Add imprecise_flops lint

Add lint to detect floating point operations that can be computed more
accurately at the cost of performance. `cbrt`, `ln_1p` and `exp_m1`
library functions call their equivalent cmath implementations which is
slower but more accurate so moving checks for these under this new lint.
This commit is contained in:
Krishna Sai Veera Reddy 2020-02-23 21:06:55 -08:00
parent e94a167508
commit ff0d44e45a
14 changed files with 67 additions and 20 deletions

View file

@ -1169,6 +1169,7 @@ Released 2018-09-13
[`ifs_same_cond`]: https://rust-lang.github.io/rust-clippy/master/index.html#ifs_same_cond
[`implicit_hasher`]: https://rust-lang.github.io/rust-clippy/master/index.html#implicit_hasher
[`implicit_return`]: https://rust-lang.github.io/rust-clippy/master/index.html#implicit_return
[`imprecise_flops`]: https://rust-lang.github.io/rust-clippy/master/index.html#imprecise_flops
[`inconsistent_digit_grouping`]: https://rust-lang.github.io/rust-clippy/master/index.html#inconsistent_digit_grouping
[`indexing_slicing`]: https://rust-lang.github.io/rust-clippy/master/index.html#indexing_slicing
[`ineffective_bit_mask`]: https://rust-lang.github.io/rust-clippy/master/index.html#ineffective_bit_mask

View file

@ -5,7 +5,7 @@
A collection of lints to catch common mistakes and improve your [Rust](https://github.com/rust-lang/rust) code.
[There are 357 lints included in this crate!](https://rust-lang.github.io/rust-clippy/master/index.html)
[There are 358 lints included in this crate!](https://rust-lang.github.io/rust-clippy/master/index.html)
We have a bunch of lint categories to allow you to choose how much Clippy is supposed to ~~annoy~~ help you:

View file

@ -16,6 +16,39 @@ use std::f64::consts as f64_consts;
use sugg::{format_numeric_literal, Sugg};
use syntax::ast;
declare_clippy_lint! {
/// **What it does:** Looks for floating-point expressions that
/// can be expressed using built-in methods to improve accuracy
/// at the cost of performance.
///
/// **Why is this bad?** Negatively impacts accuracy.
///
/// **Known problems:** None
///
/// **Example:**
///
/// ```rust
///
/// let a = 3f32;
/// let _ = a.powf(1.0 / 3.0);
/// let _ = (1.0 + a).ln();
/// let _ = a.exp() - 1.0;
/// ```
///
/// is better expressed as
///
/// ```rust
///
/// let a = 3f32;
/// let _ = a.cbrt();
/// let _ = a.ln_1p();
/// let _ = a.exp_m1();
/// ```
pub IMPRECISE_FLOPS,
nursery,
"usage of imprecise floating point operations"
}
declare_clippy_lint! {
/// **What it does:** Looks for floating-point expressions that
/// can be expressed using built-in methods to improve both
@ -34,12 +67,9 @@ declare_clippy_lint! {
/// let _ = (2f32).powf(a);
/// let _ = E.powf(a);
/// let _ = a.powf(1.0 / 2.0);
/// let _ = a.powf(1.0 / 3.0);
/// let _ = a.log(2.0);
/// let _ = a.log(10.0);
/// let _ = a.log(E);
/// let _ = (1.0 + a).ln();
/// let _ = a.exp() - 1.0;
/// let _ = a.powf(2.0);
/// let _ = a * 2.0 + 4.0;
/// ```
@ -53,12 +83,9 @@ declare_clippy_lint! {
/// let _ = a.exp2();
/// let _ = a.exp();
/// let _ = a.sqrt();
/// let _ = a.cbrt();
/// let _ = a.log2();
/// let _ = a.log10();
/// let _ = a.ln();
/// let _ = a.ln_1p();
/// let _ = a.exp_m1();
/// let _ = a.powi(2);
/// let _ = a.mul_add(2.0, 4.0);
/// ```
@ -67,7 +94,10 @@ declare_clippy_lint! {
"usage of sub-optimal floating point operations"
}
declare_lint_pass!(FloatingPointArithmetic => [SUBOPTIMAL_FLOPS]);
declare_lint_pass!(FloatingPointArithmetic => [
IMPRECISE_FLOPS,
SUBOPTIMAL_FLOPS
]);
// Returns the specialized log method for a given base if base is constant
// and is one of 2, 10 and e
@ -156,7 +186,7 @@ fn check_ln1p(cx: &LateContext<'_, '_>, expr: &Expr<'_>, args: &[Expr<'_>]) {
span_lint_and_sugg(
cx,
SUBOPTIMAL_FLOPS,
IMPRECISE_FLOPS,
expr.span,
"ln(1 + x) can be computed more accurately",
"consider using",
@ -215,18 +245,21 @@ fn check_powf(cx: &LateContext<'_, '_>, expr: &Expr<'_>, args: &[Expr<'_>]) {
// Check argument
if let Some((value, _)) = constant(cx, cx.tables, &args[1]) {
let (help, suggestion) = if F32(1.0 / 2.0) == value || F64(1.0 / 2.0) == value {
let (lint, help, suggestion) = if F32(1.0 / 2.0) == value || F64(1.0 / 2.0) == value {
(
SUBOPTIMAL_FLOPS,
"square-root of a number can be computed more efficiently and accurately",
format!("{}.sqrt()", Sugg::hir(cx, &args[0], "..")),
)
} else if F32(1.0 / 3.0) == value || F64(1.0 / 3.0) == value {
(
IMPRECISE_FLOPS,
"cube-root of a number can be computed more accurately",
format!("{}.cbrt()", Sugg::hir(cx, &args[0], "..")),
)
} else if let Some(exponent) = get_integer_from_float_constant(&value) {
(
SUBOPTIMAL_FLOPS,
"exponentiation with integer powers can be computed more efficiently",
format!(
"{}.powi({})",
@ -240,7 +273,7 @@ fn check_powf(cx: &LateContext<'_, '_>, expr: &Expr<'_>, args: &[Expr<'_>]) {
span_lint_and_sugg(
cx,
SUBOPTIMAL_FLOPS,
lint,
expr.span,
help,
"consider using",
@ -264,7 +297,7 @@ fn check_expm1(cx: &LateContext<'_, '_>, expr: &Expr<'_>) {
then {
span_lint_and_sugg(
cx,
SUBOPTIMAL_FLOPS,
IMPRECISE_FLOPS,
expr.span,
"(e.pow(x) - 1) can be computed more accurately",
"consider using",

