StrSlice should have split_at

[Manishearth]

ImmutableSlice has a split_at method. This would probably be useful for StrSlice as well.

(It might be a good idea to share some of the other methods between StrSlice and ImmutableSlice too)

[Diggsey]

It would have to make sure not to split the string across a multi-byte codepoint.

[Manishearth]

In the context of Rust strings n is de facto the Unicode-safe index, not the byte index.

[Diggsey]

That would make it inconsistent with the other methods on "str", which use byte indices unless explicitly specified otherwise by the presence of "chars" in the method name.

It would also introduce a hidden complexity cost, making the string version of "split_at" be O(n) instead of O(1) - the obvious use-case of splitting a string at multiple places then becomes O(n^2), which is unnacceptable for many applications.

[Manishearth]

split_at_char then.

[djmally]

I'll take a look into this, if no one else is already.

[Manishearth]

Sounds good. @gankro does this need an RfC?

[Gankra]

Uh, what are the traits called now?

[djmally]

That was going to be my next question, I think StrSlice is no longer a trait as of 1.0alpha

[djmally]

Does this still need to be done (i.e. has the trait been renamed), or should this be closed?

[steveklabnik]

http://doc.rust-lang.org/nightly/std/primitive.slice.html#method.split_at <-

[steveklabnik]

(that trait no longer exists, and this is an inherent impl now)

[bluss]

.split_at() on &str would be very useful. To be consistent it would use a byte offset though. I'd like this.

[steveklabnik]

oh whoops! slice != str slice...