Bug 43973 – <input type=number> should not have an invalid number as its value

• Status: RESOLVED
• Resolution: FIXED
• Component: Forms

[Kent Tamura]

Making <input type=number> conformable to the current HTML5 spec.
Discussed at http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2010-August/027896.html

[Kent Tamura]

Created attachment 64346 [details]
Patch

[Kent Tamura]

Created attachment 65892 [details]
Patch 2 (rebased)

[Dimitri Glazkov (Google)]

Comment on attachment 65892 [details]
Patch 2 (rebased)

View in context: https://bugs.webkit.org/attachment.cgi?id=65892&action=prettypatch

> WebCore/dom/InputElement.h:64
> +    virtual bool isCommittableValue(const String&) const = 0;
How about isAcceptableValue? Committable seems unclear.

> WebCore/rendering/RenderTextControlSingleLine.cpp:681
> +        else if (!node()->isHTMLElement() || !static_cast<HTMLInputElement*>(node())->formControlValueMatchesRenderer())
What does this check mean?

[Kent Tamura]

Created attachment 66343 [details]
Patch 3

[Kent Tamura]

Thank you for reviewing!

(In reply to comment #3)
> > WebCore/dom/InputElement.h:64
> > +    virtual bool isCommittableValue(const String&) const = 0;
> How about isAcceptableValue? Committable seems unclear.

Done.

> > WebCore/rendering/RenderTextControlSingleLine.cpp:681
> > +        else if (!node()->isHTMLElement() || !static_cast<HTMLInputElement*>(node())->formControlValueMatchesRenderer())
> What does this check mean?

Good catch.  This was experimental code for localized numbers and is not needed at this moment. I removed it.

[Dimitri Glazkov (Google)]

Comment on attachment 66343 [details]
Patch 3

yay, team!

[Kent Tamura]

Comment on attachment 66343 [details]
Patch 3

Clearing flags on attachment: 66343

Committed r67164: <http://trac.webkit.org/changeset/67164>

[Kent Tamura]

All reviewed patches have been landed.  Closing bug.