View file

@ -538,6 +538,7 @@ pub fn register_plugins(store: &mut rustc_lint::LintStore, sess: &Session, conf:
&fallible_impl_from::FALLIBLE_IMPL_FROM,
&float_literal::EXCESSIVE_PRECISION,
&float_literal::LOSSY_FLOAT_LITERAL,
&floating_point_arithmetic::IMPRECISE_FLOPS,
&floating_point_arithmetic::SUBOPTIMAL_FLOPS,
&format::USELESS_FORMAT,
&formatting::POSSIBLE_MISSING_COMMA,
@ -1648,6 +1649,7 @@ pub fn register_plugins(store: &mut rustc_lint::LintStore, sess: &Session, conf:
store.register_group(true, "clippy::nursery", Some("clippy_nursery"), vec![
LintId::of(&attrs::EMPTY_LINE_AFTER_OUTER_ATTR),
LintId::of(&fallible_impl_from::FALLIBLE_IMPL_FROM),
LintId::of(&floating_point_arithmetic::IMPRECISE_FLOPS),
LintId::of(&floating_point_arithmetic::SUBOPTIMAL_FLOPS),
LintId::of(&missing_const_for_fn::MISSING_CONST_FOR_FN),
LintId::of(&mutable_debug_assertion::DEBUG_ASSERT_WITH_MUT_CALL),

View file

@ -6,7 +6,7 @@ pub use lint::Lint;
pub use lint::LINT_LEVELS;
// begin lint list, do not remove this comment, its used in `update_lints`
pub const ALL_LINTS: [Lint; 357] = [
pub const ALL_LINTS: [Lint; 358] = [
Lint {
name: "absurd_extreme_comparisons",
group: "correctness",
@ -749,6 +749,13 @@ pub const ALL_LINTS: [Lint; 357] = [
deprecation: None,
module: "implicit_return",
},
Lint {
name: "imprecise_flops",
group: "nursery",
desc: "usage of imprecise floating point operations",
deprecation: None,
module: "floating_point_arithmetic",
},
Lint {
name: "inconsistent_digit_grouping",
group: "style",

View file

@ -1,5 +1,5 @@
// run-rustfix
#![warn(clippy::suboptimal_flops)]
#![warn(clippy::imprecise_flops)]
fn main() {
let x = 2f32;

View file

@ -1,5 +1,5 @@
// run-rustfix
#![warn(clippy::suboptimal_flops)]
#![warn(clippy::imprecise_flops)]
fn main() {
let x = 2f32;

View file

@ -4,7 +4,7 @@ error: (e.pow(x) - 1) can be computed more accurately
LL | let _ = x.exp() - 1.0;
| ^^^^^^^^^^^^^ help: consider using: `x.exp_m1()`
|
= note: `-D clippy::suboptimal-flops` implied by `-D warnings`
= note: `-D clippy::imprecise-flops` implied by `-D warnings`
error: (e.pow(x) - 1) can be computed more accurately
--> $DIR/floating_point_exp.rs:7:13

View file

@ -1,6 +1,6 @@
// run-rustfix
#![allow(dead_code, clippy::double_parens)]
#![warn(clippy::suboptimal_flops)]
#![warn(clippy::suboptimal_flops, clippy::imprecise_flops)]
const TWO: f32 = 2.0;
const E: f32 = std::f32::consts::E;

View file

@ -1,6 +1,6 @@
// run-rustfix
#![allow(dead_code, clippy::double_parens)]
#![warn(clippy::suboptimal_flops)]
#![warn(clippy::suboptimal_flops, clippy::imprecise_flops)]
const TWO: f32 = 2.0;
const E: f32 = std::f32::consts::E;

View file

@ -53,6 +53,8 @@ error: ln(1 + x) can be computed more accurately
|
LL | let _ = (1f32 + 2.).ln();
| ^^^^^^^^^^^^^^^^ help: consider using: `2.0f32.ln_1p()`
|
= note: `-D clippy::imprecise-flops` implied by `-D warnings`
error: ln(1 + x) can be computed more accurately
--> $DIR/floating_point_log.rs:25:13

View file

@ -1,5 +1,5 @@
// run-rustfix
#![warn(clippy::suboptimal_flops)]
#![warn(clippy::suboptimal_flops, clippy::imprecise_flops)]
fn main() {
let x = 3f32;

View file

@ -1,5 +1,5 @@
// run-rustfix
#![warn(clippy::suboptimal_flops)]
#![warn(clippy::suboptimal_flops, clippy::imprecise_flops)]
fn main() {
let x = 3f32;

View file

@ -47,6 +47,8 @@ error: cube-root of a number can be computed more accurately
|
LL | let _ = x.powf(1.0 / 3.0);
| ^^^^^^^^^^^^^^^^^ help: consider using: `x.cbrt()`
|
= note: `-D clippy::imprecise-flops` implied by `-D warnings`
error: exponentiation with integer powers can be computed more efficiently
--> $DIR/floating_point_powf.rs:14:13