Removes littering `freeze` calls with Ruby 2.3's magic comment.

Bring Boolean Attributes list for AV Tags helper upto speed with current spec

Expand list of void elements to match spec

Freeze string for default content for tag_strings, which later goes on to build a new one

This is based on https://github.com/kangax/html-minifier/blob/6b2d4536d82819143b468b41a89c700b6c61631f/src/htmlminifier.js#L197 and
spec from https://www.w3.org/TR/html51/single-page.html.

Couple of other changes to tests due to support update:
- autobuffer has been dropped in favour of preload attribute, ref: https://msdn.microsoft.com/en-us/library/ff974743(v=vs.85).aspx
- pubdate attribute has been dropped from spec, ref:  https://www.w3.org/html/wg/tracker/issues/185

Freeze string for default content for tag_strings, which later goes on to create new content. This should also be inline with content being passed should not be mutable

Expand list of void elements to match spec from https://html.spec.whatwg.org/multipage/syntax.html#void-elements

Modifies mime-registration test not to interfere with real mime types

Fix default key length on cipher

We default to using aes-256-cbc as our verification/signing cipher. It can accept key lengths of 128, 192 or 256-bit, whereas currently we were providing twice the acceptable value.
ruby < 2.4 allowed accepting these values, as extra key bits were ignored. Since https://github.com/ruby/ruby/commit/ce635262f53b760284d56bb1027baebaaec175d1 this now has a strict checking on key length.

Default to key length 32 bytes, to match the compatible length for  aes-256-cbc

Fixes #25185

Document `simple_format` handling of `\r`

`simple_format` handles `\r\n` and `\r` but previously did not document
this behavior. The handling is done with a `gsub` in `split_paragraphs`.

[ci skip]

New syntax for tag helpers i.e. tag.br instead of tag('br') #25195

[ci skip] Clarify DDL term in ActiveRecord::Transactions

Remove conflict line [ci skip]

Conflicts:
	actioncable/README.md

Fix typo in exception class name

Define `Pathname#as_json`

remove unecessary comments [ci skip]

These comments were removed in 8b69f1ee.

Remove unnecessary `assert_valid_default`

client test have not using Pathname and hashwithindifferent access

Fix a tiny typo and vertical-align some return results in the
ActiveModel::Errors documentation.

When the Pathname object is converted as JSON,
it should be a string that means itself.

Expected:
```
>> Pathname.new('/path/to/somewhere.txt').as_json
"/path/to/somewhere.txt"
```

Actual:
```
>> Pathname.new('/path/to/somewhere.txt').as_json
{"path"=>"/path/to/somewhere.txt"}
```

This was added at c7c3f73f but it never raised because MySQL cannot
create text/blob columns with a default value.

Define `URI::Generic#as_json`

Then just yield the location for the place where we need some extra processing.

`if !var.nil?` is the same as just `if var`

make `as` option work with get parameters

When the URI object is converted as JSON,
it is expected that it is a string that means its URI.

Expected:
```
>> URI.parse('http://example.com').as_json
"http://example.com"
```

Actual:
```
>> URI.parse('http://example.com').as_json
{"scheme"=>"http",
 "user"=>nil,
 "password"=>nil,
 "host"=>"example.com",
 "port"=>80,
 "path"=>"",
 "query"=>nil,
 "opaque"=>nil,
 "fragment"=>nil,
 "parser"=>
  {"regexp"=>
    {"SCHEME"=>"(?-mix:\\A[A-Za-z][A-Za-z0-9+\\-.]*\\z)",
     "USERINFO"=>"(?-mix:\\A(?:%\\h\\h|[!$&-.0-;=A-Z_a-z~])*\\z)",
     "HOST"=>
      "(?-mix:\\A(?:(?<IP-literal>\\[(?:(?<IPv6address>(?:\\h{1,4}:){6}(?<ls32>\\h{1,4}:\\h{1,4}|(?<IPv4address>(?<dec-octet>[1-9]\\d|1\\d{2}|2[0-4]\\d|25[0-5]|\\d)\\.\\g<dec-octet>\\.\\g<dec-octet>\\.\\g<dec-octet>))|::(?:\\h{1,4}:){5}\\g<ls32>|\\h{,4}::(?:\\h{1,4}:){4}\\g<ls32>|(?:(?:\\h{1,4}:)?\\h{1,4})?::(?:\\h{1,4}:){3}\\g<ls32>|(?:(?:\\h{1,4}:){,2}\\h{1,4})?::(?:\\h{1,4}:){2}\\g<ls32>|(?:(?:\\h{1,4}:){,3}\\h{1,4})?::\\h{1,4}:\\g<ls32>|(?:(?:\\h{1,4}:){,4}\\h{1,4})?::\\g<ls32>|(?:(?:\\h{1,4}:){,5}\\h{1,4})?::\\h{1,4}|(?:(?:\\h{1,4}:){,6}\\h{1,4})?::)|(?<IPvFuture>v\\h+\\.[!$&-.0-;=A-Z_a-z~]+))\\])|\\g<IPv4address>|(?<reg-name>(?:%\\h\\h|[!$&-.0-9;=A-Z_a-z~])*))\\z)",
     "ABS_PATH"=>
      "(?-mix:\\A\\/(?:%\\h\\h|[!$&-.0-;=@-Z_a-z~])*(?:\\/(?:%\\h\\h|[!$&-.0-;=@-Z_a-z~])*)*\\z)",
     "REL_PATH"=>
      "(?-mix:\\A(?:%\\h\\h|[!$&-.0-;=@-Z_a-z~])+(?:\\/(?:%\\h\\h|[!$&-.0-;=@-Z_a-z~])*)*\\z)",
     "QUERY"=>"(?-mix:\\A(?:%\\h\\h|[!$&-.0-;=@-Z_a-z~\\/?])*\\z)",
     "FRAGMENT"=>"(?-mix:\\A(?:%\\h\\h|[!$&-.0-;=@-Z_a-z~\\/?])*\\z)",
     "OPAQUE"=>"(?-mix:\\A(?:[^\\/].*)?\\z)",
     "PORT"=>
      "(?-mix:\\A[\\x09\\x0a\\x0c\\x0d ]*\\d*[\\x09\\x0a\\x0c\\x0d ]*\\z)"}}}
```

Use correct timezone when parsing date in json

Currently, if path is a relative path, add format without the discrimination of the query.
Therefore, if there is a query, format at end of the query would been added,
format was not be specified correctly.

This fix add format to end of path rather than query.