Fix unprecised float number parsing

Author: Hubert de La Jonquière

grammar/fr/src/rules_number.rs

@@ -195,27 +195,27 @@ pub fn rules_numbers(b: &mut RuleSetBuilder<Dimension>) -> RustlingResult<()> {
                           FloatValue::new(value)
                       });
     b.rule_3("number dot number",
-             number_check!(|number: &NumberValue| !number.prefixed()),
+             integer_check!(|integer: &IntegerValue| !integer.prefixed),
              b.reg(r#"virgule|point"#)?,
-             number_check!(|number: &NumberValue| !number.suffixed()),
+             integer_check!(|integer: &IntegerValue| !integer.suffixed),
              |a, _, b| {
-                 let power = b.value().value().to_string().chars().count();
-                 let coeff = 10.0_f64.powf(-1.0 * power as f64);
+                let value: f64 = format!("{}.{}", a.value().value, b.value().value).parse()?;
                  Ok(FloatValue {
-                     value: b.value().value() * coeff + a.value().value(),
+                     value,
                      ..FloatValue::default()
                  })
              });
     b.rule_4("number dot zero ... number",
-             number_check!(|number: &NumberValue| !number.prefixed()),
+             integer_check!(|integer: &IntegerValue| !integer.prefixed),
              b.reg(r#"virgule|point"#)?,
              b.reg(r#"(?:(?:z[eé]ro )*(?:z[eé]ro))"#)?,
-             number_check!(|number: &NumberValue| !number.suffixed()),
+             integer_check!(|integer: &IntegerValue| !integer.suffixed),
              |a, _, zeros, b| {
-                 let power = zeros.group(0).split_whitespace().count() + b.value().value().to_string().chars().count();
-                 let coeff = 10.0_f64.powf(-1.0 * power as f64);
+                let zeros_string =  std::iter::repeat("0").take(zeros.group(0).split_whitespace().count()).collect::<String>();
+                println!("{:?}", format!("{}.{}{}", a.value().value, zeros_string, b.value().value));
+                 let value: f64 = format!("{}.{}{}", a.value().value, zeros_string, b.value().value).parse()?;
                  Ok(FloatValue {
-                     value: b.value().value() * coeff + a.value().value(),
+                     value,
                      ..FloatValue::default()
                  })
              });