Prevents two string allocations per method call for common REST verbs
plus a ~1.5x speedup for :get in particular

```ruby
begin
  require "bundler/inline"
rescue LoadError => e
  $stderr.puts "Bundler version 1.10 or later is required. Please update
                your Bundler"
  raise e
end

gemfile(true) do
  source "https://rubygems.org"

  gem "benchmark-ips"
  gem "rails"
end

def allocate_count
  GC.disable
  before = ObjectSpace.count_objects
  yield
  after = ObjectSpace.count_objects
  after.each { |k,v| after[k] = v - before[k] }
  after[:T_HASH] -= 1 # probe effect - we created the before hash.
  GC.enable
  result = after.reject { |k,v| v == 0 }
  GC.start
  result
end

@html_options = {}

def master_version
  if @method && @method.to_s.downcase != "get" && @html_options["rel".freeze] !~ /nofollow/
    @html_options["rel".freeze] = "#{@html_options["rel".freeze]} nofollow".lstrip
  end
  @html_options["data-method".freeze] = @method
end

def fast_version
  if method_not_get_method?(@method) && @html_options["rel".freeze] !~ /nofollow/
    @html_options["rel".freeze] = "#{@html_options["rel".freeze]} nofollow".lstrip
  end
  @html_options["data-method".freeze] = @method
end

STRINGIFIED_COMMON_METHODS = {
  get:    'get',
  delete: 'delete',
  patch:  'patch',
  post:   'post',
  put:    'put',
}.freeze

def method_not_get_method?(method)
  return false unless method
  (STRINGIFIED_COMMON_METHODS[method] || method.to_s.downcase) != 'get'
end

puts 'get'
@method = :get

puts "master_version"
puts allocate_count { 1000.times { master_version } }
puts "fast_version"
puts allocate_count { 1000.times { fast_version } }

Benchmark.ips do |x|
  x.report("master_version") { master_version }
  x.report("fast_version")     { fast_version }
  x.compare!
end

puts 'delete'
@method = :delete

puts "master_version"
puts allocate_count { 1000.times { master_version } }
puts "fast_version"
puts allocate_count { 1000.times { fast_version } }

Benchmark.ips do |x|
  x.report("master_version") { master_version }
  x.report("fast_version")     { fast_version }
  x.compare!
end
```

```
get
master_version
{:FREE=>-1819, :T_STRING=>2052}
fast_version
{:FREE=>-1}
Warming up --------------------------------------
      master_version   140.839k i/100ms
        fast_version   175.639k i/100ms
Calculating -------------------------------------
      master_version      2.683M (± 7.1%) i/s -     13.380M in   5.013447s
        fast_version      3.988M (±10.1%) i/s -     19.847M in   5.039580s

Comparison:
        fast_version:  3988340.3 i/s
      master_version:  2683336.2 i/s - 1.49x  slower

delete
master_version
{:FREE=>-5003, :T_STRING=>3003, :T_MATCH=>999, :T_IMEMO=>1000}
fast_version
{:FREE=>-3002, :T_STRING=>1001, :T_MATCH=>1000, :T_IMEMO=>1000}
Warming up --------------------------------------
      master_version    47.221k i/100ms
        fast_version    44.153k i/100ms
Calculating -------------------------------------
      master_version    597.881k (±11.4%) i/s -      2.975M in   5.047200s
        fast_version    686.014k (±11.6%) i/s -      3.400M in   5.036564s

Comparison:
        fast_version:   686014.5 i/s
      master_version:   597881.4 i/s - same-ish: difference falls within error
```

