Bug 103869 – INPUT_MULTIPLE_FIELDS_UI doesn't render digits well in RTL locales

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

[Kent Tamura]

http://code.google.com/p/chromium/issues/detail?id=163108

Many RTL locales use LTR digits. So, we should show "23:59" (visual order) for "23:59" (logical order) even in an RTL locale.  WebCore applies such bidi re-ordering automatically. However,  we wrap these characters with display:inline-block elements in INPUT_MULTIPLE_FIELDS_UI and WebCore can't apply the bidi re-ordering to block elements. We should use display:inline in INPUT_MULTIPLE_FIELDS_UI.

[Kent Tamura]

Created attachment 177412 [details]
Patch

[Kentaro Hara]

Comment on attachment 177412 [details]
Patch

morrita-san: I'm not familiar with the code. Could you take a look?

[Hajime Morrita]

Comment on attachment 177412 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=177412&action=review

> Source/WebCore/dom/Element.h:398
> +    virtual bool isDateTimeFieldElement() const;

Nit: OK to be inlined like other isXxxElement(). Having extra if/endif in .cpp file seems a bit mess.

> Source/WebCore/html/shadow/DateTimeEditElement.cpp:358
> +            width += static_cast<DateTimeFieldElement*>(childElement)->maximumWidth(style->font());

You might able to override didStyleRecalc() instead of this one. It allows you to modify createdStyle.

> Source/WebCore/html/shadow/DateTimeEditElement.cpp:362
> +            width += style->font().width(childElement->textContent());

Doesn't textContent() change?

[Hajime Morrita]

Comment on attachment 177412 [details]
Patch

I hope there were better way to handle this kind of workaround. I have no good idea though.

[Kent Tamura]

Comment on attachment 177412 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=177412&action=review

>> Source/WebCore/html/shadow/DateTimeEditElement.cpp:362
>> +            width += style->font().width(childElement->textContent());
> 
> Doesn't textContent() change?

Right. It is fixed.

[Kent Tamura]

(In reply to comment #4)
> (From update of attachment 177412 [details])
> I hope there were better way to handle this kind of workaround. I have no good idea though.

The calculation is a kind of layout.  So the ideal solution might be to add a new renderer ;-(

[Kent Tamura]

Committed r136808: <http://trac.webkit.org/changeset/136808>