feat: improve tag support for parity with formatjs

[ramsey]

Description

It turns out that FormatJS doesn't parse real HTML/XML tags in messages. Instead, they use custom tags that can be associated with a callback to replace the tag with a computed value.

For example:

$formatphp->formatMessage([
    'id' => 'myMessage',
    'defaultMessage' => 'Hello, <profileLink>{name}</profileLink>!',
], [
    'name' => 'Ben',
    'profileLink' => fn ($text) => '<a href="https://example.com/user/1234">' . $text . '</a>',
]);

When rendered, this will produce:

Hello, <a href="https://example.com/user/1234">Ben</a>!

See: https://formatjs.io/docs/intl#defaultrichtextelements

In the Config object, we're also able to pass in a "global" array of callbacks as the defaultRichTextElements parameter. Any named callbacks configured this way will apply to all formatted messages.

For example:

$locale = new Locale('en-US');
$config = new Config($locale, null, [
    'homeLink' => fn (string $text): string => '<a href="https://example.com>' . $text . '</a>',
    'boldface' => fn ($text) => "<strong>$text</strong>",
    'italicized' => fn ($text) => "<em>$text</em>",
]);

$message = new Message('myMessage', '<homeLink>Go <boldface>home</boldface></homeLink>, {name}!');
$messageCollection = new MessageCollection([$message]);
$formatphp = new FormatPHP($config, $messageCollection);

echo $formatphp->formatMessage([
    'id' => 'myMessage',
    'defaultMessage' => '<homeLink>Go <boldface>home</boldface></homeLink>, {name}!',
], [
    'name' => 'Samwise',
]);

This will produce:

<a href="https://example.com>Go <strong>home</strong></a>, Samwise!

Product requirements and context

To support pseudo-locale generation, we have implemented a parser that is identical to that of FormatJS. This means our parser can't properly handle attributes and all the inside guts of an HTML tag. As a result, we can't pass <a> tags with HREFs. So, we need to implement the same tag functionality as FormatJS.

I think this is the last requirement before I can finally finish pseudo locales.

How has this been tested?

PR Checklist

• I have added tests to cover my changes.

[qlty-cloud-legacy]

Code Climate has analyzed commit edf14dc and detected 1 issue on this pull request.

Here's the issue category breakdown:

Category	Count
Complexity	1

The test coverage on the diff in this pull request is 100.0% (80% is the threshold).

This pull request will bring the total coverage in the repository to 96.4% (0.1% change).

View more on Code Climate.

[jrode]

Nice tests.