Remove `String#casecmp` part

The `String#casecmp` method does not work with Unicode (even in Ruby 2.4.1), this is written in the [documentation](http://ruby-doc.org/core-2.4.1/String.html#method-i-casecmp).

Example:

```ruby
'Привет'.casecmp('привет') # => -1
```

There is a method `String#casecmp?`, which [works with Unicode](http://ruby-doc.org/core-2.4.1/String.html#method-i-casecmp-3F).

Example:

```ruby
'Привет'.casecmp?('привет') # => true
```

But `String.casecmp?` is slower:

```
Warming up --------------------------------------
String#downcase + == 233.440k i/100ms
String#casecmp 274.247k i/100ms
String#casecmp? 219.906k i/100ms
Calculating -------------------------------------
String#downcase + == 5.746M (± 1.7%) i/s - 28.947M in 5.039252s
String#casecmp 6.942M (± 1.8%) i/s - 34.829M in 5.019073s
String#casecmp? 4.517M (± 2.6%) i/s - 22.650M in 5.017864s

Comparison:
String#casecmp: 6941676.9 i/s
String#downcase + ==: 5745893.8 i/s - 1.21x slower
String#casecmp?: 4517314.3 i/s - 1.54x slower
```

Code

```ruby
require 'benchmark/ips'

SLUG = 'ABCD'

def slow
SLUG.downcase == 'abcd'
end

def fast
SLUG.casecmp('abcd') == 0
end

def another
SLUG.casecmp?('abcd')
end

Benchmark.ips do |x|
x.report('String#downcase + ==') { slow }
x.report('String#casecmp') { fast }
x.report('String#casecmp?') { another }
x.compare!
end
```

So, `String#downcase + ==` is good compromise.