Souheil Chelfouh
Souheil Chelfouh
This issue/request is interesting. This would mean we have to handle the connection upgrade request and give back an I/O over which the websocket's handler can communicate with the browser....
This is no longer the case, with bs5 markings. I'm not sure this qualifies for a backport.
Is this still an issue we need to keep alive ?
the notation "${oid}HelpBlock" in camel-case seems a bit of out what is currently done with the rest. Maybe "help-${oid}" ?
current, in the mapping_item template, we have "req-${oid}" and "item-${oid}". I thought it would be aligned with this kind of notation
The wrapping div, currently, allows us to group the label and the input/input-group, the feedback and the error in a logical container, named #item-${oid}. I don't think removing it would...
As a note : there are some exceptions. For the complex widgets, some wrapping divs still need to hold the error marker (is-invalid). Without this, the invalid-feedback block does not...
The question arises for "required". Without it, there's no marker to show a field being required.
The ":has(+...)" does not work on every browser. It does not, on my firefox but does on my chromium.
After fiddling around in the templates, I also fail to see how we'd use a purely CSS solution with composite fields, like the date parts. The label is then linked...