Some change in this range caused this test to fail on Wk2: http://trac.webkit.org/log/?action=stop_on_copy&mode=stop_on_copy&rev=132343&stop_rev=132333&limit=100
It's working fine in my tests here if I only run this specific test. Perhaps some flakyness related to my focus commit?
I've finally reproduce it here. Let the fun begin.
Going back to r132332 did not help, and the test still failed most of the time. My guess is that it is somewhat flaky right, and has been so for a while.
This is turning out to be interesting. I can only reproduce it here if I set the focus model in my window manager to a variation of "focus follows mouse" plus "automatically move the mouse pointer to newly created windows" and run WebKitTestRunner with the pointer outside of the region the window is going to occupy when it is created...
I don't think there is an easy fix for this -- the test fails if you manage to move the mouse pointer in the window while it is in the process of selecting the elements in the box (which is made easier when the window manager automatically moves the mouse pointer for you as I described in comment #4). Since the test seems to be consistently passing in the bot again these days, I'll just unskip it again and hope for the best.
Committed r133010: <http://trac.webkit.org/changeset/133010>