Closed donquixote closed 5 months ago
I am not sure I understand the needs for the "better detection of empty elements".
For this you can check my other comment..
I do not agree on having a dedicated formatter per variation of classes, with tiny differences like attributes/tags.
The alternatives would be:
ui_patterns_field_group
module (see below).The motivation here was to be able to quickly configure content type displays with many different sections, where a "section pattern" is the natural default formatter. The most common behavior should be achievable without going into the options subform.
The ui_patterns_field_group
is the opposite of this, and requires lots of maintenance to achieve the basic behavior.
The need for an extra field for the title..
What I did in the branch POC/OEL-3000 is:
- install ui_patterns_field_group module.
- created an extra field that renders a label only if the section fields are not empty.
We need to avoid that individual websites need to create extra fields where they have to hard-code the field names. And I think your emptiness check in the extra field is not even fully reliable, because it does not cover the case where a field has content, but the visitor has no access to it.
- created a field group using the pattern formatter. Moved the fields above like in this picture: [...] With this little piece of code and the configuration in the page, the section was rendered only when the inner fields were not empty.
But this "little piece of code" would have to be replicated across different website projects, for different field groups.
And if you add a new field into the group, now you need to update the field mapping in the group settings, and you need to add the new field in the condition of the extra field.
The classes can be customised per field group instance.
We should avoid having people customize classes for each field group, if the main use case is to have standardized section and heading levels with standardized spacing.
The goal of the new patterns that we will be implementing will be to move as much towards this approach.
I think the pattern integration modules shipped with ui_patterns are a dead end, because they have to accommodate a wide range of possible pattern parameter signatures. This always deletes to complex configuration forms, with many degrees of freedom, and no validation to determine if a field mapping makes sense with the parameter type.
The goal should be to reduce complexity and unnecessary choice in the website projects,
Changes:
Up for discussion:
.mb-5
a good choice for the section margin? In our project we use.my-5
, but we have a trick to clear the top margin (the.eat-margin-top
trick).