Better copy for lint message

Since x.log(y) is actually implemented as x.ln() / y.ln()

Author: thiagoarrais

clippy_lints/src/floating_point_arithmetic.rs

@@ -618,7 +618,7 @@ fn check_log_division(cx: &LateContext<'_>, expr: &Expr<'_>) {
                 cx,
                 SUBOPTIMAL_FLOPS,
                 expr.span,
-                "division of logarithms can be calculated more efficiently and accurately",
+                "log base can be expressed more clearly",
                 "consider using",
                 format!("{}.log({})", Sugg::hir(cx, &largs[0], ".."), Sugg::hir(cx, &rargs[0], ".."),),
                 Applicability::MachineApplicable,

tests/ui/floating_point_logbase.stderr

@@ -1,24 +1,24 @@
-error: division of logarithms can be calculated more efficiently and accurately
+error: log base can be expressed more clearly
   --> $DIR/floating_point_logbase.rs:7:13
    |
 LL |     let _ = x.ln() / y.ln();
    |             ^^^^^^^^^^^^^^^ help: consider using: `x.log(y)`
    |
    = note: `-D clippy::suboptimal-flops` implied by `-D warnings`
 
-error: division of logarithms can be calculated more efficiently and accurately
+error: log base can be expressed more clearly
   --> $DIR/floating_point_logbase.rs:8:13
    |
 LL |     let _ = x.log2() / y.log2();
    |             ^^^^^^^^^^^^^^^^^^^ help: consider using: `x.log(y)`
 
-error: division of logarithms can be calculated more efficiently and accurately
+error: log base can be expressed more clearly
   --> $DIR/floating_point_logbase.rs:9:13
    |
 LL |     let _ = x.log10() / y.log10();
    |             ^^^^^^^^^^^^^^^^^^^^^ help: consider using: `x.log(y)`
 
-error: division of logarithms can be calculated more efficiently and accurately
+error: log base can be expressed more clearly
   --> $DIR/floating_point_logbase.rs:10:13
    |
 LL |     let _ = x.log(5f32) / y.log(5f32);