037bc5
fmt
2019-11-24 21:17:19 +01:00
68b2190a86
add conversions to/from {be,le,ne}_bytes
2019-11-24 20:13:26 +01:00
aac0353bb1
cleaner is_power_of_two
2019-11-24 18:53:45 +01:00
e46bf604f3
remove deprecated methods and modules
2019-11-24 18:53:07 +01:00
25510f1525
use std abs, wrapping_abs, overflowing_abs
2019-11-24 18:52:57 +01:00
221e7d22eb
use std euclid operations
2019-11-24 18:47:39 +01:00
7b74e7333b
require rustc 1.39.0
2019-11-24 18:47:06 +01:00
27380f19ca
add Euclidean division/remainder by an integer to Fixed trait
2019-10-17 15:28:41 +02:00
9880f75c42
add Euclidean division/remainder by an integer
2019-10-17 15:00:37 +02:00
4ef5f192ed
version 0.4.6
2019-10-16 12:08:30 +02:00
a674c77051
add bf16 support
2019-10-16 11:32:53 +02:00
5e7473b7f5
make saturating ops const fn
2019-10-11 14:59:38 +02:00
0207311b6c
fmt
2019-10-10 23:31:05 +02:00
f3dc9cdb04
implement az casts from bool to fixed-point
2019-10-10 19:57:44 +02:00
a99655c809
implement az casts with both source and destination as fixed-point
2019-10-10 19:28:42 +02:00
b6c1974401
clippy
2019-10-10 17:19:10 +02:00
cfe198f087
document and test az option
2019-10-10 17:13:36 +02:00
d6c634c915
make az support optional
2019-10-10 17:13:27 +02:00
019c1eeb23
add az support
2019-10-10 16:47:10 +02:00
3248e354ba
version 0.4.5
2019-08-30 22:32:17 +02:00
687a5521a7
Wrapping nits
2019-08-30 21:55:26 +02:00
e5afe1d6a4
add more tests in {display,from_str}.rs
2019-08-30 21:30:40 +02:00
1960c702c6
remove unnecessary helper methods
2019-08-30 21:28:19 +02:00
55a58bf795
typos
2019-08-30 13:01:42 +02:00
a4c3fdeeed
add methods to Wrapping
2019-08-30 12:48:31 +02:00
e31802dd23
some more display/from_str tests
2019-08-29 22:30:56 +02:00
24f197b930
remove unnecessary macros
2019-08-29 17:24:36 +02:00
75df9749ff
add compare_frac17_float test
2019-08-29 14:20:56 +02:00
1a81cef6eb
improve test for previous commit
2019-08-29 12:38:22 +02:00
f314735cc9
bug fix: decimal display panics or includes leading zero
2019-08-29 11:50:18 +02:00
de6ecb5782
add Wrapping::round_ties_to_even
2019-08-29 10:12:17 +02:00
0711850c26
doc nits
2019-08-26 11:10:07 +02:00
980804b1c7
do not use cfg_attr for clippy allows
2019-08-24 16:35:03 +02:00
ee98c16b28
simplify ceil_log10_2_times
2019-08-24 16:32:28 +02:00
06af80c9ed
nits
2019-08-24 14:52:32 +02:00
50ca556686
version 0.4.4
2019-08-24 13:15:28 +02:00
82c7781f0e
update README and doc main page
2019-08-24 13:10:49 +02:00
97e57bad0e
fix unused warning
2019-08-23 14:16:36 +02:00
c6e0563257
clippy + fmt
2019-08-23 14:06:28 +02:00
72f619b757
do not use f16::from_{f32,f64} from half crate
2019-08-23 14:03:39 +02:00
348463d204
state aims at top of documentation
2019-08-23 10:03:59 +02:00
e24a0963f6
add Frac type to Fixed trait
2019-08-22 19:30:52 +02:00
3292a2da26
some minor doc improvements
2019-08-22 14:26:08 +02:00
87d30cd8b1
reorder some methods
2019-08-22 14:03:32 +02:00
9e333d99c2
add round_to_zero
2019-08-22 13:39:51 +02:00
7d08654e0a
cleaner rounding examples
2019-08-22 13:04:39 +02:00
0eade78071
add round_ties_to_even and checked variants
2019-08-22 13:01:28 +02:00
b3796a46f0
make is_power_of_two a const fn
2019-08-21 22:15:36 +02:00
696361a8fa
macro formatting
2019-08-21 16:57:52 +02:00
381af22e96
make {,wrapping_,overflowing_}abs const
2019-08-21 15:46:24 +02:00
895d59611d
fix and expand rounding test
2019-08-21 15:45:55 +02:00
ba594f24e1
document bug fix in release notes
2019-08-21 13:30:49 +02:00
08fc9881e5
make display code clearer
2019-08-21 13:02:43 +02:00
8176a36a95
version 0.4.3
2019-08-20 19:40:53 +02:00
f423fe0159
impl {Shl,Shr}{,Assign}<all primitives> for Wrapping
2019-08-20 17:07:13 +02:00
fa271a4a9d
doc updates
2019-08-20 16:42:37 +02:00
46049bb7ac
tweak consts example
2019-08-20 15:31:39 +02:00
0b4dac8d33
reorganize modules
2019-08-20 15:30:23 +02:00
96107159f5
some cleanups in tests inside display.rs
2019-08-20 01:18:22 +02:00
daa600bb71
properly handle precision in Display
2019-08-19 23:53:43 +02:00
2d5a5c3ada
properly handle precision in Binary, Octal, LowerHex, UpperHex
2019-08-19 18:49:40 +02:00
76a0abe468
relax some trait bounds in Wrapping
...
* Now bitwise and shifting operations do not require F: Fixed, only
F: Op. For example
impl BitAnd<Wrapping<F>> for Wrapping<F>
is now bound by
where F: BitAnd<F, Output = F>
instead of
where F: Fixed
* Also multiplication and division by integers do not require Frac
bounds. For example
impl Mul<i8> for Wrapping<FixedI8<Frac>>
now has no bounds on Frac, instead of being bound by
where Frac: LeEqU8
2019-08-19 14:15:06 +02:00
a501829cc6
change rhs type of Wrapping shift operators from usize to u32
2019-08-19 13:18:54 +02:00
960a782799
consolidate the trait impls for Wrapping
2019-08-19 02:02:06 +02:00
Trevor Spiteri
Shane Pearman