Closed toojays closed 5 months ago
I should have mentioned, this is with the main branch as at cae6fad80.
Can we use the same mechanism to print the maximum as is used to print the input value?
Yes, I don't see why not. Could you propose a patch and a test?
Consider the following schema:
An interesting case is where a 32-bit floating point representation of PI is converted to a double and serialised:
The error message is confusing. The maximum has been truncated to six decimal places, which makes it look larger than the value it is complaining about. Can we use the same mechanism to print the maximum as is used to print the input value?