Error on large matrix powers

[mikmoore]

julia> VERSION
v"1.10.3"

julia> [1 1e-10; 0 1] ^ 1e20
ERROR: InexactError: Int64(1.0e20)

The issue is that there are multiple places where our exponentiation routines attempt to convert (or round, in some places) the exponent to an Int without first checking that such a conversion is possible. For example, in ^, we call LinearAlgebra.integerpow on isinteger powers (including all large float powers) which attempts Integer(p). In LinearAlgebra.schurpow we attempt floor(Integer, p) which fails for the same reasons. There may be additional sites with the unchecked assumption that one can convert to Int, but I noticed those two.