Telemetry fixit: use a minimal ws server for testSockOpts.

Telemetry fixit: use a minimal ws server for testSockOpts.

Avoid using external servers.

BUG=361658
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=266160

[bulach, 2014-04-24 10:54:13]

ptal

[tonyg, 2014-04-24 12:54:20]

lgtm

beautiful :)

https://codereview.chromium.org/258623002/diff/1/tools/telemetry/telemetry/co...
File tools/telemetry/telemetry/core/backends/chrome/websocket_unittest.py
(right):

https://codereview.chromium.org/258623002/diff/1/tools/telemetry/telemetry/co...
tools/telemetry/telemetry/core/backends/chrome/websocket_unittest.py:41: httpd =
BaseHTTPServer.HTTPServer(('', 6565), _FakeWebSocketHandler)
Instead of 6565, can we use 0 to let the OS choose a port, then use httpd.port
in the create_connection()s below?

https://codereview.chromium.org/258623002/diff/1/tools/telemetry/telemetry/co...
tools/telemetry/telemetry/core/backends/chrome/websocket_unittest.py:54:
ws.sock.getsockopt(socket.IPPROTO_TCP, socket.TCP_NODELAY), 0)
Now that we have a whole server like this it seems a shame we don't verify
something else about the connection being established. But not necessary for
this patch.

[bulach, 2014-04-24 13:40:37]

thanks tony! inline, I'll wait for dtu before CQ:

https://codereview.chromium.org/258623002/diff/1/tools/telemetry/telemetry/co...
File tools/telemetry/telemetry/core/backends/chrome/websocket_unittest.py
(right):

https://codereview.chromium.org/258623002/diff/1/tools/telemetry/telemetry/co...
tools/telemetry/telemetry/core/backends/chrome/websocket_unittest.py:41: httpd =
BaseHTTPServer.HTTPServer(('', 6565), _FakeWebSocketHandler)
On 2014/04/24 12:54:20, tonyg wrote:
> Instead of 6565, can we use 0 to let the OS choose a port, then use httpd.port
> in the create_connection()s below?

what's wrong with favorite flaky number?! :)
you're of course right, using 0.

https://codereview.chromium.org/258623002/diff/1/tools/telemetry/telemetry/co...
tools/telemetry/telemetry/core/backends/chrome/websocket_unittest.py:54:
ws.sock.getsockopt(socket.IPPROTO_TCP, socket.TCP_NODELAY), 0)
On 2014/04/24 12:54:20, tonyg wrote:
> Now that we have a whole server like this it seems a shame we don't verify
> something else about the connection being established. But not necessary for
> this patch.

indeed! a few caveats though:
1) the connect actually does a full-blown HTTP request, so this is exercising a
somewhat big surface already.

2) the response has to be framed, which is not something truly trivial... I
could import the "ABNF" class from third_party/websocket-client, but then it'd
end up testing itself. :)

I think for the time being this should be enough, but let me know if you were
thinking of any specific cases and I'd be happy to follow up.

[dtu, 2014-04-24 18:54:57]

lgtm

Very cool.

[tonyg, 2014-04-24 18:58:03]

The CQ bit was checked by tonyg@chromium.org

[commit-bot: I haz the power, 2014-04-24 21:49:57]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/bulach@chromium.org/258623002/20001

[commit-bot: I haz the power, 2014-04-25 08:37:57]

Change committed as 266160