This allows us to prevent an extra string allocation when there is a rel
argument and performs better/within error of the key check for other
scenarios such as passing in rel: nil

```ruby
begin
  require "bundler/inline"
rescue LoadError => e
  $stderr.puts "Bundler version 1.10 or later is required. Please update
                your Bundler"
  raise e
end

gemfile(true) do
  source "https://rubygems.org"

  gem "benchmark-ips"
  gem "rails"
end

def allocate_count
  GC.disable
  before = ObjectSpace.count_objects
  yield
  after = ObjectSpace.count_objects
  after.each { |k,v| after[k] = v - before[k] }
  after[:T_HASH] -= 1 # probe effect - we created the before hash.
  GC.enable
  result = after.reject { |k,v| v == 0 }
  GC.start
  result
end

@hash = {}

def master_version
  "#{@hash["rel"]} nofollow".lstrip
end

def key_version
  if @hash.key?("rel")
    "#{@hash["rel"]} nofollow".lstrip
  else
    "nofollow"
  end
end

def present_version
  if @hash["rel"].present?
    "#{@hash["rel"]} nofollow"
  else
    "nofollow".freeze
  end
end

def nil_version
  if @hash["rel"].nil?
    "nofollow".freeze
  else
    "#{@hash["rel"]} nofollow"
  end
end

def blank_version
  if @hash["rel"].blank?
    "nofollow".freeze
  else
    "#{@hash["rel"]} nofollow"
  end
end

def test
  puts "master_version"
  puts allocate_count { 1000.times { master_version } }
  puts "key_version"
  puts allocate_count { 1000.times { key_version } }
  puts "present_version"
  puts allocate_count { 1000.times { present_version } }
  puts "nil_version"
  puts allocate_count { 1000.times { nil_version } }
  puts "blank_version"
  puts allocate_count { 1000.times { blank_version } }

  Benchmark.ips do |x|
    x.report("master_version")  { master_version }
    x.report("key_version")     { key_version }
    x.report("present_version") { present_version }
    x.report("nil_version")     { nil_version }
    x.report("blank_version")     { blank_version }
    x.compare!
  end
end

puts 'no rel key'

test

puts 'rel key with real stuff'

@hash['rel'] = 'hi'.freeze

test

puts 'rel key with nil'

@hash['rel'] = nil

test

puts 'rel key with ""'

@hash['rel'] = ""

test
```

```
no rel key
master_version
{:FREE=>-2818, :T_STRING=>3052}
key_version
{:FREE=>-1}
present_version
{:FREE=>-1}
nil_version
{:FREE=>-1}
blank_version
{:FREE=>-1}
Warming up --------------------------------------
      master_version   124.677k i/100ms
         key_version   227.992k i/100ms
     present_version   208.068k i/100ms
         nil_version   235.272k i/100ms
       blank_version   176.274k i/100ms
Calculating -------------------------------------
      master_version      1.968M (±10.8%) i/s -      9.725M in   5.010763s
         key_version      7.734M (±11.2%) i/s -     38.075M in   5.001613s
     present_version      5.688M (±11.4%) i/s -     28.089M in   5.019560s
         nil_version      6.965M (±10.2%) i/s -     34.585M in   5.024806s
       blank_version      6.139M (±18.7%) i/s -     29.085M in   5.010919s

Comparison:
         key_version:  7734058.3 i/s
         nil_version:  6965050.2 i/s - same-ish: difference falls within error
       blank_version:  6138744.3 i/s - same-ish: difference falls within error
     present_version:  5688248.4 i/s - 1.36x  slower
      master_version:  1967932.3 i/s - 3.93x  slower

rel key with real stuff
master_version
{:FREE=>-2001, :T_STRING=>2000}
key_version
{:FREE=>-2001, :T_STRING=>2000}
present_version
{:FREE=>-1001, :T_STRING=>1000}
nil_version
{:FREE=>-1002, :T_STRING=>1000, :T_IMEMO=>1}
blank_version
{:FREE=>-1001, :T_STRING=>1000}
Warming up --------------------------------------
      master_version    93.351k i/100ms
         key_version    89.747k i/100ms
     present_version    91.963k i/100ms
         nil_version   103.370k i/100ms
       blank_version    74.845k i/100ms
Calculating -------------------------------------
      master_version      2.179M (±21.4%) i/s -     10.362M in   5.044668s
         key_version      2.345M (± 9.8%) i/s -     11.667M in   5.030982s
     present_version      1.738M (±14.8%) i/s -      8.553M in   5.056406s
         nil_version      2.485M (±19.1%) i/s -     11.888M in   5.015940s
       blank_version      1.951M (±12.3%) i/s -      9.580M in   5.011932s

Comparison:
         nil_version:  2484704.1 i/s
         key_version:  2344664.8 i/s - same-ish: difference falls within error
      master_version:  2178975.8 i/s - same-ish: difference falls within error
       blank_version:  1950532.0 i/s - same-ish: difference falls within error
     present_version:  1737866.7 i/s - 1.43x  slower

rel key with nil
master_version
{:FREE=>-3001, :T_STRING=>3000}
key_version
{:FREE=>-3001, :T_STRING=>3000}
present_version
{:FREE=>-1}
nil_version
{:FREE=>-1}
blank_version
{:FREE=>-1}
Warming up --------------------------------------
      master_version   112.655k i/100ms
         key_version   105.048k i/100ms
     present_version   136.219k i/100ms
         nil_version   192.026k i/100ms
       blank_version   184.846k i/100ms
Calculating -------------------------------------
      master_version      1.893M (±12.6%) i/s -      9.238M in   5.002621s
         key_version      1.672M (±13.5%) i/s -      8.194M in   5.021197s
     present_version      4.484M (±20.5%) i/s -     21.114M in   5.002982s
         nil_version      5.294M (±18.1%) i/s -     25.155M in   5.020721s
       blank_version      5.588M (± 6.7%) i/s -     27.912M in   5.019305s

Comparison:
       blank_version:  5588489.6 i/s
         nil_version:  5293929.9 i/s - same-ish: difference falls within error
     present_version:  4484493.7 i/s - same-ish: difference falls within error
      master_version:  1892919.0 i/s - 2.95x  slower
         key_version:  1672343.9 i/s - 3.34x  slower

rel key with ""
master_version
{:FREE=>-2001, :T_STRING=>2000}
key_version
{:FREE=>-2001, :T_STRING=>2000}
present_version
{:FREE=>-1}
nil_version
{:FREE=>-1001, :T_STRING=>1000}
blank_version
{:FREE=>-1}
Warming up --------------------------------------
      master_version   140.499k i/100ms
         key_version   124.738k i/100ms
     present_version   186.659k i/100ms
         nil_version   148.063k i/100ms
       blank_version   178.707k i/100ms
Calculating -------------------------------------
      master_version      1.826M (±24.2%) i/s -      8.289M in   5.026603s
         key_version      1.561M (±15.3%) i/s -      7.609M in   5.005662s
     present_version      3.622M (±19.9%) i/s -     17.173M in   5.042217s
         nil_version      2.438M (±11.5%) i/s -     12.141M in   5.053335s
       blank_version      4.911M (±15.5%) i/s -     23.768M in   5.009106s

Comparison:
       blank_version:  4910741.1 i/s
     present_version:  3622183.5 i/s - same-ish: difference falls within error
         nil_version:  2437606.2 i/s - 2.01x  slower
      master_version:  1825652.2 i/s - 2.69x  slower
         key_version:  1560530.5 i/s - 3.15x  slower
```

