Address floating point precision issue when reading functions

[finsberg]

Add better error message and more docs for adressing float precision issue when reading functions.

• Update the error message emitted when you hit this error.
• Add a test that make sure the new error message is diplayed
• Add a new page to the docs about debugging where this problem and a solution is presented

Address https://github.com/jorgensd/adios4dolfinx/issues/186