Hi webkit-dev,
If the following two tests are failing, it's probably because an old
pywebsocket process is still alive and preventing the new one from starting:
http/tests/websocket/tests/client-close.html
http/tests/websocket/tests/server-close.html
You can observe the following traceback in the layout test output:
Traceback (most recent call last):
File "Tools/Scripts/new-run-webkit-websocketserver", line 108, in
<module>
main()
File "Tools/Scripts/new-run-webkit-websocketserver", line 103, in main
pywebsocket.start()
File
"/Volumes/Big/WebKit-BuildSlave/leopard-intel-release-tests/build/Tools/Scripts/webkitpy/layout_tests/port/websocket_server.py",
line 220, in start
(self._server_name, self._port))
webkitpy.layout_tests.port.websocket_server.PyWebSocketNotStarted:
Failed to start PyWebSocket server on port 8880.
If this happens, please kill the old pywebsocket process, or reboot the
machine.
*Dear WebKit buildbot administrators:*
A couple of buildbot slaves are affected by this issue (as far as I know).
Specifically:
- apple-macpro-4 (
http://build.webkit.org/builders/Leopard%20Intel%20Release%20%28Tests%29/builds/32745
)
- apple-xserve-7 (
http://build.webkit.org/builders/SnowLeopard%20Intel%20Release%20%28WebKit2%20Tests%29/builds/12121
)
Could you kill old pywebsocket processes in these bots? After killing them,
the above two tests should start to pass.
*Dear webkitpy hackers:*
run-webkit-websocketserver probably should check if there's any existing
WebSocket server running, and kill it before starting the new one.
However, as I don't know very much about how run-webkit-websocketserver
works, I don't know if this is feasible or not. Do you think this is a good
idea? Is it easily implementable?
Sorry for all the mess, and thank you for your cooperation.
Regards,
Yuta
_______________________________________________
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
