Bug 65875 – [Qt][WK2] MiniBrowser is firing twice the QDesktopWebView::mousePressEvent method

• Component: Tools / Tests
• Status: RESOLVED FIXED

[Igor Trindade Oliveira]

Using MiniBrowser i noticed it is firing twice the QDesktopWebView::mousePressEvent method when the moused is pressed anywhere on the screen.

[Igor Trindade Oliveira]

Created attachment 103291 [details]
Patch.

Proposed patch

[Benjamin Poulain]

Comment on attachment 103291 [details]
Patch.

View in context: https://bugs.webkit.org/attachment.cgi?id=103291&action=review

> Tools/MiniBrowser/qt/MiniBrowserApplication.cpp:107
>              if (m_touchPoints.contains(touchPoint.id()))
>                  touchPoint.setState(Qt::TouchPointMoved);
> -            else
> -                touchPoint.setState(Qt::TouchPointPressed);
> +            else {
> +                if (mouseEvent->modifiers().testFlag(Qt::ControlModifier))
> +                    touchPoint.setState(Qt::TouchPointPressed);
> +            }

It seems to me that the problem might be that m_touchPoints.contains(touchPoint.id()) == false.
If the touchpoint is still active because, it should be in m_touchPoints.

[Igor Trindade Oliveira]

Created attachment 103384 [details]
Test Case

Adding Test case.

[Igor Trindade Oliveira]

Probably we are talking about two different things. I added a test case where onmousedown event is firing two times the callback. It happens because  QDesktopWebView::mousePressEvent is called twice.

(In reply to comment #2)
> (From update of attachment 103291 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=103291&action=review
> 
> > Tools/MiniBrowser/qt/MiniBrowserApplication.cpp:107
> >              if (m_touchPoints.contains(touchPoint.id()))
> >                  touchPoint.setState(Qt::TouchPointMoved);
> > -            else
> > -                touchPoint.setState(Qt::TouchPointPressed);
> > +            else {
> > +                if (mouseEvent->modifiers().testFlag(Qt::ControlModifier))
> > +                    touchPoint.setState(Qt::TouchPointPressed);
> > +            }
> 
> It seems to me that the problem might be that m_touchPoints.contains(touchPoint.id()) == false.
> If the touchpoint is still active because, it should be in m_touchPoints.

[Igor Trindade Oliveira]

Created attachment 103418 [details]
Patch

Updating proposed patch with a better ChangeLog explaining the modifications.

[WebKit Review Bot]

Attachment 103418 [details] did not pass style-queue:

Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'Tools/ChangeLog', u'Tools/MiniBrowser/qt/M..." exit_code: 1

Tools/MiniBrowser/qt/MiniBrowserApplication.cpp:104:  Missing space before ( in if(  [whitespace/parens] [5]
Total errors found: 1 in 2 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Igor Trindade Oliveira]

Created attachment 103419 [details]
Patch

Proposed patch.

[Benjamin Poulain]

Comment on attachment 103419 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=103419&action=review

This has nothing to do with the ControlModifier. The problem is we still send the original mouse event after we send the touch events. QWidget also generate a fake mouse event for the touch event so we end up with 2 mouse events. The original event should be discarded when a touch event is generated.

> Tools/ChangeLog:9
> +        On some windowing systems, mouse events are also sent for the primary touch point. This
> +        means it is possible for your widget to receive both QTouchEvent and QMouseEvent for the
> +        same user interaction point. MiniBrowser when setting the Qt::TouchPointPressed for a touch

Qt is supposed to always give you a mouse event for the primary touch point.

> Tools/ChangeLog:12
> +        to be called twice for a single mouse click. Now the Qt::TouchPointPressed is used just
> +        when Ctrl modifier is pressed.

This is not at all the intended behavior of the code of touch mocking.

[Igor Trindade Oliveira]

Created attachment 103500 [details]
Patch

Proposed patch.

[Igor Trindade Oliveira]

Created attachment 103509 [details]
Patch

Updated patch.

[Benjamin Poulain]

Created attachment 103513 [details]
Patch

[Benjamin Poulain]

Shouldn't this approach just work?

[Benjamin Poulain]

Comment on attachment 103513 [details]
Patch

I cq-, let's not land that before Igor has a look.

[Benjamin Poulain]

Created attachment 103520 [details]
Patch

[Kenneth Rohde Christiansen]

Comment on attachment 103520 [details]
Patch

Looks OK. If this solves it, r=me

[Benjamin Poulain]

Comment on attachment 103520 [details]
Patch

Let's land that. Freaking QWidget hacks... :)

[WebKit Review Bot]

Comment on attachment 103520 [details]
Patch

Clearing flags on attachment: 103520

Committed r92799: <http://trac.webkit.org/changeset/92799>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.