"Technically" out-of-range but actually-ok range indexing "unnecessarily" errors

[nsiccha]

Again a somewhat weird title - but as always it comes with an example. The following errors, but I think it should actually be fine.
However, whether or not this errors is probably a decision that can be made arbitrarily.

transformed data {
    vector[2] tmp = [1., 2.]';
    // This is fine and prints `[]`
    print(tmp[2:1]);
    // This is not fine and errors with `vector[min_max] min indexing: accessing element out of range. index 3 out of range; expecting index to be between 1 and 2`
    print(tmp[3:2]);
}

For example, the equivalent in python apparently never errors (which I find weird):

[1, 2][1:0] # fine
[1, 2][2:1] # fine
[1, 2][1:2] # fine?!
[1, 2][2] # not fine

The equivalent in Julia (with the indices shifted due to Julia's 1-based indexing) does behave how I'd like Stan to behave (pretty please):

[1, 2][2:1] # fine, returns []
[1, 2][3:2] # fine, returns []
[1, 2][2:3] # not fine
[1, 2][3] # not fine

[WardBrian]

I'd like to confirm with @bob-carpenter what we intended, but my recollection is that we have said before that an empty range should be fine (e.g. Julia's behavior). If so, this is an easy fix I believe

[nsiccha]

That would be nice!

[bob-carpenter]

Yes, an empty range should be fine. At least @WardBrian and I couldn't come up with an example that allowing it would break.

[nsiccha]

Cool, thanks!