They are in Julia:

julia> for a in 1
       @show a
       end
a = 1

AHH, BUT WHYYY?

It's considered practical by some people. The question is rather whether this has drawbacks or not. Else there's no reason to remove that.

FWIW, please use the mailing list for questions in the future, and keep GitHub for bug reports.

FWIW, I consider this a mistake, it has bitten me way too many times.

I also find it a bad idea, but github may not be the place to revive the debate

This just reeks of MATLAB to me.

@samuela the supported iteration / indexing interface is to basically treat scalars as zero dimensional arrays.

It is way too easy to miss a colon in a for loop and only do one iteration - I remember vividly how I spent a week debugging why my locality sensitive hash wasn't local - I had for i = x instead of for i = 1:x for the number of rounds boom.

Yeah, I'm not a great fan of this either. :-/

Yes I really think we should experiment with removing iteration of numbers. I think the array indexing redesign might make this possible.

It is way too easy to miss a colon in a for loop and only do one iteration

This is the bug I most often hit these days.

Here was my last 15-minute attempt at making numbers non-iterable: #10331 (comment)

Just to balance the opinions here, I'll mention I use this feature a lot and find it quite handy.

@carlobaldassi what does a typical use look like?