27 Apr
2023
27 Apr
'23
2:31 p.m.
On Thu Apr 27 13:22:42 2023 +0000, Giovanni Mascellani wrote:
This means that expressions of type other than `float` are casted to `float`. In principle that should be correct in the end, because AFAIU the sign should always be mapped correctly, but it makes me a bit nervous. Could the type of zero be set to coincide with the type of the argument?
Pushed an update to do this, since it seems like if anything it'd be an optimization - as part of the update I've also added int versions of all the tests, to make sure it's doing the right thing for float and non-float types.