Removed unneccessary _iter suffixes from the various APIs

[Kimundi]

This PR removes almost all _iter suffixes in various APIs of the codebase that return Iterators, as discussed in #9440.

As a summarize for the intend behind this PR:

• Iterators are the recommended way to provide a potentially lazy list of values, no need to name them painfully verbose. If anything, functions that return a specific container type should have more verbose names.
• We have a static type system, so no need to encode the return value of a constructor function into its name.

Following is a possibly incomplete list of all renamings I performed in the codebase. For a few of them I'm a bit unsure whether the new name still properly expresses their functionality, so feedback would be welcome:

&str : word_iter()             -> words()
       line_iter()             -> lines()
       any_line_iter()         -> lines_any()
       iter()                  -> chars()
       char_offset_iter()      -> char_indices()
       byte_iter()             -> bytes()
       split_iter()            -> split()
       splitn_iter()           -> splitn()
       split_str_iter()        -> split_str()
       split_terminator_iter() -> split_terminator()
       matches_index_iter()    -> match_indices()
       nfd_iter()              -> nfd_chars()
       nfkd_iter()             -> nfkd_chars()

&[T] : split_iter()        -> split()
       splitn_iter()       -> splitn()
       window_iter()       -> windows()
       chunk_iter()        -> chunks()
       permutations_iter() -> permutations()

extra:bitv::Bitv :  rev_liter()    -> rev_iter()
                    common_iter()  -> commons()
                    outlier_iter() -> outliers()

extra::treemap::{...} : lower_bound_iter() -> lower_bound()
                        upper_bound_iter() -> upper_bound()

std::trie::{...} : bound_iter()       -> bound()
                   lower_bound_iter() -> lower_bound()
                   upper_bound_iter() -> upper_bound()

rustpkg::package_id::{...} : prefixes_iter() -> prefixes()

std::hashmap::{...} : difference_iter()           -> difference()
                      symmetric_difference_iter() -> symmetric_difference()
                      intersection_iter()         -> intersection()
                      union_iter()                -> union()

std::path::{posix, windows} : component_iter()     -> components()
                              str_component_iter() -> str_components()

... not showing all identical renamings for reverse versions

---

I'm also planning a few more changes, like removing all unnecessary _rev constructors (#9391), or reducing the split variants on &str to a more versatile and concise system.

[emberian]

👍 awesome

[alexcrichton]

I am very much in favor of this. If I hear no objections in the next day or so, I will r+ this.

[huonw]

(@Kimundi if you do receive feedback and change things, could you keep the PR description up-to-date, so it acts as a cheatsheet/FAQ for people updating to the new names.)

[Kimundi]

Rebased and updated, and also updated the table to list the string normalization iterators, because I somehow missed them.

[brendanzab]

Woops, I might have introduced another _iter in #10688. Any idea for a better name for recv_iter?

[emberian]

.messages() ?