Possible bug in DoubleParser.cs

In Microsoft.ML.Core\Utilities\DoubleParser.cs near lines 143 and 195

```csharp
int ichEnd;
if (!DoubleParser.TryParse(span.Slice(ich, span.Length - ich), out value, out ichEnd, flags))
{
value = default(Double);
return Result.Error;
}

// Make sure everything was consumed.
while (ichEnd < span.Length)
{
if (!char.IsWhiteSpace(span[ichEnd]))
return Result.Extra;
ichEnd++;
}
```
the ichEnd is indexed on the sliced span.
If ich is not 0 there will be an offset when for "ichEnd < span.Length" and "span[ichEnd]"

Example : for an input like " 1.234 " the method will return Result.Extra

Pössible solutions:
* ichEnd += ich;
* modify the tryparse to accept an offset and use the original span