> It's true that the looser scoping rules of langauges like matlab and
python
> can be convenient. On the whole I prefer tighter scoping rules like C++
> though: they make code easier to reason about by making the data
> flow more local.
The situation I met was not as clear as you describe it here.
For some testing purposes I had defined i before defining the function:
julia> i = 10;
julia> function testfun1(x::Vector{Float64})
for i = 1:length(x)
if x[i] == 0.0
break
end
end
return i-1
end;
julia> testfun1([1.0, 2.0, 3.0, 0.0, 1.0])
9
and it took me some time before I realized that the answer came from
outside the function.
That behavior can really lead to very difficult testing situations.
I don't know whether something like this can happen in C/C++.
On Saturday, May 17, 2014 2:37:13 PM UTC+2, Chris Foster wrote:
>
> On Sat, May 17, 2014 at 9:59 PM, Hans W Borchers
> <[email protected]<javascript:>>
> wrote:
> > Yesterday I implemented a function calculating arc length of curves (to
> the
> > last digit) when I came across the following stumbling blocks. Image the
> > following function where I leave a for-loop with a 'break' statement:
> >
> > function testfun1(x::Vector{Float64})
> > for i = 1:length(x)
> > if x[i] == 0.0
> > break
> > end
> > end
> > return i-1
> > end
>
> For this particular example, I'd suggest you just return i-1 directly
> inside the loop:
>
> function testfun1(x::Vector{Float64})
> for i = 1:length(x)
> if x[i] == 0.0
> return i-1
> end
> end
> return length(x)-1 # Or other special case for when 0 isn't found
> end
>
> Using an early return like this is really quite powerful, especially when
> you have a deeply nested set of loops. I use it all the time in C++.
>
> > I understand that the scope of the loop variable is restricted to the
> loop
> > itself. What is the best way to "export" i to the outside? For the
> moment
> > I settled with defining i before the loop.
>
> Yes, if you want i accessible outside a loop, you'll just have to declare
> it in the outer scope. AFAIK you've done the right thing there, though
> initializing i to a default would be more concise and wouldn't fail if the
> input array is empty (unless that's intentional).
>
> > This works, but I must admit it runs against my gut feeling and
> > experience with other scientific programming languages.
>
> It's true that the looser scoping rules of langauges like matlab and
> python
> can be convenient. On the whole I prefer tighter scoping rules like C++
> though: they make code easier to reason about by making the data
> flow more local.
>
> ~Chris
>
