Why does wcalc output rounding symbol (~=) for integer operations (subtraction)?

I have noted, that man wcalc says:

--ints
           Toggles whether long integers will be abbreviated or not.  This  conflicts  with  engineering
           notation for large numbers, but not for decimals.

...

Rounding in the command-line version is denoted by a  tilde  before  the  equals  sign  (~=).

So, I try this:

$ wcalc --version
wcalc 2.5

$ wcalc 1588890124670286510-1588890124626020545
~= 4.4266e+07

Ok, so shown in scientific notation, as expected; then I try:

$ wcalc --ints 1588890124670286510-1588890124626020545
~= 44265965

So, it shows as full integer - again, as expected - but for some reason, it says ~=, meaning the result is "rounded". This kind of confuses me, because shouldn't integer operations (subtraction, at least), be exact? Why is the rounding symbol there, then - what is it implying?