Linux: Don't unparent unselected tab contentses....

Linux: Don't unparent unselected tab contentses.

Instead, hide them. Sometimes an unselected tab contents can get unhandled keyboard events after it's been deselected (if the renderer was really slow in handling and returning the event).

http://crbug.com/12178


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=16329

[willchan no longer on Chromium, 2009-05-18 21:01:02]

Took me a second to grok this.  I'm still confused as to how the focus is
getting ahead of the selection, but given that you say it is so, this lgtm.  For
my own clarification, can you explain how this happens though?

[Evan Stade, 2009-05-18 21:16:23]

yea. After investigating I found that something slightly different is happening:

1) user presses ctrl-pgdn, we forward to webkit
2) user presses ctrl-pgdn, we forward to webkit
3) webkit returns with UnhandledKeyboardEvent for (1). We switch to a new tab
contents.
4) webkit returns with UnhandledKeyboardEvent for (2), but by this time the tab
contents has been deselected and unparented, hence we get a null toplevel
window.

so, fix is still the same :). This should also make getting the top level for
the tab contents view more reliable in general (which is used for various things
like focus).

[willchan no longer on Chromium, 2009-05-18 21:34:21]

On 2009/05/18 21:16:23, Evan Stade wrote:
> yea. After investigating I found that something slightly different is
happening:
> 
> 1) user presses ctrl-pgdn, we forward to webkit
> 2) user presses ctrl-pgdn, we forward to webkit
> 3) webkit returns with UnhandledKeyboardEvent for (1). We switch to a new tab
> contents.
> 4) webkit returns with UnhandledKeyboardEvent for (2), but by this time the
tab
> contents has been deselected and unparented, hence we get a null toplevel
> window.
> 
> so, fix is still the same :). This should also make getting the top level for
> the tab contents view more reliable in general (which is used for various
things
> like focus).

Ok, this makes sense.  Thanks for the explanation.