Suggest replacing `split(...).collect().rev()` with `rsplit()`

What it does

A user working with a string might forget that str::rsplit exists, and design an API like this:

     /// Returns this DNS name's labels, in reversed order
     /// (from top-level domain to most-specific subdomain).
     fn rlabels(&self) -> impl Iterator<Item = &'_ str> {
        self.as_str()
            .split('.')
            .collect::<Vec<_>>()
            .into_iter()
            .rev()
     }

which could be rewritten as:

    /// Returns this DNS name's labels, in reversed order
    /// (from top-level domain to most-specific subdomain).
    fn rlabels(&self) -> impl Iterator<Item = &'_ str> {
        self.as_str().rsplit('.')
    }

Clippy does not currently detect this.

Advantage

• Simpler (fewer functions invoked)
• Fewer internal iterator transforms
• Fewer allocations (no round-trip through a Vec<_> or other owned container type)
• Reuses an intended standard API instead of reimplementing it

Drawbacks

None that I can think of.

Example

mystr.split('.').collect::<Vec<_>>.into_iter().rev()

Could be written as:

mystr.rsplit('.')

Start from the issue's split(...).collect::<Vec<_>>().into_iter().rev() example and locate Clippy's existing iterator-chain lint entry points and tests. Define the supported equivalent patterns and verify that the lint suggests rsplit() only where the transformation preserves behavior, with tests covering the accepted and rejected cases.