Bug 89637 – [Chromium] keyDown event in JavaScript: event.shiftKey, event.ctrlKey and event.altKey are always false

• Component: DOM
• Status: RESOLVED FIXED

[Yusuke Sato]

Chromium for Linux does not set .shiftKey, .ctrlKey, and .altKey correctly. For example, when Control key is pressed, the ctrlKey attribute of the event is false, and when it's released the attribute becomes true. Safari, Firefox, and Chrome for Windows/Mac/ChromeOS do the opposite, true on keyDown and false on keyUp, which should be correct.

http://code.google.com/p/chromium/issues/detail?id=127142

[Yusuke Sato]

Created attachment 148988 [details]
Patch

[Tony Chang]

Comment on attachment 148988 [details]
Patch

Can we write a layout test for this or does eventSender mock the events lower than this?

[Tony Chang]

Comment on attachment 148988 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=148988&action=review

> Source/WebKit/chromium/ChangeLog:8
> +        This is a Gtk port of http://crrev.com/142209.
> +        
> +        Reviewed by NOBODY (OOPS!).

The "Reviewed by..." line goes below the bug link. See the other entries in the ChangeLog.

> Source/WebKit/chromium/src/gtk/WebInputEventFactory.cpp:277
> +// Normalizes event->state to make it Windows/Mac compatible. Since the way
> +// of setting modifier mask on X is very different than Windows/Mac as shown
> +// in http://crbug.com/127142#c8, the normalization is necessary.

I would move this comment into the ChangeLog.

[Yusuke Sato]

Checked DumpRenderTree/chromium/EventSender.cpp. Since the sender mocks an event by directly setting event->modifiers, I think it's hard to write a layout test for my patch.

[Yusuke Sato]

Created attachment 149730 [details]
Patch

[WebKit Review Bot]

Comment on attachment 149730 [details]
Patch

Clearing flags on attachment: 149730

Committed r121397: <http://trac.webkit.org/changeset/121397>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.