methods(TT)
call(::Type{TT}, x::Float64, y::Float64) at In[22]:2
call(::Type{TT}, x, y) at In[22]:2
call{T}(::Type{T}, arg) at essentials.jl:56
call{T}(::Type{T}, args...) at essentials.jl:57
So I guess the rule is applied in call(::Type{TT}, x, y)
What does it do? Does it try to convert(Float64, x) & convert(Float64, y)?
I.e does it always try to convert each argument to the type of the field
defined in the type?
On 5 April 2016 at 15:11, Yichao Yu <yyc1...@gmail.com> wrote:
> On Tue, Apr 5, 2016 at 10:09 AM, FANG Colin <colinf...@gmail.com> wrote:
> > Sorry if this has been discussed somewhere as I am unable to find the
> > relative post.
> >
> > immutable TT
> > x::Float64
> > y::Float64
> > end
> >
> > function tt(x::Float64, y::Float64)
> > x + y
> >
> > end
> > tt(1,2) # doesn't work
> > TT(1,2) # works
>
> There are automatic constructors defined with conversion not promotion.
> See `methods(TT)`
>
> >
> > What rule applies here for TT(1,2)?
>
