Relevant portion of the HTML spec: https://html.spec.whatwg.org/multipage/forms.html#attr-fe-maxlength If I have a text <input> whose initial value exceeds the <input>'s maxlength, and the user then deletes one or more characters from the <input>'s value, but not enough characters to bring it into compliance with the maxlength, then I believe the spec requires HTMLInputElement.validity.tooLong (aka ValidityState.tooLong) to be `true` in this case. Firefox currently gives `false` in this case. Testcase: http://output.jsbin.com/xutife/1 Source code: http://jsbin.com/xutife/1/edit?html,js,output Quoting from the spec (numbering added by me): > Constraint validation: If an element > (1) has a maximum allowed value length, > (2) its dirty value flag is true, > (3) its value was last changed by a user edit (as opposed to a change made by a script), > and (4) the code-unit length of the element's value is greater than the element's maximum allowed value length, > then the element is suffering from being too long. ("suffering from being too long" = "ValidityState.tooLong is true"; see https://html.spec.whatwg.org/multipage/forms.html#dom-validitystate-toolong ) (1) is true since the <input> has a `maxlength` attribute with a valid value (namely: "5"). (2) is true since the user has edited the value of the <input> by deleting a character from it (see https://html.spec.whatwg.org/multipage/forms.html#concept-input-value-dirty-flag ). (3) is true for the same reason as (2) and since there's no JavaScript on the page that messes with the <input>'s value. (4) is true since the new value ("123456789") has a length of 9, and 9 > 5. And since all 4 conditions are true, their conjunction is thus also true.

We basically don't implement the "tooLong" validity state thing at all. There's a TODO comment in the code pointing to bug 613016 and bug 613019. The latter seems like a prerequisite for the behavior this bug is asking for.

(In reply to Boris Zbarsky [:bz] from comment #1) > We basically don't implement the "tooLong" validity state thing at all. Well, you do implement it insofar as the property exists (with a constant value of `false`) rather than being `undefined`, which makes trying to feature-detect proper support for `.tooLong` more trouble than it's worth. > The latter [bug 613019] seems like a prerequisite for the behavior this bug is asking for. Agreed.

This bug should be fixed now that bug 613019 just landed.

Indeed. Test passes in current Nightly.

Do we have tests for this?

Yes, a test for this was added as part of bug 613019. See https://hg.mozilla.org/mozilla-central/file/e2bb8f329f10/dom/html/test/test_bug613019.html