hustf: curious... why are if-statements non-Julian?  How would you do it?

I think a macro-version would be best, and fairly easy to write

On Fri, Jan 8, 2016 at 12:55 PM, Josh Langsfeld <[email protected]> wrote:

> Here's an idea for a more general and scalable approach rather than
> checking each argument's type with an if-statement:
>
> function f(args...)
>   argdict = Dict(zip(map(typeof, args), args))
>   f(argdict[A], argdict[B]) #etc...
> end
>
> Note that it won't work if any of your A,B,etc are abstract types because
> argdict will always end up with concrete types for keys. Additional error
> checking possible at the discretion of the developer.
>
> On Friday, January 8, 2016 at 12:17:47 PM UTC-5, hustf wrote:
>>
>> The "if"-sentences look kind of non-Julian, and it's very easy to
>> introduce bugs that way in my humble experience. Makes for very long code,
>> too, generally speaking. Would writing a macro for establishing the
>> alternative sequences be shorter codewise and quicker to run? In a way,
>> that is what is done implicitly with keyword arguments? I believe e.g.
>> @vectorize_2 arg can be used for a roughly similiar purpose.
>>
>

Reply via email to