Fix description of sign to match the logic found in the constructor

Author: mborland

doc/decimal/basics.adoc

@@ -23,7 +23,7 @@ constexpr decimal32(T coeff, T2 exp, bool sign = false) noexcept;
 Where types `T` and `T2` are integral types (signed and unsigned are both allowed).
 Lastly the sign follows the convention of `signbit` where `false` is positive and `true` is negative.
 If both a negative coefficient and a sign are passed then the resulting decimal number will be negative.
-The final number constructed is in the form (sign ? -1 : 1) x coeff x 10^exp.
+The final number constructed is in the form (sign || coeff < 0 ? -1 : 1) x coeff x 10^exp.
 
 [souce, c++]
 ----