While working on bug #255518 one of the checks I wrote for the API tests
was something like this:

   // Check that enabled status is the same as the declared default.
   GRefPtr<WebKitSettings> settings = adoptGRef(webkit_settings_new());
   for (gsize i = 0; i < allFeaturesCount; i++) {
        auto* feature = webkit_feature_list_get(allFeatures, i);
        g_assert(webkit_settings_get_feature_enabled(settings.get(), feature) == webkit_feature_get_default_value(feature));
   }

This fails for some settings for which the default value declared in
the UnifiedWebPreferences.yaml file does not match the default value for
the corresponding WebKitSettings property. As en example:

  # From UnifiedWebPreferences.yaml
  EncryptedMediaAPIEnabled:
    type: bool
    status: embedder
    condition: ENABLE(ENCRYPTED_MEDIA)
    defaultValue:
      WebKitLegacy:
        default: false
      WebKit:
        default: true
      WebCore:
        default: false


  // WebKitSettings.cpp
  sObjProperties[PROP_ENABLE_ENCRYPTED_MEDIA] =
      g_param_spec_boolean(
          "enable-encrypted-media",
          _("Enable EncryptedMedia"),
          _("Whether EncryptedMedia should be enabled."),
          FALSE,
          readWriteConstructParamFlags);

Note how the YAML file uses “true” as the default value on the WebKit
layer, but then the WebKit GLib API code uses “FALSE” as the default.

We should either derive the values at the GLib API layer from the YAML
file, or update the YAML file so it reflects the reality imposed by the
GLib API code. This would be good make the new API introduced in
bug #255518 behave consistently with the existing WebKitSettings
properties which happen to control WebKit features.