Add plugin for tracking non-LinkIt links in WYSIWYG fields

@BryanDeNijs thanks! Patch looks great, just some minor remarks:

1. +++ b/src/Plugin/EntityUsage/Track/HtmlFieldLink.php
   @@ -0,0 +1,155 @@
   + *
   + * @EntityUsageTrack(
   + *   id = "html_field_link",
   

   How about just html_link as the ID?

2. +++ b/src/Plugin/EntityUsage/Track/HtmlFieldLink.php
   @@ -0,0 +1,155 @@
   + *   label = @Translation("HTML Fields links"),
   

   Maybe just "HTML Links" as label?

3. +++ b/src/Plugin/EntityUsage/Track/HtmlFieldLink.php
   @@ -0,0 +1,155 @@
   + *   description = @Translation("Tracks usage of entities linked HTML (WYSIWYG)
   + *   fields."),
   

   The description inside the annotation should be in a single line, even if longer than 80chars.

   Also, maybe "Tracks usage of entities referenced from regular HTML Links" could be a bit clearer?

It would be great to have test coverage for this plugin as well. Can you work on this? Let me know if not.

Thanks!

`not(starts-with(@href, 'http://')) and not(starts-with(@href, 'https://'))`

If you want to support unmanaged links then you also should consider that unqualified users also paste links to own site incl domain. Pathologic tries to identify them by supporting a domain list. Maybe support pathologic or have own setting?

IMHO you should never use a current host in any processing. The host can vary based on context, for instance with a per-language domain setup or while cron run.

Adding this will cause pain as it contains risk of unstable conditions with various results.

So IMHO this is oversimplified and not ready, decision up to the maintainer. ;-)

Maybe add better support for multiple hosts in a followup?

I'm OK with including a feature that only partly addresses the need, but I would try to avoid including something we know is going to cause issues in some scenarios.

Why don't we only support relative links for now, and create a follow-up to extend that to support absolute URLs as well in the future? As long as we document it on the project page / documentation, I believe this would be the best option.

Quick review.

1. +++ b/src/Form/EntityUsageSettingsForm.php
   @@ -190,6 +190,12 @@ class EntityUsageSettingsForm extends ConfigFormBase {
   +      '#description' => $this->t('A comma or new-line seperated list of domain names that should be assumed as a domain for this website. When a user inserts an absolute URL to an entity with one of these domains, it will be processed as an internal entity usage.'),
   

   One type: seperated > separated

   Some suggestions:
   that should be assumed as a domain for this website > for this website

   When a user inserts an absolute URL to an entity with one of these domains, it will be processed as an internal entity usage. > Absolute URLs in content will be checked against these domains to allow usage tracking.

2. +++ b/src/Form/EntityUsageSettingsForm.php
   @@ -205,8 +211,13 @@ class EntityUsageSettingsForm extends ConfigFormBase {
   +    $site_domains = preg_replace('/[\s, ]/', ',', $form_state->getValue('site_domains'));
   +    $site_domains = array_filter(explode(',', $site_domains));
   

   Can we add this above the block of code setting the config values and let it site nicely below each other for readability?

3. +++ b/src/Plugin/EntityUsage/Track/HtmlLink.php
   @@ -0,0 +1,162 @@
   +            $href = preg_replace($host_pattern, '/', $href);
   

   I think this is missing a break;?

4. +++ b/tests/src/Kernel/HtmlLinkKernelTest.php
   @@ -0,0 +1,234 @@
   +   * Modules to install.
   +   *
   +   * @var array
   

   {@inheritdoc}

5. +++ b/tests/src/Kernel/HtmlLinkKernelTest.php
   @@ -0,0 +1,234 @@
   +
   

   Useless newline

I'm ok with this, @marcoscano could you do another check?

Great work! I think we're almost there.

1. +++ b/config/schema/entity_usage.schema.yml
   @@ -19,8 +21,11 @@ entity_usage.settings:
   diff --git a/interdiff-14-16.txt b/interdiff-14-16.txt
   
   diff --git a/interdiff-14-16.txt b/interdiff-14-16.txt
   new file mode 100644
   
   new file mode 100644
   index 0000000..7416252
   
   index 0000000..7416252
   --- /dev/null
   
   --- /dev/null
   +++ b/interdiff-14-16.txt
   
   +++ b/interdiff-14-16.txt
   +++ b/interdiff-14-16.txt
   @@ -0,0 +1,74 @@
   

   seems like the interdiff got unadvertently into the patch :)

2. +++ b/src/Plugin/EntityUsage/Track/HtmlLink.php
   @@ -0,0 +1,163 @@
   +   * Constructs display plugin.
   

   I would use a generic text here, or the actual name of the object we are constructing.

3. +++ b/src/Plugin/EntityUsage/Track/HtmlLink.php
   @@ -0,0 +1,163 @@
   +    // Loop trough all the <a> elements that link to internal URLs and don't
   +    // have the LinkIt attributes.
   

   Seems like this comment is a bit obsolete now?

4. +++ b/src/Plugin/EntityUsage/Track/HtmlLink.php
   @@ -0,0 +1,163 @@
   +        // Strip of the scheme and host, so we only get the path.
   

   s/of/off

5. +++ b/src/Plugin/EntityUsage/Track/HtmlLink.php
   @@ -0,0 +1,163 @@
   +          if ($files) {
   +            // File entity found.
   +            $target_type = 'file';
   +            $target_id = array_keys($files)[0];
   +          }
   

   We currently don't have a usage listing for a target file entity. So this would only be useful for people using this as an API. Nothing against letting it in though.

\o/

This is a nice one, thanks!