better message for raising integer to a negative integer

[macd]

Is throwing a DomainError intended behavior when raising to a negative integer?

julia> VERSION
v"0.3.0-rc1+86"

julia> 2^-1
ERROR: DomainError
in power_by_squaring at intfuncs.jl:60
in ^ at intfuncs.jl:84

julia> 2^-1.
0.5

[ViralBShah]

Yes, when both inputs are integers. You could do 2.0^-1 as well.

[ivarne]

It even has it's own FAQ entry

[macd]

Sorry, missed that one. was looking the manual. I do think this will
trip up many folks coming from MATLAB.

On Wed, Jul 23, 2014 at 4:00 AM, Ivar Nesje [email protected]
wrote:

It even has it's own FAQ entry
http://docs.julialang.org/en/latest/manual/faq/#why-does-julia-give-a-domainerror-for-perfectly-sensible-operations

—
Reply to this email directly or view it on GitHub
#7702 (comment).

[timholy]

Yep. If you have any good suggestions (that haven't been discussed before; please search old issues first), please let us know.

[ivarne]

It is not mentioned in http://docs.julialang.org/en/latest/manual/noteworthy-differences/#noteworthy-differences-from-matlab. It probably should, because this is really a noteworthy difference.

[ivarne]

We can also probably add some text to the printing of DomainError for power_by_squaring as we did in #4249

[StefanKarpinski]

Note that raising a floating-point value to a negative integer is fine, so technically this is not different than Matlab – it's just that in Matlab 2 is a floating-point number, not an integer.

[mbauman]

In Matlab:

>> int64(2)^-1
Error using  .^ 
Integers can only be raised to positive integral powers.

The notable thing (which isn't mentioned) is that 2 isn't a Float64. The domain error could then be mentioned as a consequence.

[ViralBShah]

We should probably have a more informative error.

[ivarne]

Reopening for a better error message. I'll assign myself.

[jakebolewski]

Closed by #8784.