Right now printn is used widely throughout the test suite, but its failure to distinguish between +inf and +/- inf impedes investigations of failing tests.
The behavior of printn itself has been discussed already, so it seems like modifying it is not a no-brainer. In that case, would it make sense to switch the preferred print function in the test suite to a version that can distinguish between a finite interval vs. the point at positive infinity vs. the point at negative infinity vs. the whole extended real line vs. an indeterminate value, as outlined in http://arblib.org/arb.html?
Right now
printn
is used widely throughout the test suite, but its failure to distinguish between+inf
and+/- inf
impedes investigations of failing tests.The behavior of
printn
itself has been discussed already, so it seems like modifying it is not a no-brainer. In that case, would it make sense to switch the preferred print function in the test suite to a version that can distinguish between a finite interval vs. the point at positive infinity vs. the point at negative infinity vs. the whole extended real line vs. an indeterminate value, as outlined in http://arblib.org/arb.html?