Constraint validation for readonly doesn't correspond implementations

[dSalieri]

What is the issue with the HTML Standard?

Constraint validation: If the readonly attribute is specified on an input element, the element is barred from constraint validation.

The thing is that this constraint validation does not correspond to reality. The specification specifies it for any input elements.

And here's what I discovered:

For chrome, elements: input, button, select, form-associated custom elements - when adding the readonly attribute, the willValidate getter changes to false.

For firefox, elements: input, form-associated custom elements - when adding the readonly attribute, the willValidate getter changes to false (the unspecified elements that were in chrome obviously didn't change the state of the willValidate getter).

That is, we see that there are inconsistencies between implementations, but not only do the implementations not correspond to each other, but none of the implementations corresponds to the specification.

It feels like the specification needs to be edited. But in which direction? Chrome or Firefox?

[annevk]

What makes you say the specification is not aligned with Firefox? I think Firefox is correct here. The readonly attribute is not even defined for button or select.

[annevk]

Keeping this open for now as it seems we might need some WPT coverage for this.

[dSalieri]

What makes you say the specification is not aligned with Firefox? I think Firefox is correct here. The readonly attribute is not even defined for button or select.

Since the specification does not say that the readonly attribute has an effect on form-associated custom elements, Firefox is very close.

@annevk

[annevk]

@dSalieri it does? It's literally three paragraphs down from your link.

[dSalieri]

@dSalieri it does? It's literally three paragraphs down from your link.

Oops, I actually didn't look there. I take back what I said about the Firefox inconsistency.

But in any case, there is a discrepancy between the browsers.