267357 - DOM value of text field value attribute vs user edit

Status: RESOLVED INVALID

(Core :: DOM: Core & HTML, defect)

Description

[Dan M]

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.3) Gecko/20041101
Firefox/1.0RC2

Editing a text field causes a disconnect between its displayed value and its
value to script. textElement.getAttribute always returns the value last set by
textElement.setAttribute, ignoring user edits. Further, textElement.setAttribute
will change the text displayed in the text element, but not after the user has
edited the text field.

Originally reported by brilliantcrank,
http://forums.mozillazine.org/viewtopic.php?t=154454

Comment 1

[Dan M]

Attachment: testcase

Comment 2

[Dan M]

Oops. I got the attribution wrong. Reported by rjm,
http://forums.mozillazine.org/viewtopic.php?t=154210

Comment 3

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

(In reply to comment #0)
> Editing a text field causes a disconnect between its displayed value and its
> value to script. textElement.getAttribute always returns the value last set by
> textElement.setAttribute, ignoring user edits.

It's supposed to work that way.  The attribute is equivalent to the
.defaultValue property.

http://www.w3.org/TR/DOM-Level-2-HTML/html.html#ID-6043025

> Further, textElement.setAttribute
> will change the text displayed in the text element, but not after the user has
> edited the text field.

This seems reasonable, but I don't think the spec says either way.  What do
other implementations do?

Comment 4

[Johnny Stenback (:jst)]

Yeah, all per design. There's been *tons* of bugs about this, trying to balance
our implementation between what's expected on the web, what the standards say,
and what IE does. Marking INVALID.

Comment 5

[Dan M]

Guess I should heed my own advice about looking for closed bugs before filing
new ones :) Another one of those W3 vs IE things, eh? Gotta say, IE's
interpretation makes more sense to me. Shutting up now.

Comment 6

[Hixie (not reading bugmail)]

Note that the current proposed WHATWG prose disagrees with Mozilla here:

   http://whatwg.org/specs/web-forms/current-work/#the-defaultvalue

It isn't final, I'm awaiting comments from a few people (hyatt in particular)
before having any sort of firm opinion on this. If you want it to change please
mail the [e-mail] list explaining what you think it should say, and
why. Compatibility with IE without losing sanity is an important priority here.
(Strict compatibility with IE is impossible without losing sanity here.)

Comment 7

[Hixie (not reading bugmail)]

Actually looks like we're going with what Mozilla does here after all. Trying to
be compatible with IE any more than Mozilla is seems nigh on impossible.