Something like all_same?

[phimuemue]

Do you think it would be helpful to have something that determines whether all elements of an iterator are the same?

[bluss]

Hm first approach, is there a neat way to do that already?

There is:

iter.tuple_windows().all(|(x, y)| x == y)
// but this requires cloning elements (irrelevant if they are integers or references)

There is (weird one):

all_equal == iter.dedup().nth(1).is_none() // only zero or one unique elements

I haven't had the need for this operation, I can see that current itertools doesn't cover it very well. .dedup() is pretty close in fact, but it's not intuitive.

[phimuemue]

Thank you for your reply. Actually, I found out that in my concrete case, there is another - more natural - way avoiding all_same, so it is not very urgent for me.

However, I still think that it might be useful, especially within assert! (imagine e.g. you're using group_by and need a representative from each group).

I do not know how the dedup actually works, so I can not tell whether dedup().nth(1) solution does more work than necessary.

[bluss]

dedup is lazy and only looks at adjacent elements so it should be close to doing the minimal work required (this is without me cross checking with the implementation).

[bluss]

This method needs a better name. With .all_equal() I think it would be a go. I'd consider implementing it using dedup.

[phimuemue]

I think you are correct that a dedup-based solution is convenient.

However, I saw that dedup internally remembers the last seen element (it works via CoalesceCore which has Ok(joined) => last = joined, i.e. assigning last in each internal step). I think remembering last might be more work than necessary, or am I mistaken?

[bluss]

I think it needs to remember last. Otherwise, why not prove it with an implementation 😄

[phimuemue]

Should we close this?

[bluss]

Sure. Fixed in #196