@rdyne, as I mentioned earlier, the option with the internal IO was just something I meant for you to consider given what you indicated in the original post. Which is that you somehow arrive at a floating-point value programmatically - 0.123456789 as you put it - and you want to represent it as-is in a different (higher) precision for whatever reason.
I did not mean it to be “clever” and please know compared to approaches following via libraries in other languages (e.g., C strtod) the use of internal IO with Fortran can be slower - see this thread.
And I was not intending with my suggestion involving internal IO to get into any messaging to you as to whether it is prudent or not from a numerical analysis involving floating-point calculations, rather leave that up to you, particularly considering you will have best knowledge about your program.