new-run-webkit-tests: stop driver on windows properly
Created attachment 88160 [details] Patch
Comment on attachment 88160 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=88160&action=review > Tools/Scripts/webkitpy/layout_tests/port/chromium.py:-524 > - # http://bugs.python.org/issue1731717 I've removed the comments about poll(), because it looks like it only applied to 2.4 and if we weren't reliably calling _proc.wait().
Created attachment 88168 [details] add unit test
Comment on attachment 88168 [details] add unit test View in context: https://bugs.webkit.org/attachment.cgi?id=88168&action=review > Tools/Scripts/webkitpy/layout_tests/port/chromium.py:-524 > - # http://bugs.python.org/issue1731717 I've deleted this comment about poll()'s thread safety because it looks like it was only relevant on Python 2.4 and when we weren't reliably calling proc.wait().
Comment on attachment 88168 [details] add unit test View in context: https://bugs.webkit.org/attachment.cgi?id=88168&action=review Did you test this on win32 python and cygwin python? > Tools/Scripts/webkitpy/layout_tests/port/chromium.py:337 > + self.KILL_TIMEOUT = 3.0 Nit: pull this out of __init__? You can still refer to it with self.KILL_TIMEOUT and shadow it in a test by setting driver.KILL_TIMEOUT.
I will move it out of the __init__. I'm stil banging on this one and some related bug fixes intended to stabilize the win ports, so there may or may not be additional changes for this patch.
Committed r83384: <http://trac.webkit.org/changeset/83384>