Makes most strings tests pass

[gampleman]

@ahankinson I took a stab at this.

I think you can still get failures with high enough --fuzz values, but the build should now pass at least most of the time, which I think is a significant milestone.

There are a few fixes/changes to behaviour that I think are generally sensible.

These are:

-- Before
String.Extra.softBreak 1 "    " --> ["    "]
String.Extra.softBreak 3 "foo    bar" --> ["foo    ", "bar"]

-- After
String.Extra.softBreak 1 "    " --> [" ", " ", " ", " "]
String.Extra.softBreak 3 "foo    bar" --> ["foo", "   ", " bar"]

-- Before
String.Extra.toTitleCase "č" --> "č"

-- After
String.Extra.toTitleCase "č" --> "Č"

-- Before
String.Extra.clean "foo\tbar" --> "foo\tbar"

-- After
String.Extra.clean "foo\tbar" --> "foo bar"

I also deleted a test that was just wrong 🤷.

[ahankinson]

Looks great to me! I did manage to get it to fail with a high fuzz rate:

Running 574 tests. To reproduce these results, run: elm-test --fuzz 100000 --seed 316688162955213

↓ String.HumanizeTest
↓ humanize
✗ It removes a trailing `_id` & replaces underscores and dashes with a single whitespace

Given "__id"

    "id"
    ╷
    │ Expect.equal
    ╵
    ""

I'm not sure if that's something you think is worth fixing or not? (c.f., your slack messages about "_id")