[Webkit-unassigned] [Bug 177449] REGRESSION(r222392): [GTK] Many forms tests are failing

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=177449

--- Comment #11 from Chris Dumez <cdumez at apple.com> ---
(In reply to Chris Dumez from comment #10)
> Ok, I think that the simulated PlatformEvents from GTK may be 0 as timestamp
> (WallTime). When it gets converted to a MonotonicTime, we may end with a
> negative value when calling WallTime::approximateMonotonicTime().
> 
> If we get a negative MonotonicTime value, the following check in
> TypeAhead::handleEvent() will always be true:
>     if (event->timeStamp() < m_lastTypeTime)
>         return -1;
> 
> .. because m_lastTypeTime is 0.

I will propose a fix making the code more robust against having PlatformEvent::timestamp() returning 0. However, if this is really the problem, I think the GTK port should make sure to to initialize it with WallTime::now() when it is 0.

I think this may be observable by firing an event using eventSender.keyDown() then checking event.timeStamp in JS.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20170925/7bd11d4b/attachment.html>