This fixes http://code.google.com/p/chromium/issues/detail?id=3007, which is ...

This fixes http://code.google.com/p/chromium/issues/detail?id=3007, which is the mouse-click-events.html
layout test failure. Fix as per below:-

1. The test shell mouse event generating mechanism now tracks the last mouse button number received. This is used to determine whether the click count continues to increment or not. 

2. We now default to middle button if the button number passed in from js contains anything other than 0 or 2. This is as per the event sending mechanism in webkit mac.

We still need to re-baseline this test as the number of events printed on the page cause a scrollbar to show up for the 4th mouse event test. This causes the middle mouse button to be eaten as per recent changes in EventHandler.cpp. 

Bug=3007
R=tony


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=2917

[tony, 2008-10-05 22:34:44]

It looks like we set the webmouseevent::layout_test_click_count in
webkit/tools/test_shell/event_sending_controller.cc.  Is that not enough?  I
think this test used to pass, it would be nice if we could fix
event_sending_controller so it passes again.

[ananta, 2008-10-06 22:24:55]

Hi Tony

The latest snapshot has been updated as per the following:-
1. The test shell mouse event generating mechanism now tracks the last mouse
button number received. This is used to determine whether the click count
continues to increment or not. 

2. We now default to middle button if the button number passed in from js
contains anything other than 0 or 2. This is as per the event sending mechanism
in webkit mac.

We still need to re-baseline this test as the number of events printed on the
page cause a scrollbar to show up for the 4th mouse event test. This causes the
middle mouse button to be eaten as per recent changes in EventHandler.cpp. 

Thanks
Ananta

[tony, 2008-10-06 22:35:56]

LGTM!

Nit: I would probably just make everything file static (or in an anonymous
namespace), but this is fine too.

On 2008/10/06 22:24:55, ananta wrote:
> Hi Tony
> 
> The latest snapshot has been updated as per the following:-
> 1. The test shell mouse event generating mechanism now tracks the last mouse
> button number received. This is used to determine whether the click count
> continues to increment or not. 
> 
> 2. We now default to middle button if the button number passed in from js
> contains anything other than 0 or 2. This is as per the event sending
mechanism
> in webkit mac.
> 
> We still need to re-baseline this test as the number of events printed on the
> page cause a scrollbar to show up for the 4th mouse event test. This causes
the
> middle mouse button to be eaten as per recent changes in EventHandler.cpp. 
> 
> Thanks
> Ananta