Do a check if the 'rel' argument is passed in, and simply set it to
'nofollow' if 'rel' was not passed in. This prevents three string
allocations for each call to `link_to` in that scenario. In the scenario
where the 'rel' argument is passed in, performance is around the same as
before as the `key?` check is very fast.

```ruby
begin
  require "bundler/inline"
rescue LoadError => e
  $stderr.puts "Bundler version 1.10 or later is required. Please update
                your Bundler"
  raise e
end

gemfile(true) do
  source "https://rubygems.org"

  gem "benchmark-ips"
  gem "rails"
end

def allocate_count
  GC.disable
  before = ObjectSpace.count_objects
  yield
  after = ObjectSpace.count_objects
  after.each { |k,v| after[k] = v - before[k] }
  after[:T_HASH] -= 1 # probe effect - we created the before hash.
  GC.enable
  result = after.reject { |k,v| v == 0 }
  GC.start
  result
end

@hash = {}

def master_version
  "#{@hash['rel'.freeze]} nofollow".lstrip
end

def fast_version
  if @hash.key?('rel'.freeze)
    "#{@hash["rel"]} nofollow".lstrip
  else
    "nofollow".freeze
  end
end

puts 'no rel key'

puts "master_version"
puts allocate_count { 1000.times { master_version } }
puts "fast_version"
puts allocate_count { 1000.times { fast_version } }

Benchmark.ips do |x|
  x.report("master_version") { master_version }
  x.report("fast_version")     { fast_version }
  x.compare!
end

puts 'rel key'

@hash['rel'] = 'hi'.freeze

puts "master_version"
puts allocate_count { 1000.times { master_version } }
puts "fast_version"
puts allocate_count { 1000.times { fast_version } }

Benchmark.ips do |x|
  x.report("master_version") { master_version }
  x.report("fast_version")     { fast_version }
  x.compare!
end
```

```
no rel key
master_version
{:FREE=>-2791, :T_STRING=>3052}
fast_version
{:FREE=>-1}
Warming up --------------------------------------
      master_version    80.324k i/100ms
        fast_version   200.262k i/100ms
Calculating -------------------------------------
      master_version      2.049M (±11.9%) i/s -     10.121M in   5.025613s
        fast_version      6.645M (±21.3%) i/s -     29.439M in   5.007488s

Comparison:
        fast_version:  6644506.3 i/s
      master_version:  2048833.0 i/s - 3.24x  slower

rel key
master_version
{:FREE=>-2001, :T_STRING=>2000}
fast_version
{:FREE=>-2001, :T_STRING=>2000}
Warming up --------------------------------------
      master_version   155.673k i/100ms
        fast_version   106.515k i/100ms
Calculating -------------------------------------
      master_version      2.652M (±20.4%) i/s -     12.610M in   5.036494s
        fast_version      2.237M (±16.8%) i/s -     10.865M in   5.035366s

Comparison:
      master_version:  2651702.2 i/s
        fast_version:  2237470.6 i/s - same-ish: difference falls within error
```

