There is a huge Drupal core issue about evaluating a render array as true or false from a Twig template: 🌱 [meta] Themes improperly check renderable arrays when determining visibility Needs work
In certain situations regions that contain empty blocks still render.
The core issue is that there is no possible way to definitely identify an 'empty' block without first
rendering it.
The ideal result is the ability to add this to a template:{% if region %} <div>{{ region }}</div> {% endif %}
Very annoying but, after 13 years, still no fix. But I hope it may be fixable at SDC level.
Let's do some tests on Drupal 10.1.1 with this variable:
{% if description %}<p class="fr-tile__desc">{{ description }}</p>{% endif %}
We expect description to be resolved at false.
$variables['description'] = [];
OK: description is false, because in PHP an empty array is false
$variables['description'] = [
"#cache" => [
"contexts" => [
"user.permissions"
],
"tags" => [],
"max-age" => -1
]
]KO: description is true
$variables['description'] = [
[
"#cache" => [
"contexts" => [
"user.permissions"
],
"tags" => [],
"max-age" => -1
]
]
]KO: description is true
Layout Builder is sometimes wrapping the render array in a mapping where the key is the block UUID:
$variables['description'] = [
"cf124739-3274-41cc-829d-c3465ee4eaa5" => [
"#cache" => [
"contexts" => [
"user.permissions"
],
"tags" => [],
"max-age" => -1
]
]
]KO: description is true.
$variables['description'] = [
"whatever" => []
];KO: description is true.
$variables['description'] = [
"#markup" => "",
];KO: description is true.
Same with #plain_text I guess...
So, it may be fixable at SDC level because we are lucky to know:
[#type => 'component'])2 informations other render elements are not able to know easily.
So, in ComponentElement::preRenderComponent() for each component slot:
#cache, #attached... what else?) to the component render element if they are bubbable#weight... what else?) if they are not bubbableSo, the renderable will be evaluated to false by Twig if empty 👍
Active
11.0 🔥
One of the challenges I see with this is that we will have to identify and manually handle all these non-printable keys. This includes bubbling, and or erasing some keys.
I would love to have this feature, but I am not sure a solid / maintainable implementation is feasible. Let's see what an initial patch looks like.