#31012 closed Bug (fixed)
Required SelectDateWidget renders invalid HTML
Reported by: | Kevin Brown | Owned by: | nobody |
---|---|---|---|
Component: | Forms | Version: | 3.0 |
Severity: | Release blocker | Keywords: | |
Cc: | Hasan Ramezani | Triage Stage: | Accepted |
Has patch: | yes | Needs documentation: | no |
Needs tests: | no | Patch needs improvement: | yes |
Easy pickings: | no | UI/UX: | no |
Description
When you render a SelectDateWidget
for a field marked as required, I noticed it produced output that does not meet the HTML standard per https://www.w3.org/TR/html5/sec-forms.html#placeholder-label-option. This was called out in the 3.0 release notes so I checked out #29056 and determined that was likely the cause.
The specific issue is that a SelectDateWidget
, when specified as a required field, will generate a placeholder
attribute on each of the <select>
boxes which it generates. The placeholder
attribute is not valid for a <select>
box and instead a "placeholder" option must be specified as an <option>
with an empty value (value=""
) with the text containing the placeholder to display.
From a Django forms perspective, this would be equivalent to setting the empty_label
for the SelectDateWidget
to (_('Year'), _('Month'), _('Day'))
, which would generate each of the placeholder label options for each of the corresponding <select>
boxes used by the widget.
It's also important to note that the HTML generated for the SelectDateWidget
in previous versions of Django was valid per the HTML standard. They did not contain a placeholder label option, but because they were not a <select multiple>
they always contained an option with a selectedness set to true (usually the first option). This meant that the required
attribute generated for them was useless, since it would never not have a value or be empty, but it did mean it was still valid HTML.
Here's some relevant parts of the standard when it comes to required select boxes.
If a select element has a required attribute specified, does not have a multiple attribute specified, and has a display size of 1; and if the value of the first option element in the select element's list of options (if any) is the empty string, and that option element's parent node is the select element (and not an optgroup element), then that option is the select element's placeholder label option.
If a select element has a required attribute specified, does not have a multiple attribute specified, and has a display size of 1, then the select element must have a placeholder label option.
Constraint validation: If the element has its required attribute specified, and either none of the option elements in the select element's list of options have their selectedness set to true, or the only option element in the select element's list of options with its selectedness set to true is the placeholder label option, then the element is suffering from being missing.
You can also find the list of accepted attributes for a <select>
at https://html.spec.whatwg.org/multipage/form-elements.html#the-select-element and https://developer.mozilla.org/en-US/docs/Web/HTML/Element/select#Attributes. Note that neither of those pages specify that a <select>
element may have a placeholder
attribute.
Change History (7)
comment:1 by , 5 years ago
Resolution: | → duplicate |
---|---|
Status: | new → closed |
comment:2 by , 5 years ago
Has patch: | set |
---|---|
Resolution: | duplicate |
Severity: | Normal → Release blocker |
Status: | closed → new |
Triage Stage: | Unreviewed → Accepted |
Type: | Uncategorized → Bug |
This isn't quite a duplicate: a regression was introduced in #31012. We'll track resolving that here.
I'm reviewing this now, but will Accept and mark as RB, since Baptiste is on the case already.
comment:4 by , 5 years ago
Cc: | added |
---|
comment:5 by , 5 years ago
Patch needs improvement: | set |
---|
As per the PR, I'm minded to revert f038214d917c982613f5a15db8dfe325b1f7479b for 3.0 and allow a new ticket to add required
handling as a New Feature.
Hi,
Thanks for the report.
I've reopened the original ticket and updated its severity to "release blocker" so I will close this one as a duplicate and the discussion can continue over there.