This basically reverts c4d1a4ef

This reverts commit 3420a145, reversing
changes made to afb66a5a.

- `check_parameters` kwargs was added to the `current_page?` method, the implementation was assuming only hashes responds to `delete`. This was causing issues when `current_page?` was called with a Active Model object
- ref https://github.com/rails/rails/pull/27549
- Fixes #28846

Since this protection is now in Parameters we can use it instead of
reimplementing again.

(I personally prefer writing one string in one line no matter how long it is, though)

Example:

For "http://www.example.com/shop/checkout?order=desc&page=1"

current_page?('http://www.example.com/shop/checkout')
=> true

current_page?(
  'http://www.example.com/shop/checkout',
  check_parameters: true
)
=> false

Removed references to `button` in the `link_to` `disable_with` api
description.

Regexp#match? should be considered to be part of the Ruby core library. We are
emulating it for < 2.4, but not having to require the extension is part of the
illusion of the emulation.

Before, an error would be raised saying that the method `to_param` was
undefined on the instance of `ActionController::Parameters`. Now, we are
checking to see if the `params` object being passed to `button_to`
responds to the `permitted?` method, and if so, we will call `to_h` on it. If it
does not respond to `permitted?`, then the `params` will remain
unchanged.

[Jon Moss, Rafael Mendonça França]

Style/SpaceBeforeBlockBraces
Style/SpaceInsideBlockBraces
Style/SpaceInsideHashLiteralBraces

Fix all violations in the repository.

The current code base is not uniform. After some discussion,
we have chosen to go with double quotes by default.

Where appropriate, prefer the more concise Regexp#match?,
String#include?, String#start_with?, or String#end_with?

reduce string allocation.
- Use freezed empty string instead of create one every time we need to return it

When `button_to 'Botton', url` form was being used the per form token
was not correct because the method that is was being used to generate it
was an empty string.

We only need action and method so pass them explicitly instead of
merging the hash with HTML options.

Fixes #23524

In e6e0579d the `params` option was added to the `button_to` helper. However, the patch doesn't support nested hashes so `{a: {b: 'c'}}` for example gets turned into a hidden form input with the name 'a' and the value being the string representation of the `{b: 'c'}` nested hash.

Since Rails supports nested hashes everywhere else (and even in the URL params of link_to and button_to), I believe this to be a bug/unfinished feature.

`link_to :back` creates a link to whatever was
passed in via the referer header. If an attacker
can alter the referer header, that would create
a cross-site scripting vulnerability on every
page that uses `link_to :back`

This commit restricts the back URL to valid
non-javascript URLs.

https://github.com/rails/rails/issues/14444

`hfvalue` parts should always be percent encoded, so lets do that!

Revert "use path escaping for email addresses"

This reverts commit 21ffef38.

Due to e25fdad2, we are correctly using
path escaping for email addresses.  This commit fixes the tests to
expect path escaping.

content_tag's first argument is will generate a string with an html tag so `:a` will generate: `<a></a>`. When this happens, the symbol is implicitly `to_s`-d so a new string is allocated. We can get around that by using a frozen string instead which

This change buys us 74,236 bytes of memory and 1,855 fewer objects per request.

No idea why on earth this hash key isn't already optimized by MRI, but it isn't. 💩

This change buys us 74,077 bytes of memory and 1,852 fewer objects per request.

This reverts commit d459b001, reversing
changes made to 4d4950fa.

😓

The block is not supposed to be passed to `link_to`. It's used for a
customized behavior of the `condtion = false` case. The docs
illustrate that like so:

```
<%=
   link_to_if(@current_user.nil?, "Login", { controller: "sessions", action: "new" }) do
     link_to(@current_user.login, { controller: "accounts", action: "show", id: @current_user })
   end
%>
```

add block to link_to_if when condition is true